|
|
|
Saturday, 1 July 2017 | Dereel | Images for 1 July 2017 |
Top of page | ||
next day | ||
last day |
Still more anniversaries
|
Topic: history | Link here |
Coincidentally, this month includes a number of life-changing anniversaries, this year neatly laid out in a 5 and 15 year framework. For once in reverse order:
10 years ago today I retired.
20 years ago, on 18 July 1997, we moved into Wantadilla, also marking the end of a 3 month move of domicile from Germany to Australia.
25 years ago today I officially left Tandem Computers and became a consultant.
35 years ago today I officially started working with Tandem Computers, in the process moving house back to the Frankfurt am Main area. In the process I met my wife Yvonne.
50 years ago I was towards the end of a 3 month move of domicile from Malaysia to Germany.
55 years ago I took not one, but two scholarship exams for King's College Taunton: one for academic performance, the other for music. I didn't get either scholarship, but I did get what they call an “exhibition”, something like a second rate scholarship, for both. I think I was the only person to manage that, and it seems that it's indicative of me: jack of all trades, master of none.
In this month 60 years ago I returned to Australia after spending the previous 3 years in Malaya and subsequently a trip round the world. My best guess was that this happened on 17 July 1957, coincidentally 40 years to the day before we moved into Wantadilla.
Even more vaguely, about 65 years ago I first moved house (with my parents, of course), from Nunawading to Parkville. My guess there is early October 1952.
Newtonian mechanics?
|
Topic: technology, opinion | Link here |
Yvonne has surprised me by taking a liking to the The Big Bang Theory TV series, apparently because the character Sheldon Cooper reminds her of me. I don't know whether I should be flattered.
Yesterday we watched a series in which Sheldon tries to explain “physics” to Penny, the straight girl. In fact, it was Newtonian Mechanics. And clearly Sheldon didn't explain it in a way that a non-techie would understand.
But that left me wondering. Clearly I understand Newtonian mechanics well, but
what are the laws? How are they stated? The only one I could remember was
“For every action there is an equal and opposite reaction”. It kept me awake at
night, and this morning I went and checked in Wikipedia.
No surprises about what they meant: they form part of the basis for my understanding of the world around me. The question was how they were stated. Wikipedia states this one (the third) as:
When one body exerts a force on a second body, the second body simultaneously exerts a force equal in magnitude and opposite in direction on the first body.
Somehow I like my version better. Of course, Newton didn't write either of those. In 1687 the language of science was Latin, so it reads:
Actioni contrariam semper et æqualem esse reactionem: sive corporum duorum actiones in se mutuo semper esse æquales et in partes contrarias dirigi.
But somehow I'm still not happy. The laws themselves are simple, elegant and (unlike the old saw) correct, at least at normal speeds. Why is it still so difficult to formulate them succinctly?
By sheer chance, while attempting to tidy up the music room, I came across my old maths
notes from late 1965. Not surprisingly, what I wrote there matches my formulation above
almost exactly: only the initial “for” was a “to”, which clearly makes more sense: thus the
crossed out for above.
But look at the first law, for example. We have several choices:
Surely there must be a more elegant way of formulation than those. A good start would be to consider motion to include zero velocity. But there should be more. It should be possible to reduce the length of the statements by half.
Towards a new teevee
|
Topic: technology, multimedia | Link here |
Yesterday's issues with old TV series made one thing more important: bring tiwi, my TV computer, up to date. For reasons that I don't understand, the sound of some YouTube videos isn't reproduced correctly when I download the videos. Yes, I could try to find the cause and fix it, but it's an old, slow machine, the currently installed system is well over a year old:
In these modern times, that's bad: it's too old for the Ports Collection. I have a faster machine waiting, so I should take the opportunity to upgrade it as well.
I've been keeping a virtual machine eureso up to date with FreeBSD STABLE. The intention is to make it the next eureka, but it could also serve as the basis for the new tiwi. That has a number of advantages: firstly, the new machine (to be called teevee) will really be new hardware, so it can coexist with tiwi until it's ready for prime time. And secondly it can serve as a test bed for my upgrade methods.
So: find an old disk, put it as a second disk in one of my test machines, and copy the root file system, which includes /usr and /var, from eureso. Grab the “new” hardware, blow the cobwebs out and install the disk. Boot.
And it Just Worked. Of course, I had forgotten to change /etc/rc.conf and other config files, but I was able to do that later.
What next? X, of course. But there I run into trouble. I don't have a graphics card for this machine, just the on-board graphics that X can't run at more than 1280×1024. So basically all I can do so far is to adapt it to be a look-alike of tiwi, documenting the steps as I do. I'm in no hurry.
Amplifier: repair or replace?
|
Topic: multimedia, opinion | Link here |
The term High Fidelity is an old, worn-out magic word. Fifty years ago every man and his dog were interested in HiFi. Now it's just the people who want to find a justification for their solid gold speaker leads, and the buzzword is “portable”.
Since November 2015 the audio from my computer is a conventional computer “speaker” (modern term for “loudspeaker”). It works, sort of: it only has one input, of course, and the cables are not really long enough to span my monitors. And the preferred way to operate the controls is via a graphic program, requiring mouse action and having it visible at all times.
Previously I used my old Hi-Fi amplifier, a Technics SU-3500, which I got some time in 1979. It was a good amplifier, but clearly ancient. Still, it had a number of advantages: manual controls (so I can turn the volume off when I get a phone call, for example), significantly better sound quality, multiple inputs, and a convenient size to put a monitor on. Its big disadvantage is that it's defective, but it's still there in its function as a monitor stand.
Still, there are others like it for sale, on eBay for example. Went looking. They're surprisingly expensive, typically over $100, and there's no guarantee that they will last long. Is it worth repairing my amplifier? Off looking on the web, and found the service manual. It's not very detailed—it would be nice, for example, to have some debug procedure or at least typical voltages at various places. As it is, the circuit diagram, at least printed on A4, is barely legible. Still, it's worth considering. In a unit of that age, I'd guess that the problem is a faulty electrolytic capacitor.
I later magnified the manual so that it's legible. And yes, there are indications of procedures and voltages. So it's quite reasonable after all.
Netflix programme
|
Topic: multimedia | Link here |
Yesterday I grumbled that I couldn't find out what Netflix had on offer. Jamie Fraser gave me this link, which appears relatively useful. And it confirmed what I had expected: almost no old TV series (specifically, all of five series from before 1983, none of which I would consider watching).
Sunday, 2 July 2017 | Dereel | Images for 2 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Half-hearted system configuration
|
Topic: technology, multimedia, opinion | Link here |
Did very little work towards configuring teevee today. Tried to build a new world and kernel, but it died with obscure error messages. It wasn't until I was looking for something else that I found this in /var/log/messages:
And that wasn't surprising: I didn't have an entry for swap in /etc/fstab.
And X? Am I really limited to the VESA drivers? A bit of looking gave me:
=== root@teevee (/dev/pts/3) /etc 9 -> pciconf -v -l
I know that there's a Radeon driver (of course), but I've always steered clear of Radeon: there are always surprises with more than one display, even with the same manufacturer, and I've found how to deal with nVidia. The nVidia driver is called x11/nvidia-driver, so clearly the Radeon driver has an analogous name. But I didn't find it. Oh well, better things to do.
Blast from the past
|
Topic: technology, opinion | Link here |
While upgrading the ports on teevee, I saw:
That's doubly amazing. First, Pilot link is a program for communicating with Palm Pilots, and I haven't used one of them for decades (the last reference was 5 September 2001, nearly 16 years ago). So this port has been slipping by with every upgrade since then.
The other thing is: why has there been an upgrade? Do people still use this stuff? No, it highlights one of the biggest issues with modern software, shared libraries:
In fact, it seems that the program itself hasn't changed since 2010, but since then there have been 25 updates to compensate for changes in the support software.
Understanding rolling shutter
|
Topic: photography | Link here |
“Rolling shutter” is a term applied to electronic shutters that roughly corresponds to the issues with focal plane shutters using shutter speeds faster than their transit time: the image is read out sequentially, so one end of the image is older than the other. This can cause distortion if the subject is moving. But this article raises it to an art form:
Pro Capture in practice
|
Topic: photography | Link here |
“Pro Capture” is the intuitive Olympus term for the high speed pre-capture technology in the Olympus OM-D E-M1 Mark II. You half-press the shutter, and the camera starts processing images at up to 60 images per second. At some point you press the shutter fully, and the camera saves the previous <mumble> images and another <mumble> images afterwards.
How many? The “instruction manual” is not saying. According to this page it's up to 14 before. You have to go into the menu to find out the real choices.
There are two versions: “ProCap L” at (presumably) 30 frames per second, and “ProCap H” (here) with 60 frames per second. And indeed, it's between 0 (what sense does that make?) and 14 frames before, and the rest of the total of 25 shots after. Admire the scale:
Nice idea. But what can I do with it? An obvious choice is Chris Bahlo's archery. Over this afternoon for a few trial shots.
The good news: it works. And I got many shots that I could never have got before. Clearly there are things I can learn that will improve the matter. The first is a way of stepping through the images, and I'm sure I can work out some JavaScript code to do that. In the meantime, here a couple of examples. They only work with JaveScript enabled, which in particular eliminates RSS. Run the cursor over the first image to see the difference:
|
|
This one was taken at 1/320 s, clearly far too slow. So for the next shots I raised the ISO sensitivity to 33°/1600, which enabled me to use a shutter speed of 1/2500 s:
|
|
|
|
|
|
This series also shows how the arrow hits the target and then straightens out. That's still something that I don't understand.
For a first attempt, I'm quite happy. But clearly there are things I can improve on. First, even 1/2500 s is only barely fast enough: the arrows aren't completely sharp when moving. Then there's the issue of display. I really need that JavaScript function to be able to move back and forward, rather than multiple displays. And that requires aligning the images. I hadn't expected the viewpoint to change that quickly. And last, sadly: 60 frames per second aren't really enough for this kind of image. The arrow moves too fast, so I don't get enough images. Looking at the first of the last collection of images, it looks as if the arrow moves about 50 cm in 1/60 s, or about 30 m/s (108 km/h, not really that fast). Still, it opens a lot of possibilities.
Monday, 3 July 2017 | Dereel | Images for 3 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Mouse pain
|
Topic: technology | Link here |
Yvonne and I have been using cheap wireless mice for the last three years, but it seems that they're gradually passing their use-by dates. Mine has taken to producing spurious clicks, resulting (usually) in copying of unintended text to the cut buffer. Today Yvonne's started producing button 9 events when any button was clicked. Time for replacements.
But what? OfficeWorks has a large selection—I bought these mice there—but an absolutely appalling web site. After checking my location with Google Location Services, which thinks I'm in Ballarat (another story), it decided that I'm in Traralgon, 300 km away. More to the point, though, there is no information about the mice, just photos, prices and random other details:
|
JB HiFi is just as bad. If I want to find out what these devices are, I'll have to go and take a look at them. At the very least I need a button 2 (middle button in the good old days) that is easy to use, which (for me, at any rate) doesn't include combined button 2 and scroll wheels. And none of these pages even tell me how many buttons the mouse has!
X on teevee, continued
|
Topic: technology, opinion | Link here |
So the motherboard graphics on the new teevee is an "RS880 [Radeon HD 4200]", so presumably I can use an ATI driver to run it at natural resolution. Where do I start? Yesterday I established that there's no obvious counterpart to x11/nvidia-driver, so off to read the manual. Yes, they're supported. The difference between the two is worth mentioning:
AMD® Radeon
2D and 3D acceleration is supported on Radeon cards up to and including the HD6000 series.
Driver name: radeon
NVIDIA
Several NVIDIA drivers are available in the x11 category of the Ports Collection. Install the driver that matches the video card.
That makes sense: only the nvidia ports are mentioned. Is radeon in the kernel? Yes! Not one module, but no less than 82 of them, most with names like radeonkmsfw_PITCAIRN_pfp.ko, radeonkmsfw_PITCAIRN_rlc.ko and radeonkmsfw_RV770_me.ko. Presumably they refer to specific chip sets, though there's no radeonkmsfw_RV880_me.ko. Hopefully the base radeon.ko would know what to do. So I loaded radeon.ko:
As the log messages suggested, I ended up with another module too:
=== grog@teevee (/dev/pts/0) ~ 3 -> kldstat
OK, try starting X again. No difference. OK, /root/xorg.conf.new still referred to the vesa driver, so I changed it to radeon. Still no go. From /var/log/Xorg.0.log:
Are we talking about two different drivers? Tried changing it to nvidia instead (the system does have the nvidia driver installed). Clearly it wouldn't work, but the error messages were interesting enough:
That's a very different message, and it shows where it was loaded from. If there's a /usr/local/lib/xorg/modules/drivers/nvidia_drv.so, there should be a /usr/local/lib/xorg/modules/drivers/radeon_drv.so too. But there isn't.
Back to the manual. Really no mention of any external drivers. Then Peter Jeremy came up with the idea to install the xf86-video-ati package. So I tried that (never mind that there's no mention of it in the manual). And how about that, it loaded!
And crashed:
What's that? Went looking on the web, and it's a not unknown bug. Ubuntu has a bug report about it, open for some years now. The discussion, which I only read superficially, suggests that it might be triggered by specific hardware.
Do I care? My intention is to put the nVidia card from tiwi into this machine. But maybe I should care. I've been configuring X for 25 years now. Over 20 years ago I wrote documentation about how to get it to work for your configuration. And here's relatively mainline hardware that has given me the best part of a day of grief, not helped by inadequate documentation.
ANZ security
|
Topic: general, technology, opinion | Link here |
We recently received new ATM cards, complete with an almost non-removable sticker saying that they first needed to be activated. I could do that online, so I braved the ANZ web site and did so for mine. Then it offered me the option of activating Yvonne's card too. OK, did that. All over and done with.
Except that the card didn't work. It seems that their validation is broken. OK, try again. But this time I got the message:
We are unable to activate your card online
To activate an ANZ Credit Card or ANZ Everyday Visa Debit Card, please call 1800 652 033.
Sigh What nonsense have they thought up now? So I called up the number, which proved to be a computer. It wanted “my” (Yvonne's) date of birth in DDMMYYYY format. OK, I know that. Typed it in, along with a # so that the computer knew that the input was complete. “Thank you, your card is validated”.
I've complained about ANZ “security” before, but this really seems to be an escalation. Anybody who steals Yvonne's wallet would get the card and her driver license, which shows her date of birth. So what is the point of the activation in the first place? And why does it have to be done over the phone (normally unencrypted) rather than via a secure web session?
Yes, activating the card doesn't mean that you can use it. You still need the PIN for that. But that begs the question what use activation is in the first place.
When will these people learn?
Indianmeal moth
|
Topic: animals, photography | Link here |
We still haven't got rid of indianmeal moths (which, bizarrely, Wikipedia prefers to call “Indian mealmoth”, as if it had something to do with India). Today I found one on a wall, time for some photos:
|
Somehow it's doesn't quite look like the one in Wikipedia:
Is this a different species, or maybe a different moth?
Tuesday, 4 July 2017 | Dereel | Images for 4 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Google location services, revisited
|
Topic: technology, opinion | Link here |
While I was trying to find out why Officeworks thought I was in Traralgon, I received a strange popup: “Edit work address”. Why work address? What does that have to do with work? Checking showed that it thought I was in Ballarat, in fact not far from where Google Location Services put me the week before last. Could that be part of the problem? In any case, for the first time I've found a way to tell the system where I am (something that I ranted about, and others didn't understand, last time). OK, enter the address and try to measure the distance from there to Traralgon:
|
|
San Diego CA? That's at the other end of the world! How on earth did it come up with that? And on the next attempt, it forgot home altogether:
|
Finally it was done, but the location services still think I'm in Ballarat.
X config insights
|
Topic: technology, opinion | Link here |
I really don't want to untangle the can of worms that is the X radeon driver, but I feel I should.
The first question: according to the documentation, the driver should be installed automatically. Clearly the documentation is insufficient, but is there more? It proves that the X port is a metaport: it just ensures that numerous subports (some of them themselves metaports) are installed. From /usr/ports/x11/xorg/Makefile (almost all of it):
OK, clearly I need xorg-drivers. And its Makefile contains:
So ATI is the option to install the ATI drivers. And by default it's not set! The same applies to everything except VESA. No wonder I always end up with the vesa driver.
That doesn't help much, though: I still get this undefined symbol exaGetPixmapDriverPrivate. It didn't take much searching to discover that it's a function in exa, a 2D acceleration architecture, along with documentation that has since gone into hiding and a man page, exa(4), which gives some configuration parameters, but refers to xorg.conf(5) for details, and of course there there are none. One of the options in radeon(4) is:
That looked promising, but setting it either way had no effect.
In the end I gave it up and tried rumours that the vesa driver can handle resolutions higher than the 1280×1024 that it provides by default. The radeon driver had at least got as far as establishing a mode line for the chip/monitor combination:
That's what it writes into /var/log/Xorg.0.log, but of course xorg.conf doesn't want to know that. It needs to be trimmed to:
And indeed, that had an effect: it lowered the resolution to 1024×768. Nothing obvious in the log files.
I knew why I didn't want to get involved in this stuff. But maybe I should read this page, which is addressed towards driver writers. If there's a bug in the radeon driver (something I'm not convinced of), it may help me to find it.
Too windy for Strelitzia nicolai?
|
Topic: gardening, opinion | Link here |
Our Strelitzia nicolai pups aren't looking obviously unhappier than they were when transplanted last week. But the leaves of the bigger one are looking a little tatty:
|
Is that the way they were last week, or is it too windy where we planted them? To be observed.
Wednesday, 5 July 2017 | Dereel → Ballarat → Dereel | Images for 5 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Shopping in Ballarat
|
Topic: health, opinion, technology, multimedia | Link here |
Into Ballarat this morning for various things. First to Dorevitch for a blood test, which was something of a record: 4½ minutes from entering the facility to leaving it again. In other places I can wait up to an hour.
Then dropped in to JB HiFi to look for mice, of which they had very few. While I was there, also looked for stereo amplifiers, which weren't much different from what they have on their web site. But I found out why: “Stereo” is an old, worn-out magic word: nowadays it's called “5.1”, and the amplifiers are, if anything, less expensive:
|
Still, more than I want to pay, especially since the user interface to most modern amplifiers is less than easy to use.
On to Officeworks, where I had the pleasure of watching a salesman checking details of laptops via their web site. In the end, I think, he gave up. I was looking for mice, and that was enough fun that I describe it separately below.
Then on to the hairdressers. If there's one thing I hate, it's waiting in for my turn. I was lucky today with the blood test, but in general I have to wait for the hairdresser, and on occasion I have just left again. But today was my lucky day: the place was empty, and I had my hair cut immediately. A case of lucky timing: while I was there, three other people arrived.
NBN ready phones
|
Topic: technology, opinion | Link here |
I'm vaguely on the lookout for new cordless phones: the ones I have are getting a little flaky, and the one that is supposed to support Bluetooth is effectively useless because I can't adjust the volume. So while at JB HiFi and Officeworks I also looked at phones, without finding anything exciting.
Except, maybe, the claims at Officeworks: “NBN ready”. What is there about a cordless phone that is NBN ready? I was going to ask, but forgot. I'll have to do so next time.
Fake road works
|
Topic: general, opinion | Link here |
On the way back through Cambrian Hill, found yet another case of road works that weren't. Signs to slow to 40 km/h, and some road works on a small side road. But nothing at all on the main road apart from the signs.
While I was thinking “there should be a law against this”, I passed a police car coming in the other direction. What speed was I doing? Under 60 km/h, I'm sure, but was it below 44 km/h? I would be really pissed off if I got a ticket for that.
So: let's gather some evidence. Firstly, the signs didn't have any ”end of speed limit” on them:
|
|
That white area at top left should show a new speed limit, but there were none.
Then the “road works” themselves, seen from the other end (the other side of the sign in the previous images):
|
Yes, there's a brow on the hill there, but I couldn't be bothered going back beyond it. There was nothing to see.
Still more Epacris
|
Topic: gardening | Link here |
Going through Enfield State Park, it seems that there are more Epacris impressa all the time:
|
|
|
I've been driving along this stretch of road for over 10 years, and I've never seen anything like this before. Have they been planted? If so, why are they red and not pink?
GPS tagging Olympus photos
|
Topic: photography, technology, opinion | Link here |
On the way home, there were a number of photos that could have done with GPS tagging: the road works and the wildflowers. But how do I do that? Tried pairing with my mobile phone, which worked well in absence of other networks. And then? OI.Share has an “Add Geotag” function, but how do you use it? Selected it and was told that “this will take a long time”, but after about 30 seconds it seemed to be happy. Clearly that can't have anything to do with the photos I'm about to take, and I can't take them with the camera until I disconnect from the phone. And if I stay connected, it creates JPEG images as well.
So I took more photos, and they came up without GPS information in the EXIF data. All I have is this, which is what I get when I press on the “Add Geotag” tag:
|
It seems that that at least tells me where I was when I took the photo. But how do I get the information out? My first reaction is “RTFM time”, but I have little faith in the FM.
More mouse fun
|
Topic: technology, opinion | Link here |
At Officeworks I found a mouse that I hadn't noticed before, because nothing in the description suggested that it had more than two real buttons (not counting the scroll wheel). But for $19 I bought a “J.Burrows Wireless Comfort Mouse Black” (is this a mouse or a vole?), with the description:
The J.Burrows Wireless Comfort Mouse is designed to make you more comfortable when using the computer. The shape of the mouse fits naturally in your hand to reduce the risk of injury and the design gives you greater control over your browsing with integrated buttons.
It has a comfortable design for your convenience.
It is powered by a built in Lithium Ion battery.
It uses 2.4 GHz wireless technology for strong connections.
Connecting the mouse to your computer is easy thanks to the USB receiver.
The mouse has a built in 4 way scroll wheel for navigating up, down, left and right.
You can enjoy quick access to the Windows start screen.
With the exception of the second and sixth items, this applies to all mice currently available. But on the side there are two buttons where I would expect them:
|
Why do they have keyboard symbols? Probably a function of the keyboard driver, but there's nothing on the package to clarify that:
|
Yes, it offers a “Windows” key, but says that it's compatible with Apple. How can that work? They don't say, of course.
To be on the safe side, took a couple of Logitech mice as well, an M560 EXTRA and an M705 MARATHON (presumably with longer legs), with the intention of only keeping one of them.
First the “Comfort Mouse”, presumably not Korean. It comes with almost a bare minimum: charger cable, receiver and mouse. No driver. No instructions. Nothing informative on the packaging. A non-technical person would be lost. But then, mice are an old, dying breed. Nowadays you smear glass.
Plug in the receiver:
Eight buttons are more than enough. But the side buttons did nothing whatsoever. xev didn't see any input. What can I do to fix that? Nothing, I think. Life is too short. Let's try Logitech instead.
The M650 EXTRA comes an even barer minimum: battery, receiver and mouse. No driver. No instructions. Nothing informative on the packaging. A non-technical person would be lost. But then, mice are an old, dying breed. Nowadays you smear glass.
And this time the buttons did something. They generated keyboard events. The front one generated a keycode 116 (Ctrl_R), and the rear one generated:
Interpreting, this is press key 115 (`, the “grave” character), then press key 40 (d) and then release them in the reverse order. What does that mean? About the only thing we could find on IRC was that the product page contains the ominous:
No matter what you’re doing, you’re always just one click away from the things you use most, like running apps, search and settings. A button located right behind the scroll wheel gives you one-touch access to the App Bar and Charms Bar.
You get one-touch access to the Windows® 8 App Bar, Charms Bar, Desktop and Start Screen.
Full-size, comfort-first designDeep, contours. Soft, rubber grips. Whether you’re a righty or a lefty, your hand is cradled naturally and comfortably—hour after hour.
Admire the punctuation.
Windows 8 edge menu button Two extra thumb buttons
So is this thing designed for “Windows” 8? Not according to the packaging. About the only information of any kind states “Windows® 10 or later... Linux® 2.6+”. But it also includes “Windows” 8, presumably later than “Windows” 10.
How do these things work? It seems that either the mouse driver or the dongle can somehow generate keyboard events, though it's not clear how. Daniel O'Connor suggested using usbdump, an analogue of tcpdump, to see what the dongle is saying to the system, but that was mainly polling, making it almost impossible to see what was really going on.
Yes, I could try wireshark, but it's gradually becoming more trouble than it's worth. Somehow I've been having trouble for the last 20 years with people redefining mice functionality to limit it to operating systems that I don't use. Yes, the scroll wheels are a good idea, but in general they've only implemented them in one direction. And in the process they've castrated button 2, which I have simulated with one of the thumb buttons. And now the thumb buttons aren't buttons at all! GRRR!
Of course, there's one more button, the little square one behind the scroll wheel. Could it (desperate hope) be a button 2? No, it latches something. On first press, I get:
Interpreting, this is press key 64 (Meta_L), press key 115 (`, the “grave” character), and then release them in the same order. It doesn't alter the function of the other buttons. This is similar to the function of the rear thumb button, but it releases in reverse sequence. And if I press it again, it just simulates button 1:
No idea what that is supposed to be good for, but it's not good for anything for me. Went looking on the web, where some Linux people have found a way to live with them, but in general the feeling was “not very compatible”, despite the claims on the packaging.
Only then did I look more carefully at the log messages:
The first time I only found what I was looking for, the last 3 lines. Yes, indeed, it's simulating a keyboard as well. O brave new world, that has such technology in 't!
Thoroughly frustrated, I left it alone for a while. Some hours later, though, I decided to open the M705 MARATHON anyway and see how it worked, though it looks almost the same as the M560 EXTRA (on the left):
|
That probed as:
Another damned keyboard! But I tried it anyway, and it worked out of the box! The side buttons appear to generate button 4 (rear) and button 9, though I had mapped button 4 to button 2. It's not clear what use the keyboard port is.
Something about the mouse seemed familiar, in particular the strange idea of inserting two batteries in parallel, presumably to keep it running 18 months or two years, depending on what you read. I have a mouse like that, but it's defective. Yes, how about that, it's the same model. The only problem is that, since I dropped it on the carpeted floor from all of 70 cm height, and despite my repair attempts, the scroll wheel no longer ratchets, which is very irritating.
And what does the square button behind the scroll wheel do? How about that, nothing. No mouse event, no keyboard event. But then I discovered that the scroll wheel no longer ratcheted. That button controls the ratchet. OK, look at the old mouse. Yes, indeed: the non-ratcheting isn't a bug, it's a feature. So now I can use that one again too.
Comparing the log output of the first mouse from 4 years ago and today's, there's really almost no difference, apart from the different device numbers:
The only other difference is the revision number: 24.00 then, 12.03 now. Why are they counting down?
In summary, then:
Two more power failures!
|
Topic: general | Link here |
Power failure at 15:09, mercifully short. So I headed to the office to write it down. And then there was another one! Normally that's a precursor to a longer failure, but fortunately that was all, apart from stress on my part.
Thursday, 6 July 2017 | Dereel | Images for 6 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Mobile phone battery: wrong again
|
Topic: technology, opinion | Link here |
Last week I had ordered a replacement battery for Petra, and she came along again today with her phone. Six months ago I had managed to buy the wrong battery model, so this time I had been particularly careful, but it still made sense to check that it's the right battery before opening it.
|
Again! How the hell did I manage that? It certainly doesn't help that Samsung has two different names for each model: the marketing name (“Samsung Galaxy 2 고대”) and the name on the phone itself (GT-I9200T):
|
Once again, it might fit, but this time I think I owe it to Petra to get the right battery.
Pension red tape
|
Topic: general, opinion | Link here |
The German word for “red tape” is the more descriptive „Papierkrieg“, “paper war”. And that's what we got today: no fewer than four letters about our pensions. Our German pensions have been raised by 28 € per month, but for that we need to prove that we're still alive, to their satisfaction. This means going in to the police station in Ballarat and showing our driver licenses to a complete stranger so that they can confirm that we're alive. I've tried in the past to get the certificate signed by my doctor, but what would he know?
The other two letters were from Centrelink: it seems that we are to receive a payment, the sum of which we need to work out for ourselves, since their computers clearly can't do it:
|
|
But of course I know how much it is, because it was paid ten days ago, exciting much concern about a drop in my pension in the process. And I had already established that I should have received a letter in advance.
So how come I only got the letter now? Australia Post again? No, the letter was dated 26 June 2017, the day of the payment. And for once Australia Post has helped: it seems that they're putting date stamps on letters again, although this one landed across the window on the envelope:
|
So the letter wasn't even posted until 8 days after the payment. Good idea, lousy implementation.
Cha Shao, ALDI style
|
Topic: food and drink | Link here |
A while back we bought a “Chinese Style Sticky BBQ” marinated pork from ALDI. The photo made it look like Cha Shao, and the price was right, so we bought one.
What do we do with it? Cha Shao is an ingredient, not a dish in itself. Decided to cook it and use some of it for noodles. That matched the appearance of the meat, not quite the way it looked on the packaging:
|
|
And yes, despite the attribute “sticky” (which it wasn't), it was quite close to cha shao, just not as red (nor as red as on the packaging). Probably worth the trouble.
Friday, 7 July 2017 | Dereel | Images for 7 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Delta visits: an end
|
Topic: animals, general | Link here |
Yvonne and Sasha off today to the Geoffrey Cutter Centre, for the last time as it proved: Kath has found somebody else to start next week. Although that doesn't apply to Nikolai and me, we still don't need to go again. So we're done. And, it seems, not too early: Sasha was really not happy today.
More mouse fun
|
Topic: technology, opinion | Link here |
Mail from Daniel Nebdal about my mouse fun: the Logitech receivers are designed for both keyboards and mice, so there's no extra hardware cost involved in adding keyboard functions to mice. That makes sense.
Still, I now have three mice, two of which have side buttons that don't work with X, at least not without some unspecified configuration. But I have a Microsoft laptop too, euroa, and the mouse I was using for that is a Logitech m705 (as it was called in the days before they grew up to be M705 MARATHONs). So why not try the J.Burrow comfort vole? Callum Gibson opined that, though no driver was delivered with the device, Microsoft would go out and look for one.
OK, we can try that. And, sure enough, it did. One point for Microsoft. And what do the side buttons do? Nothing! The down side with Microsoft is that there's no documentation for the driver that I can find. Effectively the side buttons are useless. I didn't really want them on Microsoft anyway, but why should I pay for something that I don't want? One point lost for Microsoft.
OK, there's still the Logitech M560 EXTRA with the keypress-creating side buttons. How does that work?
First thing that surprised me was that it didn't load a driver. Did it like the vole driver? Anyway, one of the side buttons (probably only coincidentally the same one indicated on the vole) now simulates the broken window key on modern keyboards and invokes the “Start” menu. The other one still does nothing.
And, of course, there's no documentation whatsoever. How am I supposed to find out what the other button does, let alone use it? I'm baffled. About the only indication I have is the symbol on the corresponding button of the vole, a running human. It reminds me of the symbol on emergency exits:
|
Is that a warning?
Still more mouse problems
|
Topic: technology, history, opinion | Link here |
In the meantime, I've found a case of nausea with the M705 MARATHON: when moving the mouse cursor from one window to another via a third, unrelated window, I have frequently seen it paste the copy of the cut buffer into the third window. That's annoying at best, and downright dangerous at worst. This is what happened with a similar problem (different mouse) decades ago:
Date: Sun, 22 Apr 2001 15:45:06 +0930
From: Greg Lehey <groggyhimself@lemis.com>
To: FreeBSD Core Team <core@FreeBSD.org>
Subject: Grrr. I've accidentally given out the key to -announce
I *hate* this mouse. It just went and did this all over #BSDcode:
> o post to announce (as I'm
> sure I have said on a couple of occasions, the first line needs to be:
> Approved: letmeout
> and it should go. Don't cc: ot bcc: it. Send a seperate copy if you want
> to.
<cmc> groggy: Space before a period?
* groggy-P murders his mouse, grinds it up and swallows the remains.
<green_> nice paste there
<kkenn> groggy-P: just commenting
* groggy-P blushes crimson.
We'd better change the password as soon as possible.
This was the (relatively uninteresting) password for posts to the FreeBSD-announce mailing list. As a result of that incident, Jordan Hubbard came up with a new password:
Date: Sat, 21 Apr 2001 23:23:17 -0700
From: Jordan Hubbard <jkh@osd.bsdi.com>
To: core@freebsd.org
Subject: new announce password
Oh yeah, for those too lazy to look in the config file on
hub, the new -announce approval password is "groggydidit"
New Hugin
|
Topic: technology, photography | Link here |
The newest version of Hugin, 2017.0, has been released. For once it was relatively trivial to update the port, and all seems to work. Hopefully it'll stay that way.
GPS tracking with OI.Share
|
Topic: photography, technology | Link here |
More playing around with OI.Share today to try to understand GPS logging. In fact, it's not difficult. All you need is documentation. I still don't have that, but I'm working on it. In the meantime, it seems that the app (or the phone) maintains a location log, and at some later point you can connect the camera and get it to update the EXIF information of the photos in the camera. That's how I managed this photo, for example:
|
The image information (run mouse over the image) shows the coordinates, and the larger (“small”) image includes the link: Display location on map.
The next thing I need is a way to add the location information manually. That shouldn't be that difficult.
Saturday, 8 July 2017 | Dereel | Images for 8 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
eBay: Pay! Pay! Pay! Pay!
|
Topic: technology, opinion | Link here |
So I've decided not to sell anything on eBay again until people can convince me that they have their act together. So what happened over the last couple of days? No fewer than four invoices!
Do they include it in the email? Of course not! That would be too easy. Admittedly, since they don't seem to understand encryption, it would also be insecure, but there's a simple solution to that non-problem. But when I look at the invoice (there's only one), it's what I would expect: nothing to pay.
Fifty years ago (really) we laughed about people getting bills and then payment demands for the sum of $0.00 or less. In those days, computerized billing was in its infancy. It seems that eBay hasn't caught up with the 1960s yet.
Hugin detect failure
|
Topic: photography, technology, opinion | Link here |
House photo day today, and a miserable start it was. It wasn't until I got inside that I discovered that light drizzle had marred many of the photos:
|
Still, part of the point of taking this series is to show the garden in different weather situations, so I started to stitch it. But things didn't turn out the way I expected. This “panorama preview” should show something like the final panorama:
|
What went wrong there? Looking at the control points, everything seems to have been recognized correctly, but two of the images have been inverted (only one shown here):
|
Something wrong with the new version (2017.0) of Hugin after all? Spent some time reinstalling the previous version (2016.2), not helped by the fact that the supporting libraries no longer existed. And when I did, exactly the same thing happened. It wasn't just this panorama: all the dull panoramas had similar issues. I suppose that's some relief; I wouldn't have relished looking for the reason why the new version didn't work.
I thought it might have something to do with the end of the focus rail that's showing in the bottom right, but looking at the control points, they were all relevant and detected correctly. There was another issue in play: these images are each merged from three individual images exposed 3 EV apart, nominally -2 EV, +1 EV and +3 EV. Normally I use a custom DxO Optics “Pro” profile to convert the component images, but today I used the standard. But why should that make any difference? In passing, it's worth noting that Hugin decided that the lens (Olympus Zuiko Digital ED 8 mm f/3.5 fisheye lens) had a focal length of 10.487 mm. It always alters that focal length, but normally to 7.7 mm or so. But probably this is a result of the underlying problem, not the cause.
In the meantime the sun had come out, and I took another series of photos. That worked fine:
|
But clearly there's something funny going on here, and I need to investigate it. I'm putting the source files and my .pto file at Day/20170708/old-garden-pano.zip for the Hugin community to look at.
Trout or salmon?
|
Topic: food and drink, general, opinion | Link here |
Chris Bahlo along for dinner this evening, as usual on Saturdays, bringing Margaret Swan with her. Yvonne had planned a soufflé for the first course, and had admonished them to be on time. And for once they were not just on time, but early.
There was a reason: due to the unseasonally dry winter, Chris was out of water, and they needed to use our showers. But we still just about made it.
Main course was trout, though the size suggested it might be a salmon trout. It's something that requires careful dismemberment:
|
Clearly the trout was not impressed:
|
|
But then, it might just have been because Chris and Margaret drank red feet wine
with the fish.
In passing, it's interesting to note the pattern on the gratin dauphinois that Yvonne served with the fish:
|
It was baked in the oven, of course, so you'd expect it to be browned evenly. I wonder what caused that.
Sunday, 9 July 2017 | Dereel | Images for 9 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
teevee: next step
|
Topic: technology, multimedia, opinion | Link here |
I've been busy with other things lately, so configuring teevee, my new TV computer, has taken a back seat. I've given up trying to configure X with native resolution, and the next task (maybe the last) is to configure LIRC.
Things didn't start well: I've mislaid the second USB receiver that I had planned to use, so I had to remove the one from tiwi. And, as half expected, nothing happened. Starting lircd was straightforward enough, and it recognized when the mplayer process started:
=== root@teevee (/dev/pts/6) /home/grog 6 -> /usr/local/sbin/lircd -n --driver=dvico --device=/dev/uhid0
mplayer said:
Strange. On tiwi I don't have that file, but I do have a file /etc/lircrc. Why didn't it try that? In any case, it was easy enough to copy a .lircrc to my home directory. Start again, and the error message no longer appeared. But that was all. Still no reaction to the remote control.
Further investigation showed that the batteries were flat. OK, new remote control with functional batteries. No reaction to the remote control.
OK, how does tiwi do it? Checking recently accessed files (ls -lu) on tiwi showed:
So more comparison, and somehow my general aversion took over, and I moved on to more pleasant things.
Propagation progress
|
Topic: gardening, opinion | Link here |
I've taken a number of cuttings of our Hibiscus rosa-sinensis over the last few weeks, mainly because the bush needs pruning. It's the wrong time of year to propagate them, but it was worth the trouble anyway. I've also done some begonias that we got somewhere—Lorraine Carranza, I think. To maximize my chances, I put some in the lounge room and some in the bathroom, which has about 2 EV less light.
Progress has been mixed. The biggest difference is in the begonias. The ones in the lounge room are looking quite happy, the ones in the bathroom considerably less so. One has died:
|
|
Why? It can't be the temperature, which is hardly different, so it must be the light. I'm gradually coming to the conclusion that light levels are much more important than people claim them to be. I'm still looking for the correct lighting for my Spathiphyllum, supposedly one of the most forgiving of plants.
The hibiscus are less obvious. I have planted cuttings over a period of time, and success rates are varied. A month ago I planted three cuttings. Here at the time and today:
|
|
The one at the top is clearly suffering, but the one at bottom left seems to be doing better, if anything.
The fourth cutting was planted two weeks later on 20 June 2017. Here's a comparison of then and now:
|
|
It, too, is not very happy, but they're all far from dead.
Curry tree recovery
|
Topic: gardening | Link here |
I've had mixed luck with my curry tree. It's 7 years old now, but it has been through several iterations of almost complete leaf loss. Last year it took months to recover. This year it started again about a month ago, presumably as a result of mites and scale insects. But this time it's recovering much better, and all the branch tips are showing new growth:
|
I need to ensure that I spray it twice a week to avoid a recurrence.
Curry tree photo
|
Topic: photography, opinion | Link here |
I'm relatively happy with this photo:
|
It was done with in-camera focus stacking with the Olympus OM-D E-M1, one of a total of 63 that I took. Why the E-M1 and not the E-M1 Mark II? Convenience, mainly. Why so many photos? The first time round I took them without flash, requiring a relatively wide aperture of f/6.3, and for that a sensitivity of 33°/1600 ISO. And that was too much depth for only 8 component images.
|
Taking the photos was an experience. It's the first time I've used my cross bar, in the process discovering that my portable tripod is too flimsy. It makes it easier to position the camera:
|
|
|
Here's a view from the side (or below, from the camera viewpoint). The subject leaf is at the very right:
|
Flash is interesting. On the E-M1 there is almost no choice of shutter speed: the lowest speed for focus stacking is 1/8 s, and the highest for flash sync is 1/13 s. That would have been a reason to use the E-M1 Mark II, which can sync up to 1/60 s with the electronic shutter, but by then I couldn't be bothered.
And once again the M.Zuiko Digital ED 60 mm f/2.8 Macro proved too long, so I used the M.Zuiko Digital ED 30 mm f/3.5 Macro. And it seems to have significant hysteresis in the focus ring, not making an already marginal focus peaking any easier.
Powerball eye damage!
|
Topic: general, opinion | Link here |
Yesterday we tidied up a couple of cupboard with a number of cleaning products which we no longer use, along with a few other things. But what's this?
|
Clearly it's a Finnish powerball. But what's it good for?
|
That's the only description, in multiple languages. “Risk of serious damage to eyes”. People, isn't that an indication that the function should be clearly defined?
Of course, it's probably something to do with dishwashing. The version for illiterates is almost more descriptive:
|
That suggests that it includes two toy cameras, a bottle of eye lotion and an unspecified sum of money.
I've ranted about this particular company before, but I can't find the article.
Monday, 10 July 2017 | Dereel → Ballarat → Dereel | Images for 10 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
To the doctor again
|
Topic: health | Link here |
Off to Ballarat yet again for a doctor's appointment. This time was different: Yvonne was with me, and we were there for an hour. Good news: my blood test results were all either in range or better than before. Even things like mean corpuscular volume, which had been puzzling us for some time, are now “essentially within normal limits” (98 fL in a range of 80 to 96, though Wikipedia claims 80 to 100 fL). About the only exception was Vitamin D, which I don't think I have had tested before: only 44 nmol/L, where it should be ≥ 50. Doubtless that's due to my indoor lifestyle. So I'm to take tablets over the winter. And apart from that, he's happy enough that I don't need more attention for 12 months, apart from an obligatory blood pressure measurement in about 6 months.
This was Yvonne's first real examination here, and he went through the same series of vaccinations as I did recently. After initially wanting to refuse, Yvonne finally accepted a total of three at a time. Here's the third:
|
The polyp she had removed 6 weeks ago proved to be hyperplastic, and thus not worth worrying about.
Ten years of Dereel
|
Topic: general | Link here |
In my mail this morning:
How time flies!
teevee progress
|
Topic: technology, multimedia | Link here |
Finished my comparison of the configuration files for LIRC today. One unexpected configuration detail was the driver specification. The startup file is /usr/local/etc/rc.d/lircd, and it contains:
The first four lines describe what parameters to load from the “real” configuration file: whether to start it at all, and if so, which device to open. And the “driver”? That's the --driver=dvico, and clearly it should be a parameter. But it isn't. At the very least I need to update the startup scripts.
After that, it worked! Well, I got correct responses from the irw program. Ran mplayer, and it still didn't react. Why? I hate this stuff.
Mouse: caught in the act
|
Topic: technology, opinion | Link here |
My new M705 MARATHON mouse still vomits over innocent windows from time to time. Today an Emacs caught it and reported:
How did that happen? With the exception of button 9, those buttons don't even exist! And according to the manual, “triple-drag” means that the button has been pressed three times in rapid succession and then dragged. Is this an issue with moused, or is it really a problem with the mouse? If the latter, how does the Microsoft driver handle it?
New glasses
|
Topic: general, food and drink, opinion | Link here |
The drinking glasses we use daily are showing their wear, though it's surprisingly difficult to get a photo that shows it clearly. Time for replacements. But the kind of glasses we used in the last century are no longer modern. New glasses are big, so that they can take up to a bottle of wine at a time, and so that they don't fit into the dish washer. We've been looking for replacements for years, for beer glasses, wine glasses and aperitive glasses.
While in town, dropped into Harris Scarfe on the off chance. No aperitive glasses, of course. There were some like the one on the right in this view:
|
But they were twice the dimensions and had an estimated volume of 400 ml, not exactly appropriate for an aperitive. We did, however, find something like normal beer glasses:
|
But they're higher, and the shape is strange. The normal ones are what you'll find in any German pub. These are made in Czechia, not exactly far from Germany, and the shape suggest that they have been modified to make them larger. But any port in a storm, so we bought some.
Where can I find normal glasses?
Tuesday, 11 July 2017 | Dereel | Images for 11 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Bread baking again
|
Topic: food and drink | Link here |
Since dropping to only two meals a day, we're not eating as much bread as we used to. Today was the first time in 5 weeks that I baked bread. It's good that it freezes and thaws well.
One issue I have always had was getting the lumps out of the flour. When I was a kid, my mother told me to sieve the flour. But with 1.3 kg of flour, that's quite an issue. My attempts to find automated sieves have failed, but a couple of weeks ago we bought a “food mill” at ALDI. It's designed for something else, such as mashed potatoes, but it looked like it would work well with flour:
|
Things weren't as good as I hoped, unfortunately. Turning the handle put about half the flour through the mesh, and pushed the other half up along the sides:
|
|
And when I was done, a lot of flour was left in the sieve and the corners of the frame:
|
And after washing in the dish washer, lumps remained:
|
What really killed it, though, was that the flour immediately formed lumps after mixing with water. So the whole thing wasn't worth the trouble.
SIM cards: pint in a quart pot
|
Topic: technology | Link here |
The credit on my old mobile phone SIM
card (0401 265 606) runs out on Wednesday. I don't like the number—I can't even
remember it easily—and I have another that I do like and can remember (0438 490
494), so it's a no-brainer to put that in the phone instead.
Not quite as easy to remember as I thought. In fact, the number is 0490 494 038.
Well, not quite a no-brainer. The new card was in my iPhone, and had thus been trimmed to Nano-SIM format. My old Samsung GT-I9100T (Galaxy S2?) takes a larger Micro-SIM format;
|
So how do I put it in? Looking at the compartment, there's a clear boundary for the card:
|
|
It's also obviously possible to put even bigger cards (“Mini-SIM”) in the phone, as indicated by the cut off corner at the top right. It's interesting to note that the micro-SIM also has a cutout, but there's nothing to align it with in the phone.
OK, how do I put a nano-SIM in there? Do the cutouts match? Looking carefully, it's clear that the large contact area in the middle serves no additional purpose: it's connected to one of the corner contacts (bottom left in the photo above), which also serves to orient it. Yes, the cutouts have the same orientation. So I put the new card in the phone. “No SIM card”. Careful examination showed that I had to align it in a non-obvious way, with the right contacts almost sticking out from the card:
|
Dying begonia: why?
|
Topic: gardening | Link here |
Peter Jeremy thought that the death of the Begonia cuttings in the bathroom might be due to different humidity levels. I didn't think so, but it was worth measuring. In the lounge room I have a pretty consistent relative humidity of about 45% to 50%. Measuring in the bathroom, I found 47% to 55%. That shouldn't be enough to kill the cuttings.
The photos show a more likely reason:
|
The light comes from the left, and that's where the one cutting is growing. It's too dark in there.
Hibiscus: time for pruning
|
Topic: gardening | Link here |
Since repotting, our Hibiscus rosa-sinensis has grown well. I had wanted to hold off pruning until the spring, so that I could use the cuttings for propagation. But it's becoming increasingly lopsided, so I bit the bullet today.
What can I do with the cuttings? Put them on the compost, or plant them. If they die, I can still put them on the compost.
So I planted one in the pot where the begonia had died (top right) and put it with the survivors in the lounge room, where there's more light:
|
Two more came into the dining room:
|
And then I planted three of them outside with differing levels of care:
|
|
|
The first was planted properly in the soil, the second in potting mix, and the third (which doesn't look up to much anyway) just planted in a pot where the previous owner had died. I'll be interested to see whether any survive.
Also put the Spathiphyllum in an area with more light, so there are now quite a few plants at the west end of the lounge room:
|
Wednesday, 12 July 2017 | Dereel | Images for 12 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Mouse: caught in the act
|
Topic: technology | Link here |
Why is my M705 MARATHON mouse vomiting all the time? Started moused with debug mode, which is very verbose. By the time I the problem recurred, I had:
The output is less than inspiring. Normally there's lots of:
I'm guessing at the meaning of this, but clearly each character is reported individually and then together (“received full packet”). Then there are some lines that currently don't tell me very much, and finally a summary that it has been interpreted as button 1. “count:1” appears to mean that the (a?) button is pressed. Then things repeat themselves. This time the first character has changed from 0x83 to 0x87, and this is interpreted as the button being released (“count:0”). dx, dy and dz are presumably motion in each direction, here representing no motion. So this output represents pressing and releasing button 1.
So what happens when the mouse vomits? Omitting the individual character reports:
Clearly there's a minor explosion here. There's motion in all three directions (possibly correct for x and y), and no less than 9 buttons pressed. The “buttons” mask appears to come from the low order bits of ”flags”. Button 4 is missing from the mask, possibly because I have remapped it to 2. What it does tell me is that there is genuine input to moused that causes it. Where does it come from? It could be the mouse, the receiver or the FreeBSD USB subsystem. In any case, there seems to be scope for rejecting any event which claims that non-existent buttons have been pressed. To be considered.
Radeon driver: should I care?
|
Topic: technology, multimedia, opinion | Link here |
I still haven't been able to get teevee to run with the X radeon driver. I'm torn between ignoring the problem and trying to find out what's wrong. In favour of the former course is that I really want a display with HDMI output, and the motherboard doesn't offer that. So after some consideration I ordered an ASUS card with the GeForce 710 (a page that I could only find by second-guessing the URL scheme on their terminally broken web site) and passive cooling. That'll set me back $57, much less than the pain with the radeon driver.
But it didn't leave me any rest. Why is this happening? It seems to be happening specifically to me. Is it related to the way that the server was built? Removed most of X (it's not easy with metaports) and rebuilt from something like scratch. No luck: I still ran into the undefined reference problem:
Do I really want to climb into this code? Or should I just enter a problem report?
eBay: Yet another bungle
|
Topic: general, technology, opinion | Link here |
Mail from eBay today:
What's that? It's a German variant of the stupidity I experienced two months ago: “We have reason to believe that your eBay account has been used fraudulently without your permission”. If they really have reason, they're keeping it close to their chest. But once again they have reset my password—that must be the fifth time this year, and my new passwords are becoming increasingly obnoxious.
Why are they doing this? Hanlon's razor suggests that it's not just to annoy me, though they're succeeding in doing this. Clearly their heuristics are broken, and since they don't listen to complaints, they're not doing anything obvious about it. In this case I tried logging in to http://ebay.de/ with an old stored password. That's all the easier since they keep making me change it.
What a company!
Use NBN for high performance!
|
Topic: technology | Link here |
Somebody posted this anonymous image on IRC today:
|
Clearly it carries a political message: “up to 150,000 homes within a single week”. Presumably that refers to the number of installations and not the number of destinations you can ping in that time, so I assume it comes from the National Broadband Network. But looking at the details is interesting. First, it's clear that this is a display from http://www.speedtest.net/, a company with a not quite unblemished record (it refuses to believe that I'm in Australia, for example). But it shows somebody apparently in the outback connecting to Internode in Adelaide. And the ping time is 598 ms! That's beyond terrible.
It seems that this is an example for Sky Muster, the NBN satellite service. It's interesting. A good satellite service should give a better time, though not much: the speed of light makes a ping time of less than 480 ms impossible. So indirectly I think it's saying that they think that 580 ms is about as good as they can do. In any case, there's nothing in that image to show it, and it can easily give the impression that it applies to more conventional technologies. Bad planning on the part of NBN.
Thursday, 13 July 2017 | Dereel | |
Top of page | ||
previous day | ||
next day | ||
last day |
More garden stuff
|
Topic: gardening | Link here |
Mick the gardener along this morning. There's not much to do at the moment, but he planted some red-flowering carpobrotus (what species? I still can't find out) in the island in the middle of the driveway:
|
|
He also moved the tall Cannas away from the window, swapping some of them with the hedychium species (Hedychium gardnerianum?) at the east end of the bed.
|
|
We had enough cannas left over to give some to Mick, Petra Gietz (who happened to be there) and Chris Bahlo.
ATA fail!
|
Topic: technology, opinion | Link here |
Petra Gietz has just changed her [IR]SP from Telstra to Aussie Broadband, and she's very happy, maybe too happy in view of her limited experience after the changeover, which, however, was seamless. One issue: what about VoIP? While we were talking about, I looked at my ATA and discovered that the power light was indicating red—error!
What caused that? I couldn't communicate with the device. I had a vague recollection of a similar issue last year. That prompted me to buy a second ATA, though I was never quite sure whether the problem was with the ATA or the plug-in power supply.
OK, what about now? (Search for and) grab the other ATA. Red light. Grab another power supply. Red light. Dammit, why are thees things so unreliable? Spent some time online looking for new ATAs, preferably a different brand, but none are available in Ballarat, and the cheapest I could find was well over $100.
Discussing the alternatives on IRC, we considered the very likely probability that there were two failures. I didn't have a third ATA, but there was a possibility that I had a third power supply (5 V, 2 A). And how about that, I found one from Cisco, apparently for one of the Sipura SPA-3000 ATAs that both failed after a power failure over 10 years ago. Plugged that in, and—it worked! So did the other ATA.
Clearly a power adapter issue. But why? I had measured the voltage and found it to be good. A load issue? In any case, time to buy a supply of cheap adapters as spares. Off to eBay. What size plugs are they? They're coaxial and appear to be relatively common—external disk drives also use them—but what are they called? Measured one and found that the outside diameter was 5.45 mm, clearly matching the offers for adapters with 5.5 mm plugs. A further investigation showed that it's a IEC 60130-10 type A connector, and there are two different kinds: one with 2.1 mm internal diameter, and the other with 2.5 mm. How do you measure that? Do the manufacturers put this in their specs? Of course not! This page came up with an obvious method: find drill bits with 2 mm and 2.5 mm diameter. The latter won't fit in the 2.1 mm plugs.
So: tried that and found that the plugs have a 2.1 mm diameter. The replacement adapter that I tried first has a 2.5 mm diameter. I'm guessing that that's enough to supply minimal power, but the contact resistance is sufficient that it can't supply full power, thus the failure indication. We're back to one real failure, the first power adapter, and at least one perceived failure. A lot of work, but worth it.
Now why don't the manufacturers write this stuff on their devices? External disk drives use
one of the same kind of connector, but they want 12 V at a lesser current. Plug one of them
into an ATA and you'll probably blow it. But at
least Sipura
Cisco
Linksys put the power rating and polarity
near the contact, in a font that I can barely read without a magnifying glass:
|
|
Friday, 14 July 2017 | Dereel | Images for 14 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Still more pump problems
|
Topic: general, opinion | Link here |
We had 4.4 mm of rain overnight, though relatively little water had gathered in the sump round the front of the house. Turned on the pump... nothing! GRRR! I'm having nothing but trouble with that pump, though this is the first time that it didn't run at all when I turned it on.
Went back into the house, and to add to my mood, Yvonne told me that there had been a power failure. OK, at least that would explain the pump not working. But no: the microwave oven had no power, but the wall oven did. Circuit breaker.
No, it turned out to be the RCD. Turned it on again. It failed, repeatedly. So is the pump to blame? It's fairly likely, since the cable is partially immersed in water. Out and disconnected the plug. RCD still trips.
Which devices in the house are connected to the RCD? Not that many, in fact: most are on the UPS, which has its own RCD. Washing machine, dryer, fridges, deep freezes, kitchen appliances, toaster and (as only occurred to me later) dishwasher. How do I pull the plug on the fridges and freezers? The connectors are behind the devices. So I disconnected what I could and checked. Yes, the RCD no longer tripped. Connected everything again. RCD still didn't trip. And the pump worked. How I hate bugs that go into hiding!
Round this point Yvonne resurfaced, and I told her what had happened. “That's OK, then”. No, of course it wasn't. We still didn't know what the problem was.
Then I saw that I hadn't reconnected the electric kettle. Reconnected that. Trip! So at least we know what it was, and the kettle was neither essential nor expansive.
But why did it fail? How can it fail? It's a heating element immersed in water. Why should there even be an earth contact? But there is. What use is it? It's only connected to the element, something that is neither easy nor a good idea to touch when power is on.
To add to the fun, eureka powered down. It's not quite clear why: it's on a UPS, and the UPS didn't fail. I had put it on a separate UPS when the main one kept failing, but now it seems that I should rearrange things.
E-PM2: The end of time
|
Topic: photography, opinion | Link here |
Yvonne took some photos today with her Olympus E-PM2, and then ran into trouble processing the photos: after reading them in, the photo directory (20170714, after the day) didn't exist. Checking showed that the photos had been written to 20170101.
Why? It seems that the camera had forgotten its concept of time. I've seen that before, when changing batteries on my Olympus E-PM1, the previous model. Yvonne recently changed the battery. Is this a design problem that they didn't fix for the E-PM2? Tried removing the battery, and it kept its date, so maybe it's only a problem if the battery is left out for some time. To be observed.
Revisiting 1964
|
Topic: history | Link here |
This online diary, kept since 1 September 2000, is not the first diary I've kept. I kept one on paper from 1 January 1963 until 1 October 1970, and over the years I've been typing in parts of it onto this site. By the beginning of last year I had about 2 years' worth online, and then I decided to do the rest, mainly 3 days' diary entry per day. My method is:
At the end of the month, I then do some automated tasks: ensure that there is a space after a comma (my keyboard likes to swallow it) and link to names of people to whom I refer frequently.
By yesterday, I had everything between 1 January 1963 and 27 July 1964, and also from 1 April 1966 until 1 October 1970. Then I came across a stretch of over two months that I had already entered. In principle I didn't need to do anything. But there were some comments in there that assumed that the pages were in isolation, and I had noted a number of typos. So today I started proof-reading the two months' entries. How hard can it be?
By evening I had managed until about 24 August 1964, and it's clear that it'll need more attention, in particular with the photos. Yes, I could leave it until later, but I don't think there ever will be a “later”.
Australia Post: yesterday, no Monday
|
Topic: general, technology, opinion | Link here |
The display card that I ordered on
Wednesday was sent with a tracking information sent to my eBay email address, so yesterday I was informed that it
arrived in black
hole Sunshine West the same
evening at 20:19, and that it should be delivered on Thursday. This information continued
until late evening, long after any hope of delivery had expired.
This morning, as expected, it had found its way to Wendouree in only 32 hours, and was available for collection at “NAPOLEON CPA”. Estimated date of arrival? Monday, 17 July:
|
|
I never cease to be amazed.
Highest honors for the president
|
Topic: politics, opinion | Link here |
Yesterday Donald Trump arrived in Paris to attend the parade at the Fête nationale celebrations. This was, of course, reported in the news media world-wide.
What was apparently missed by the fake news were nuances that only a Frenchman could have thought out: in the morning, the president met with Angela Merkel, and a decision was made to build a new European fighter jet. Why? Clearly they needed a reliable supplier. And he met Donald Trump at the Invalides. Not at Orly airport.
It seems that the reporters didn't twig to these subtle details. Did Trump? He's not subtle, but he's very sensitive when his own person is involved.
Saturday, 15 July 2017 | Dereel | Images for 15 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
DxO: We support your lens
|
Topic: photography, technology, opinion | Link here |
Two months ago I discovered that DxO Optics “Pro” reports my Leica DG Vario-Elmarit 12-60 mm f/2.8-4 as a Zuiko Digital ED 12-60 mm f/2.8-4.0 SWD. Why? Same focal length range, same aperture range. It must be the same lens. Who cares that the EXIF lens type code is different? My ticket on the subject was closed without action (or, as they put it, “solved”).
My Olympus Zuiko Digital ED 8 mm f/3.5 fisheye lens is also not supported, as I noted most recently last month. Well, that was the case until today. Now I'm given the opportunity to download a correction module—for a Panasonic LUMIX G Fisheye 8 mm f/3.5 (or, as they prefer to misspell it, “8mm / F3.5”, or even H-F008E). We don't need no steenking EXIF! And, of course, once again the two lenses have nothing in common except focal length and aperture.
More panorama issues
|
Topic: photography, opinion | Link here |
I attributed last week's problems stitching panoramas to the fact that the contrast was so low. But this week it happened again, despite the good contrast. And once again it had “autodetected” the focal length of the 8 mm lens to be 10.487 mm. What happens if I change that to 7.7 mm?
Success! It was the incorrect focal length that did it. But how did it get set that way, and why doesn't it get reset when I stitch an image? I'm pretty sure that this has something to do with lensfun, but what? To be investigated, some time.
teevee: NFS mount problems
|
Topic: technology | Link here |
Apart from my X problems, I've been having issues where teevee doesn't mount NFS file systems on startup. It's some kind of race condition: mount can't look up the remote system name that early on in the boot. I've worked around the issue by putting eureka into /etc/hosts, but I don't like the “solution”.
X: 25 years of pain
|
Topic: technology, history, opinion | Link here |
Chris Bahlo brought me my new GeForce 710-based display card today, so finally I can make progress with the new teevee machine.
Put it in and booted. No display. Neither from the onboard chipset, nor from the new card.
Took it out again and looked at
the BIOSUEFI settings. Set for
display board first, on PCI.
There were also settings for PEG and PEG1. Clearly this board
is PCIe (“PEG”), but which slot is which?
And why should it make a difference? The probe output showed that the board was recognized:
In the end, disabled the onboard display altogether. This is dangerous stuff: how do you get a display if it's disabled, and the other card doesn't want to work? And indeed it didn't want to work. Spent some time fiddling around, and in the end it worked. I decided that it was a combination of things ganging up on me, including the monitor not wanting to autoscan the inputs.
After some time and much fiddling, found a method that worked: the new card connected via HDMI and displaying correctly. Next, run X -configure. Not quite the results I had expected:
Google gets quite confused by terms like “X -configure”, but it seems that the problem is known. There's even a “solution” in the Arch Linux forums: don't use X -configure, you don't need it. Wrong! You do need a configuration file if you're doing anything non-trivial, like looking for display resolutions beyond 1280x1024 (aren't they all)? And of course, if it's broke, fix it. In this case, the default configuration loads the nv driver, and not the nvidia driver.
Still, there are alternatives. tiwi has a functional configuration for an nVidia card, so try that. The problem there, of course, is that the model number is different, and so is probably the PCI ID. OK, there's a second configuration program, nvidia-xconfig. Try that.
Yes, it generated a new config file. Removed the monitor identifier. Reorganized two entries in the Monitor section, moved a couple from the Device section to the Screen section—and that was that. It didn't even change the board name.
Starting X brought up a 640x480 display.
That's something to compare, of course. But why do I have to? Looking back at yesterday's article on my 1964 diary, it's sobering to think that I've been using X for more than half the intervening time, 25 years of it with BSD. It caused me pain then, and it's still doing so.
Garden flowers in mid-winter
|
Topic: gardening | Link here |
Today marks the middle of the middle month in winter, time for the monthly garden flowers.
In many ways, things look much the same as they did last month. We still have roses and Cannas, though the Gladioli are finally done, and the others don't look as good as last month:
|
|
|
And the Hebes are also still with us:
|
|
Indoors, the Begonias continue too
|
So do the Hibiscus, and our Cymbidium has just produced the first flower of the new season:
|
|
And the Aloe in the garden is producing a flower spike:
|
An outside frond of the newly planted Strelitzia nicolai is going brown, but that might have happened anyway. And the Strelitziae reginae are developing new flower spikes:
|
|
More power issues
|
Topic: general | Link here |
Another couple of transient power issues today, at 11:12 and 12:43. In each case my office UPS screamed, but there was no other problem. Is this the UPS or a power surge?
Chicken cooking times revisited
|
Topic: food and drink, opinion | Link here |
Roast chicken again for dinner today. That takes about an hour per kilogram, as noted in my cooking times page. This one weighed 2.075 kg and took 118 minutes to reach 82° in the breast. But on reflection the breast was overcooked. I think I'll cook at 170° and go for 80°, or maybe even 78°, next time.
Sunday, 16 July 2017 | Dereel | Images for 16 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
X pain, continued
|
Topic: technology, opinion | Link here |
So where do I go from here with my X configuration issues? Do Linux versions have the same problem with this hardware? There's an easy way to find that out: the biggest issue is to identify a disk on which to install a Linux release. Chose Ubuntu 17.04, which took its time installing, notably a period of several minutes after I agreed to install third party drivers (thinking, of course, of the nvidia driver). But it installed, and came up out of the box, sort of:
|
Ubuntu is trying to be a better Microsoft—they even use the term “folder”—and it seems that they come up with a different window manager interface every release. Once I discovered that “super” is mapped to the broken windows key, it made sense, and just pressing that key brings up the key code display. That's nice, but why should I have to re-learn all the keystrokes stored in my cerebellum? Maybe I could install fvwm2, but then I'm in uncharted territories again. And the lack of even basic utilities gets on my nerves:
|
Still, I have X running on my hardware. Now I just need the xorg.conf. But there wasn't one. /etc/X11 contained all sorts of files, but not a configuration file. And /var/log/Xorg.0.log showed that it didn't use a configuration file. Maybe that is why it used the NOUVEAU driver and not the nvidia driver.
Where do I go from here? All my attempts to use Linux have rubbed me up the wrong way, and, like Microsoft, the installations tend to want me to do things their way. But it is really beginning to look as if X under FreeBSD is seriously broken. Or maybe it's just a POLA violation, and there's a completely different new way to configure it. Or again, maybe I have broken something in my ports installation.
OK, how about duplicating the Ubuntu approach? Download a FreeBSD 11.0 ISO and install that. Started that, but didn't finish.
Monday, 17 July 2017 | Dereel | Images for 17 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Which OS?
|
Topic: technology, opinion | Link here |
As planned, downloaded the lasted FreeBSD ISO image today. And then Peter Jeremy asked why I didn't download PC-BSD. OK, that makes sense: it's specifically tailored to a desktop, and it is based on FreeBSD, so I don't have any painful adaptation.
But where is it? http://pcbsd.org/download/ shows many links, all crossed out. You can still access them, but they're ancient—the latest one is based on FreeBSD 10.3 and was made in March 2016. Not a good sign.
Then somebody pointed out that PC-BSD is an old, worn-out magic word, and it's now called TrueOS. That was—barely—recognizable on the home page, but there's no mention of it on the download page. OK, go to https://www.trueos.org/downloads/. And that doesn't have release numbers at all, only dates. Tried downloading the most recent, http://download.trueos.org/master/amd64/TrueOS-2017-07-05-x64-DVD.iso. And it hung after about 75 MB. Repeatedly. I wasn't able to download it.
Somehow this is very sad. PC-BSD wasn't for me, but the idea of a supported desktop version of FreeBSD sounded good, and it was just what I was looking for. But it's beginning to look more like another distribution that, for whatever reason, just didn't catch on.
X configuration: solved?
|
Topic: technology, opinion | Link here |
Installing FreeBSD from DVD is pretty straightforward: I have a second root partition on teevee's disk, so I can just install to that and ignore the rest. But there's an even simpler way: copy the disk image from eureso, my one-step-ahead VM. That would also help identify whether my messing around with ports installations last week had an influence.
Did that and restarted X. No difference: still only 1024x768. But then I saw, in the log files:
OK, that's correct. I had loaded the driver for the old card that I have in tiwi. But why didn't the module mention this when it was loaded? What I got was:
=== root@eureso (/dev/pts/0) /etc/X11 9 -> kldload nvidia
=== root@eureso (/dev/pts/0) /etc/X11 10 -> echo $?
=== root@eureso (/dev/pts/0) /etc/X11 11 -> kldstat
At the very least it should have unloaded again. Can kldload return a non-0 status? According to the man page, yes:
OK, remove the driver and install the correct one. And there I read:
To use these drivers, make sure that you have loaded the NVidia kernel
module, by doing
# kldload nvidia (or nvidia-modeset, see below)
or adding
nvidia_load="YES" (or nvidia-modeset_load="YES", see below)
to /boot/loader.conf, or putting ``nvidia'' (or ``nvidia-modeset'', see
below again) on ``kld_list'' variable in /etc/rc.conf.
...
Starting with version 358.09, some important functionality of the driver
has been broken out into a separate kernel module, nvidia-modeset.ko.
Users that experience hangs when starting X11 server, or observe
(II) NVIDIA(0): Validated MetaModes:
(II) NVIDIA(0): "NULL"
in /var/log/Xorg.0.log should replace ``nvidia'' with ``nvidia-modeset''
in their /boot/loader.conf or /etc/rc.conf configuration files, depending
on how they prefer to load NVidia driver kernel module.
See /usr/local/share/doc/NVIDIA_GLX-1.0/README for more information.
The whole of this message is at /usr/ports/x11/nvidia-driver/files/pkg-message.in (why? It used to be pkg-message in the parent directory.
This last is exactly what I found on Saturday. OK, try loading nvidia-modeset. And how about that, I got a 1920x1080 display! Success! Or, as we used to say, “It compiles! Ship it!”.
Did some testing, and then switched back to /dev/vty0, notionally with Ctrl-Alt-F1 (in fact, on this Sun Type 7 keyboard it's mapped to Shift-AltF1, where AltF1 is the F1 key above the keyboard). And I got:
|
I've seen that before, but not for a long time. The X server (or component, probably driver) doesn't restore the chip set configuration correctly. Whichever way you look at that, it's a bug, and potentially a serious one. I couldn't find a way to fix it. Does this happen on Ubuntu? Definitely not: they have added a server option to ensure that you can't switch back. Could it be that this restrictive attitude has resulted in code atrophy?
In any case, finally I have a working X configuration. It has only taken me 16 days, helped by 25 years of experience.
What's wrong with this picture?
Somehow I find this incredibly depressing. PC-BSD is dead. TrueOS appears to be dying. X appears to be dying. And I'm not too sure about the health of FreeBSD. How many people run X on it any more? Do they all have the same pain that I do?
DBDRV responds
|
Topic: Stones Road house | Link here |
Call from Joshua Brooksbank relating to my dispute with Domestic Buildings Dispute Resolution Victoria last month. Had a long, sensible talk, and he sent me a document to fill out—without asking, in Microsoft .docx format! Tried that with OpenOffice, and made a mess of it. OK, “see attachment”. Hopefully that won't make his life too much more difficult.
Next anniversary: 20 years
|
Topic: history | Link here |
In today's mail:
How time flies! And it's 60 years since we first returned to Australia.
Sound on teevee
|
Topic: technology, opinion | Link here |
Now I have X up and limping on teevee, the next step is to get sound working. First problem: which driver? Tried kldload snd_driver, which loaded something like 20 drivers, none of which probed anything. OK, unload, try loading each in sequence. Nothing. No matter what I did, I didn't get a probe result.
What is the sound hardware, anyway?
=== grog@teevee (/dev/pts/2) ~ 4 -> pciconf -lv
That second one must be on the display card, presumably to drive the HDMI output. But it's the first one on the motherboard that I want to connect to. Spent some time fiddling around, but nothing I could do produced any results.
Then I discovered that I had plugged the output plug into the wrong connector (why can't they write the purpose on these candy-coloured sockets? I know it should be the green one, but in dim light it's difficult to distinguish). Put it in the right one, and my current configuration worked. Unloaded it. Still worked. Rebooted without sound driver. Still worked. So it seems that a suitable (dare I hope universal?) sound driver is in the GENERIC kernel. That's a reverse POLA, I suppose.
Now all I have to do is get remote control working for mplayer, and I'll almost be ready to switch over.
Dishwasher run times
|
Topic: general | Link here |
Our Bosch dishwasher, called SI6P6B in the part of the Tower of Babel that creates the labels, uses rinse aid at a phenomenal rate. It's almost certainly adjustable, but I have mislaid the instruction manual, and I can't work out how to do so. And Bosch doesn't know its own model number either. Maybe the web site is in a different part of the tower.
Still, do we need rinse aid? It's designed to help hard water drain off glass. But our water is rainwater, with no hardness at all. So for the while we're trying without rinse aid.
The surprising first result: the “fast” wash cycle has increased in time from 29 minutes to 33 minutes. Why?
Strelitzia damage
|
Topic: gardening | Link here |
Found the dogs (not just Nikolai) inspecting the larger of the Strelitzia nicolai that we have transplanted. One of the fronds has broken off:
|
|
Did the dogs do that? I don't think so. It's been very windy, and that's a much more likely cause. And is it the newest frond? I can't decide. The left-hand one is clearly past its prime, but that happens anyway. We'll see if it survives or not.
Walking in the forest
|
Topic: gardening, animals | Link here |
Went walking with the dogs in the forest today to the “Big Linden tree” (really a conifer). Here photos from last month.
|
|
We couldn't get much further: a big tree with an interestingly twisted branch had fallen across the way:
|
On the way back, I saw a number of interesting leaves:
|
This is a designated place for finding orchids, and I have a recollection that I've seen orchids with leaves like this. Now I just need to remember where.
Dog bath fun
|
Topic: general, animals, opinion | Link here |
When we had the dog bath pump repaired in April, they recommended running the pump once a month to ensure that it didn't seize up. No worries, says Yvonne, we'll bathe a dog at least once a month.
The last time was on 18 April, 3 months ago, so it's high time to run it again. Put some water in the bath and turned on the pump. It ran, but it didn't pump: airlock.
Grrr! You'd think that a pump that is completely below the water level would self-prime. It took me 5 minutes to manually prime through the outlet hose. I can avoid the problem by leaving the water in the bath, as I did, but will that cause other problems?
Tuesday, 18 July 2017 | Dereel | Images for 18 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
teevee: done?
|
Topic: technology, multimedia, opinion | Link here |
It seems that teevee is now as good as ready to run. The only issue was with the remote control. Reading my diary entry for 23 November 2006 was instructive: it seems that in lircrc, lirc assigns every button to a specific program. To avoid search issues, I had put a copy in ~/.lircrc, and it contained entries like:
It's understandable that's not going to talk to mplayer. But the entry in /usr/local/etc/lircrc referred to mplayer, so all I had to do was to remove ~/.lircrc and everything worked. Something simple, for once.
Still, quite some discussion on IRC. Do you need a configuration file or not? How do you specify which driver to use? Ubuntu configured out of the box without a config file, though I wasn't given the choice of driver. I had ticked “download third-party drivers”, but it didn't install the nvidia driver, and I don't know how to do it. Like everything here, it seems that documentation is wanting. Even X.org doesn't have any clear installation instructions, though once the X documentation was considered exemplary.
Peter Jeremy tells me that X -configure “works for him”, and thought that maybe the crash was due to the presence of the nvidia kernel module. OK, we can check that, and in the process consider whether I can use the on-board graphics to run the vtys, and use the nVidia card for X. That would have the advantage that I could display both vty and X at the same time on two different monitors. So I set the primary display to the on-board graphics in the UEFI setup and tried to start X:
The hardware was probed, so it's not quite clear why there should have been a problem, but I've had enough of this mess already.
While I had the on-board graphics active, however, I revisited the issue of running X -configure. As I suspected, it did generate an xorg.config.new before crashing. So I tried running X using that configuration:
So the crash was not the result of creating the config file, but of trying to use it. I suppose I should look at it more carefully, but again, this is enough pain already.
Other mods for new teevee
|
Topic: technology, multimedia | Link here |
Also addressed some other issues needed to get teevee to run in the TV environment. I'm starting a HOWTO page describing the modifications necessary. Today it was mainly how to ensure that X gets started automatically on boot:
Log in automatically as grog. Modify /etc/ttys to contain:
/etc/gettytab gets:
Ensure that ~/.bashrc contains:
I still have an issue with late mounting NFS file systems. With the late option in /etc/fstab, it now works, but it takes 60 seconds to do so. It works on tiwi, but I recall having problems there, too, so I should check what I did there.
Wednesday, 19 July 2017 | Dereel | Images for 19 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
teevee: going live
|
Topic: technology, multimedia, opinion | Link here |
So everything seems to be OK with teevee. Time to take it out of the office and connect it to the TV in the lounge room.
I need to keep tiwi running along with teevee until I'm sure that everything is OK. That means new cables and stuff, of course. Where have all the power cables gone? Found some new ones, and even an additional power strip. That was the easy part. Next, Ethernet. Found my old 5 port Ethernet switch, but where did the power supply go? I was using it only a few months ago, until I gave up on TV. But I couldn't find it.
Never mind, I had a second one, an 8 port switch, and a power supply to go with it. But it never got out of its self test. Power supply or switch? Should I care?
Three more switches, all only 100 MB/s. Any port (out of 8) in a storm. More power supply problems! They all have strange ratings, like 5 V, 2.4 A (which I had, but only 100-120 V input) and 7.5 V (which I had, but not with the correct connector). Finally I looked at the last one: a 24 port switch with (thank God) a normal IEC 63020 C13 power connector. So I plugged that in, and it recognized tiwi, but not the wall plug. It seemed that the cable I had put in there was defective. Crossover? I thought that modern switches could handle that. So a new cable, and that worked.
And then there were audio and HDMI cables, all placed so that it's almost impossible to plug them in. But finally I was done, and I turned it on, connected to the TV via HDMI connector 2.
It worked, but the image went beyond the edges of the screen. Xorg.0.log suggested that it wasn't getting any EDID information, and I had some recollection of that. Plugged it in to HDMI 1 (more cursing the location of the sockets), and it worked. But not well. This CPU has roughly twice the grunt of the old one, but the display was jerky, and top showed that it was maxed out.
Comparing the Xorg.0.log files from tiwi and teevee was interesting: the latter contained repeating messages:
Why? Searching the web suggested that this only occurs in conjunction with problems that stop things working altogether. But clearly the issue relates to the display, since I didn't have them with the normal monitor.
So off to the nvidia site to look for the driver documentation. I recall seeing it before, and finding it helpful, but I couldn't find a link on the web site, and I couldn't find anything in this diary either. Finally, after a roundabout way that I forgot to record, I did find it: Configuring X for the NVIDIA Driver and nvidia X configuration options. And there I found:
"NoMaxPClkCheck": each mode has a pixel clock; this pixel clock is validated against the maximum pixel clock of the hardware (for a DFP, this is the maximum pixel clock of the TMDS encoder, for a CRT, this is the maximum pixel clock of the DAC). This argument disables the maximum pixel clock checking stage of the mode validation pipeline.
So I added that to the config file, which, in addition to things I changed long ago, now contains:
Or does it? When I looked the next morning, the last line was commented out. But the messages no longer appeared. It looks as if this is still not over and done with, especially since the performance still seems less than stellar.
Apart from that, there are other minor things I need to look at. Why is the firefox display size so different? The xterm display colours don't match. And a lot of other minor stuff that doesn't stop me using it. But (here we go again) after 25 years, this stuff a real pain.
First camellia of spring
|
Topic: gardening | Link here |
Our Camellia japonica is starting to flower:
|
Only one small flower so far, but there are plenty of buds, and the leaves are looking better. Looking back, we've had it for nearly 7 years, and the only time it flowered anywhere near this early was 5 years ago. I'm hoping that things will be better now that it's planted in the ground.
Panorama problems solved?
|
Topic: photography | Link here |
It seems that I'm not the only person to suffer from the changed focal length syndrome that caused me pain earlier this month. Now T.Modes (as he styles himself) has produced this document, which I must read.
Thursday, 20 July 2017 | Dereel | Images for 20 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Strange egg yolks
|
Topic: food and drink | Link here |
Why are egg yolks so consistent? The genetic code, of course. But from time to time there are differences, like this one, with eggs from the same box:
|
Hugin alignment issue
|
Topic: photography, technology | Link here |
As planned yesterday, tried the instructions to solve the problem that I had earlier this month aligning panoramas. The issue was supposedly related to an incorrectly calculated focal length, and the instructions showed how to reset it.
It didn't work. It recalculated the incorrect focal length, and the panorama still was a complete mess. Another thing to investigate more deeply.
Strelitzia growth
|
Topic: gardening | Link here |
It seems I paid too much attention to the Strelitzia nicolai with the broken front. The other one that we transplanted is developing a new shoot:
|
|
That's interesting because of its almost complete lack of roots.
MyNetFone escalation
|
Topic: technology, opinion | Link here |
For reasons that I don't understand, CJ Ellis has genuine issues with his VoIP connection with MyNetFone. I've been using them for years, and have had relatively few problems. But a couple of years ago CJ had a strange problem: incoming calls were diverted immediately to voice mail, and MyNetFone support claimed—incorrectly—that the ATA wasn't registered. Outgoing calls worked normally. It took forever for them to fix it.
Now CJ has a problem: incoming calls are diverted immediately to voice mail. Outgoing calls worked normally, but MyNetFone support claims—incorrectly—that the ATA wasn't registered. There's clearly something really strange in their configuration. Got CJ to authorize me, and today I called up, in the process getting the this message while waiting for an answer:
While you are waiting, please consider this: Please stay on the line for the next available operator.
Finally I spoke to Manish, who couldn't add anything to the matter: once again, their system claimed that he wasn't registered. Asked him to escalate it, but he couldn't. Asked to speak to his supervisor, but he couldn't connect me, and I could expect a call back in 24 to 48 hours! This sounds like Telstra all over again.
Left him with the message that they should hurry, or they'd find that the service had been cancelled. Only then did I confirm with CJ that he was prepared to pay a little more for the corresponding Aussie Broadband service. For about $25 a year it should be worth it for his peace of mind.
ALDI pasta machine
|
Topic: food and drink, opinion | Link here |
Yesterday Yvonne bought a pasta machine that was on special at ALDI:
|
Do we need it? I don't know, but it seemed worth trying, though certainly not to save money: you can buy a lot of pasta for $129. Still, we can always bring it back...
The thing didn't look bad, and it comes with attachments to make all sorts of different pasta:
|
But of course the instructions were as confusing as ever:
|
Of course it came with its own volumetric flour measure (“cup”), but how much fits in there? And what's the issue with water, vegetable juice, stock and egg/water mix? This device is suited to wheat only (why?), and what kind of wheat noodles do you make without eggs? And what is the proportion of eggs and water?
So the first thing is to establish how much flour fits into the flour measure. That's a variable, of course:
|
|
|
|
So what's correct? Add the weights to the egg-water mix, and you end up with between 218 g and 257 g of pasta, and not the 160 g that the table claims. What's wrong here? On a different page I read (clearly IMPORTANT):
NOTE; DO NOT EXCEED THE RECOMMENDED MAXIMUM FLOUR CAPACITY OF 640G.
Elsewhere again I read:
Note: Do not exceed 4 cups of flour or 640g.
That looks like a hard limit, not a recommendation. But 640 g is the maximum amount of pasta in the table above (4 “cups”). So reading between the lines, “Makes g of pasta” in the table translates to “weight of flour”.
OK, we can do that:
|
|
Next, what proportion of egg and water? The noodles I know have no water at all, but maybe the machine needs it. One thing's clear: we don't want to add fractional eggs. So what does an egg weigh?
|
|
OK, 51 g egg, 29 g water, a total of 70 g. Put in the machine according to the instructions, turn on: “Appliance will begin to knead the dough for 3 minutes - this depends on capacity being mixed”. In fact, for one cup it starts with a display showing 11 minutes, and kneads for 4. I can't imagine the time being shorter for larger loads.
But despite all attempts, the quantities were completely wrong. At the end of the 4 minutes, I just had crumbs:
|
OK, another egg:
|
But that was too much:
|
Not much I could do there, so I pulled the mess out and kneaded another 20 g in by hand. Then back into the machine, and let it run. At the end of the 4 minutes, not much had changed:
|
Then it started trying to extrude the noodles. Complete and utter failure. At the end we had:
|
Nothing at all came out! On the face of it, that's not surprising. The chamber is far too large for a single “cup” of flour, and there's nothing in the mechanism to propel it to the output conveyor:
|
So how can this thing work at all? My best guess is that if you fill it with at least 480 g of flour and other ingredients to suit, some will end up in the conveyor, and the elasticity of the (wheat) dough will pull the rest through. But I'm not prepared to waste anything else on the device. It's a particularly bad failure. Back it goes, with a flea in its ear.
21-gun salute?
|
Topic: food and drink, opinion | Link here |
So what do we do with the perfectly good noodle dough? Put it through one of our manual noodle machines. But where are they? Searched the whole kitchen and pantry, but couldn't find them. It seems that we haven't unpacked them since moving here over 2 years ago.
What I did find, though, was “champagne” from Perrini in Meadows, vintage 2007. We had bought 2 cases of 12 bottles each, and drunk exactly 3. So we have 21 bottles left over.
How good is it? We'll find out. The thought occurred to me that if we popped all the corks at once, we could have something like a 21-gun salute.
Friday, 21 July 2017 | Dereel | Images for 21 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Shelducks
|
Topic: animals | Link here |
For some time there have been a large number of ducks across the road from the “schoolyard”, the property on the corner of Stones Road and Bliss Road where, on the 1889 map, there was apparently a school of sorts, here at the top left:
|
The ducks are in the blank area to the east of the schoolyard. And occasionally I've seen herons there too, so when we walk the dogs I normally take my Leica DG Vario-Elmar 100-400 mm telephoto lens with me. So far, though, I haven't had both the lens and the herons.
There were no herons today either, but even at a distance we saw some orange colouring we hadn't seen before. Australian shelducks, and they kept their distance:
|
|
Those crops correspond to full-frame focal lengths of between 1500 and 2000 mm, and the camera tells me they were about 72 m away. They could at least have taken to the air.
Roast cauliflower
|
Topic: food and drink | Link here |
Yvonne has borrowed a cookbook from Chris Bahlo, titled after the author (whose name I have forgotten) rather than the kind of cuisine. It's currently full of yellow Post-It stickers, and today we had the first one, roast cauliflower with pine nuts and other stuff:
|
Saturday, 22 July 2017 | Dereel | Images for 22 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
teevee network issues
|
Topic: multimedia, technology, opinion | Link here |
So it seems that most of the functions of teevee are working correctly. Today I moved the disk with the recordings from tiwi to teevee, so I no longer need to have it running.
That also means that I no longer need the switch in the lounge room, so I disconnected it and plugged teevee directly into the wall socket. No carrier! After some messing around, discovered that it would only work with the switch. Why? At one point I thought it might be a crossover cable issue, but it was the same cable, and just to be sure I tried a couple of different cables.
Could it be that the switch at the other end (in the pantry) remembers the polarity of the connection? I left it turned off for about 20 minutes with no change. Could it simply that the switch is defective? I've had one of them fail on me.
Discussing later on IRC, with no new insights. If it's the switch, does it affect other ports as well? Went into the dining room with euroa, my Microsoft laptop and thus not the best thing to test with, and plugged it into the wall socket. It connected immediately. OK, what about the lounge room? Same thing. Plugged teevee in in its place, using the cable I had been using previously, and it, too, worked.
So what went wrong here? I still don't understand. We can rule out cable problems, because I tried several, and the one that was in place when it ultimately worked had been in place before when it didn't work. Contact issues? Possible, but unlikely. I had tried wobbling the connectors to be sure. I still don't know, but hopefully it won't happen again.
In the meantime I should start looking for a replacement switch, maybe an 16 port one for the office. In the process I came across this gem in the specs for some switch:
Silently delivers up to 2000 Mbps on each port
Arguably that's wrong, of course: “deliver” suggests a direction.
teevee: current status
|
Topic: multimedia, technology, opinion | Link here |
So we have teevee substantially working as intended. But there are still some rough edges that I need to attend to, and it's worth noting them. Some are trivial, others less so.
And that's all I can think of at the moment. On the positive side, I can now transfer files between eureka and teevee at effectively wire speed. That didn't use to happen. Somebody must have changed the default TCP parameters.
The web meets Dereel
|
Topic: general, technology, opinion | Link here |
A while back, while looking for builders, I signed up with an online business search site that promised me three quotes within a ridiculously short space of time. The results were meagre at best: no online reply, though one person who contacted me may have come from them.
But they keep spamming me with more offers for things I don't need:
From: hipages <newsletter@newsletter.hipages.com.au>
To: Greg <groggyhimself@lemis.com>
Subject: Here's Dereel's Gutter specialist
X-Mailer: Create Send
Message-ID: <cm.1231261328404.ktydktl.bhljryhki.t@cmail20.com>
Reply-To: getquotes@hipages.com.au
Guttering and downpipes need to work correctly to protect your home in winter. If you haven't done so already, clean your guttering.
AAA Roofing & Restorations??? has 90 Recommendations and has been a member since 16 August 2013
OK, I'll bite. To the best of my knowledge Dereel (population 666, clearly a sinister implication) has no companies at all. So where are AAA? Ballarat? No, they're in Carrum Downs, 186 km away on the other side of Melbourne. Clearly Hipages' spam algorithms need considerable tuning.
Another cup size
|
Topic: food and drink, opinion | Link here |
I've been maintaining a page about weights and measures for nearly 13 years now. In particular, cups are a useless unit of measure: they can range in size between 80 ml and 284 ml, though the smallest used in cooking is 180 ml. But what about the measuring cup that came with the ALDI “pasta machine”?, clearly designed for left-handed people. It holds about 265 ml:
|
|
That's only marginally interesting, of course, since it's intended to be used for measuring flour, where it's wildly inaccurate:
|
|
|
|
Champagne: worth the price
|
Topic: food and drink, opinion | Link here |
So today we tried the champagne that I found on Thursday. We knew that we were in trouble when the cork hardly made a pop; it was only about half the diameter it should have been, barely wider than the neck of the bottle. And it tasted of cork. Yvonne and Chris sum it up:
|
|
|
|
Sunday, 23 July 2017 | Dereel | Images for 23 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Hugin stitch problems: solved
|
Topic: photography, technology | Link here |
For the past couple of weeks I've been puzzling about why my first panoramas two weeks ago didn't stitch correctly. Today I finally found out why. cpfind was finding control points on the end of the panorama bracket:
|
|
That makes sense, and it's something I had been concerned about some time ago. The real question is why it didn't cause a problem before. My best bet is that the contrast was better in other images, so the control point detectors found better results (or, as in this case, none of any use). Time to rearrange the bracket.
Scanning old negatives
|
Topic: photography, technology, history | Link here |
As I go through my 1964 diary, I'm re-scanning the photos I took at the time. I had already scanned them with the ill-fated Canon 9900F scanner, but I can do better now.
Or can I? After letting the Epson “Perfection” 4990 Photo spend hours scanning the negatives, I ended up with:
|
What the hell went wrong there? One of the scanner parameters? But which one? The settings in the scanner program are:
|
Which are useful? I've established that “Backlight Correction” improves the shadows, and for colour photos “Color Resotration” really does help a lot. And the others? Grain is always a problem in old films (here Kodak Plus-X Pan), so I do that too. And the dust is everywhere, so how about “Dust Removal”? Ah, that's an old, worn-out magic word (so why is it there?). The new one is “DIGITAL ICE Technology”.
Is one of these causing my problems? Went through most combinations, and found that colour restore didn't do much, though it did slightly darken the image:
|
|
In the images above, as well as many below, run the cursor over an image to compare it with its neighbour. As expected, backlight correction improved things:
|
|
And then “Digital ICE”:
|
So that was the culprit! What about dust removal?
|
|
Some dust removed, but not much. OK, what about removing the “colour restore”? The difference is subtle.
|
|
But there are other issues. Not all of the dust appears to be dust, and there are artefacts that appear at random, here visible on the shutters in front of Ananda's face. They're present in both images, but in different places:
|
|
|
|
That looks like an issue with the scanner or the software, possibly related to the dust removal. Given the quality of the originals, I can live with that.
DxO fisheye
|
Topic: photography, technology, opinion | Link here |
My Olympus Zuiko Digital ED 8 mm f/3.5 fisheye lens suffers from considerable chromatic aberration, and DxO Optics “Pro” doesn't have a profile for it. As a result, the best correction I can get is this (from the image shown below):
|
But thanks to its simplistic approach to recognizing lenses (“We don't need no steenking EXIF”), DxO thinks that it's a Panasonic Lumix G fisheye, because it has the same focal length and maximum aperture. And it keeps nagging me to download the module.
OK, who knows? Maybe it can help reduce the chromatic aberration. So I downloaded the module. The results were remarkable. No, it didn't reduce the chromatic aberration, quite the opposite, it would seem. Here an image from today's garden shots, firstly corrected manually, and then automatically with the DxO module:
|
|
That's not correcting optical faults. It's correcting a world view, converting a fisheye projection to (I think) rectilinear. For my purposes, that's useless. And why does it do it? DxO has another product, ViewPoint, that does that kind of correction. That's annoying, because it means that when I do have a fisheye with support, I won't be able to use it.
And the chromatic aberration? Significantly exaggerated:
|
|
That's mainly an indication that the lens properties differ significantly.
teevee progress
|
Topic: technology, multimedia | Link here |
More playing around with my teevee issues today. The colour difference on xterms proved to have nothing to do with bash: I had forgotten the .Xdefaults file, and that fixed that. It also made the window manager menus more legible, but it didn't change the font sizes in firefox. And to add to the list, mail isn't being delivered:
That's not overly serious, since I don't really need mail from this system, but I suppose I should fix it.
New network hardware
|
Topic: technology, opinion | Link here |
There's another issue with teevee: slow NFS mounts, is gradually beginning to make sense: after configuring the network interface, it takes nearly 10 seconds before it comes up. During this time the system attempts NFS mounts and fails. And, as I noted, this happened on tiwi before it, a machine with completely different hardware—in the box. It still connected to the same switch in the pantry, which has given me other cause for concern. So high time for a new switch.
I don't really understand switch pricing. I can buy a 5 port switch for as little as $8.73 including postage from Hong Kong, as long as I like white. A black version costs $1.18 more.
That makes for a price of $1.75 per (white) port. But come to 8 ports, and the minimum price goes up to $16.65, or $2.08 per port. And for 16 ports the minimum price is $83.00, or $5.19 per port. Does the architecture of a port mean that the price rises proportionately to the square of the number of ports? Or is it related to the fact that most 5 port switches are for a domestic environment, while 16 port switches are exclusively “professional”? No idea.
In any case, it seems that switches are particularly susceptible to failure, so it makes sense to buy one from a reputable manufacturer. Found a used Netgear JGS516 going for $49, including postage. That seems a reasonable choice: they have a lifetime warranty (my life, not the life of the switch), and though it's not clear that it applies to me, it is a certain indication of their confidence in their product. It also has a built-in power supply, which I'm beginning to consider an advantage.
Typesetting units
|
Topic: technology, general, opinion | Link here |
Andy Farkas came up on IRC today with an interesting quote from Wikipedia:
Later Apollo missions used Nikon F 35mm cameras with a special f1.2 aperture 55m lens in order to take photographs in low light conditions...
He thought that could do with improvement. So did I: f/1.2, not f1.2. But he was really looking at the focal length, and 55 m really did seem rather long. Clearly it should be 55 mm.
Or should it? Andy thought “55mm”. And that's a question that's been in the back of my head for a while. I (now) always put a space between numbers and units, but it is correct? I dragged out my style guides, and found:
A word space is used between the numeral and abbreviation
That's no longer the correct paragraph in the current editions, but since they want to be paid to show the content, I can't be bothered.
11.23: Unit names and symbols should be separated from any associated numeral by a thin space.
In all cases, there are exceptions for degrees, minutes and seconds.
The interesting thing, though, is that most people didn't like it. Despite clear and consistent rules, most people don't want to put a space between the number and the unit. Why not?
Monday, 24 July 2017 | Dereel | Images for 24 July 2017 |
Top of page | ||
previous day | ||
next day | ||
last day |
Fake saffron
|
Topic: food and drink, opinion | Link here |
A month ago I bought some very cheap “saffron” on eBay. At the price ($0.28 per gram, comparing to $5 odd per gram for normal quality saffron) it was clear that the quality would not be spectacular, but I was curious.
It arrived today:
|
|
|