Δαφνη Ευτυχια Αρθουρος

Okay, it should quickly become obvious that I'm not camping in Enchanted Ground[*] ...

The latest American Science & Surplus catalog arrived in todays snailmail. "Oh, right," I thought, "they sell photovoltaic solar cells! I've been meaning to look up whether it'd be cost effective to get some of those for Pennsic."

Eep! On p.52, a panel rated at 12V/1A that they said they got 19V/750mA out of unloaded ... on sale for $144.50. %wince% Okay, not in the cards for this year, and probably not any of the next several years for me, unless I can find a way to turn it into a profit-generating thang as well as an electricity-generating thang. So looking up whether their shipping could get it to me before I leave or not became moot.

OTOH, they've got a bunch of rechargeable batteries for what looks like good prices so far (i.e., before checking the AH:price ratio of car batteries down the street at Advance Auto Parts): 6V/12AH Pb-acid $7.50, 15ct. "grab bag" assortment of NiCd $3.95, 10 AA NiCd packaged together as 12V/1AH (but easily cut apart to use as individual cells, so maybe I should get these for me DigiCam (but I'd better check the camera's manual to see whether it warns against using NiCd cells)) $5.95, single AA 1.2V/2100mAH NiMH $2.25 (I think those are $13.95 per 4-pack at Rite Aid -- dunno about other local retailers).

The main thing I'm pondering how to run is rated at 5V/2A (I tried four NiCd D-cells and got less than two hours), so either the 6V/12AH Pb battery plus enough solar cells to both recharge the battery and run the device (not cost effective), or a much better idea of how often I'll actually want to be making active use of the device to get a better idea of charge/drain times (not enough info just yet), would be useful. I think the grand plan there will wait a bit longer to get into seriously trying to implement it.

So the secondary concern is wanting to extend the life of the battery+inverter I'll be leeching off of to recharge my PDA and cell phone (and the laptop I borrow, if I do borrow one) and to help me feel less guilty about leeching off of syntonic_comma's battery (I do wish I'd remembered to yank the battery from my car before it was towed). Of course, being able to skip the inverter stage would help ... the catalog list a "7-cell Li-ion pack" with an assortment of connectifiers and output switchable "between 3V/6A up to 8.4V/2A" (I hope they mean there arew a bunch of voltages available in between the two extremes) but it doesn't list an AH rating, which makes it harder to judge the cost-effectiveness ... and neither my phone nor my PDA has a nice, standard, barrel or phone connector for power. (I can charge either directly from 12VDC with a cigarette lighter socket -- easy enough to wire up if I get a chance to swing by a Radio Shack or similar store, or if I find the one I used to use as an extender when I drove a car with a lip on the ash tray that prevented the plug from the phone I was using then from fitting into the lighter.)

Third on the list is powering the digital camera. Not wanting to drain the laptops/phones/PDA master battery for that, especially with inverter losses (wait, just how efficient are those itty bitty consumer-grade 100W inverters these days, anyhow?) I was planning to bring a Big Pile O' Batteries and just run it on alkaline AA cells when the two sets of NiMH cells I have run out. But I wonder: if I can afford to buy a mess of NiMH or Li-ion AA cells in time (and run them all through my charger), will they hold their charge long enough in the camera bag to still do me any good in the second week, or should I stick with alkaline (or maybe NiCd) for Pennsic?

Golly, if only I were going to a party known to be frequented by electronics geeks, scientists, and engineers between now and War! ;-) This sounds like a problem to discuss over ice cream, n'est-ce pas? Gotta remember to take the catalog with me to Baitcon this weekend.

(Whoah. I just noticed a 10-pack of A76 cells -- which suck in film cameras, which prefer S76, but are great in laser pointers -- for $1.50. That's less than a single cell costs at the drug store.)

[*] This footnote grew into a whole entry of its own, about the many different ways of enjoying Pennsic and how I feel about bringing electronics, which I'll post a little later this afternoon.

Made it to pharmacy. Made it to nail salon after very annoying wait for wrong bus due to misinformation. Stopped at Appalachian Bluegrass for a nut blank because the slot for the bottom string on the mandolin I was practicing on while waiting for buses crumbled. (Blank is way bigger than I need; gonna have to cut it lengthwise. Anybody local have a band saw and not mind the smell of bone dust too badly?) Stopped at grocery on the way back. Am very tired.

Especially annoying numbers of rude, immature people out and about today. Funny thing, when I hear people laughing and turn around and see a cameraphone pointed at me (with such an exaggerated gesture that it seems my seeing them is half the point), and I bring an SLR to my eye pointed back at them, people freak out. Dayum, either it's acceptable behaviour or it's not, idjits; make up your minds! But suddenly I go from curiosity to something-to-flee once the lens is aimed at them.

Have lined up rides for tomorrow. Tried to avoid pushing myself so hard today that I'll be too beat to cope tomorrow, but with the amount of walking required because of where the bus lines do and don't go, and just being Out And Active fort so long, I'm pretty darned beat. Hope I didn't push toooo hard. Gotta manage to be functional Sunday as well!

Want ABC 2.0 <-> MusicXML converter. Found an ABC 1.6-> MusicXML tool, but a) it dropped three measures when I tested it, b) it barfs on V: tags, and c) the source seems to have been lost and the author is unreachable. Half a mind to start banging one out in C (and wondering whether I could to the MusicXML->ABC direction as a 'sed' script), but already have Too Many Projects in the queue so should try a little longer to find existing tool maintained by Somebody Not Me. Also want shareware or o-s OMR (aka 'music OCR'); closest I've found is 30-day free evaluation of $190 program. Will try that, but will try to time the start of the eval period to maximize usefulness. (Probably worth $190, if I had the spare dough for that sort of thing.

Have found annoying amounds of ABC FUD on chat boards while Googling for tools, as well as an awful lot of "since I don't need that, I can't imagine it's of use to anyone else on the planet either; they should just all use the same proprietary notation tool that I do" attitudes. Feh.

Doh! Should remember to ask about XML conversion on ABC mailing list!

Tentative lead on Volvo wagon. More news ifwhen more news exists.

Difference between now and the long-tiny-nested-paren paragraph before, is today I realize I'm this tired.

Perrine is much more tolerant than I thought she'd be, of the "cat-on-keyboard bad, so if you choose to occupy keyboard zone, it's keyboard-on-cat" rule I've instituted. Though her acceptance or disgusted departure depends somewhat on my typing speed.

I do wish she'd tell me where she hid the mouse I can smell rotting. *sigh*

Made it to pharmacy. Made it to nail salon after very annoying wait for wrong bus due to misinformation. Stopped at Appalachian Bluegrass for a nut blank because the slot for the bottom string on the mandolin I was practicing on while waiting for buses crumbled. (Blank is way bigger than I need; gonna have to cut it lengthwise. Anybody local have a band saw and not mind the smell of bone dust too badly?) Stopped at grocery on the way back. Am very tired.

Especially annoying numbers of rude, immature people out and about today. Funny thing, when I hear people laughing and turn around and see a cameraphone pointed at me (with such an exaggerated gesture that it seems my seeing them is half the point), and I bring an SLR to my eye pointed back at them, people freak out. Dayum, either it's acceptable behaviour or it's not, idjits; make up your minds! But suddenly I go from curiosity to something-to-flee once the lens is aimed at them.

Have lined up rides for tomorrow. Tried to avoid pushing myself so hard today that I'll be too beat to cope tomorrow, but with the amount of walking required because of where the bus lines do and don't go, and just being Out And Active fort so long, I'm pretty darned beat. Hope I didn't push toooo hard. Gotta manage to be functional Sunday as well!

Want ABC 2.0 <-> MusicXML converter. Found an ABC 1.6-> MusicXML tool, but a) it dropped three measures when I tested it, b) it barfs on V: tags, and c) the source seems to have been lost and the author is unreachable. Half a mind to start banging one out in C (and wondering whether I could to the MusicXML->ABC direction as a 'sed' script), but already have Too Many Projects in the queue so should try a little longer to find existing tool maintained by Somebody Not Me. Also want shareware or o-s OMR (aka 'music OCR'); closest I've found is 30-day free evaluation of $190 program. Will try that, but will try to time the start of the eval period to maximize usefulness. (Probably worth $190, if I had the spare dough for that sort of thing.

Have found annoying amounds of ABC FUD on chat boards while Googling for tools, as well as an awful lot of "since I don't need that, I can't imagine it's of use to anyone else on the planet either; they should just all use the same proprietary notation tool that I do" attitudes. Feh.

Doh! Should remember to ask about XML conversion on ABC mailing list!

Tentative lead on Volvo wagon. More news ifwhen more news exists.

Difference between now and the long-tiny-nested-paren paragraph before, is today I realize I'm this tired.

Perrine is much more tolerant than I thought she'd be, of the "cat-on-keyboard bad, so if you choose to occupy keyboard zone, it's keyboard-on-cat" rule I've instituted. Though her acceptance or disgusted departure depends somewhat on my typing speed.

I do wish she'd tell me where she hid the mouse I can smell rotting. *sigh*

Dear Lazyweb,

What's the proper lubrication for this type of AC motor (from a box fan)? I'm guessing that neither WD-40 nor olive oil, the two oils I have on hand, are correct.

Graphite powder seems iffy to me in an electric motor too, but if'n I knew fer sure, I wouldn't need to ask. Oh, I might have some 10W30 around here someplace, if that would work as a substitute for the correct stuff. Or is this one of those "once it starts squeaking and rattling and stalling it's time to throw it away" situations? It would run normally for a few seconds, squeak really loudly for a few minutes, then slow and stop. I cleaned a decade or two worth of accumulated dust and hair out of it but it's still not happy.

Also, when I switch it on and it locks itself up -- an electromagnetic situation, not friction -- does that mean it's not worth trying to lubricate it to eke a little more life out of it? It doesn't lock up every time, just once in a while.

It's pretty cool how neatly it comes apart and goes back together anyhow. I hadn't taken apart this type of motor before. Getting the fan blades off the shaft was the hardest part.

Dear Lazyweb,

What's the proper lubrication for this type of AC motor (from a box fan)? I'm guessing that neither WD-40 nor olive oil, the two oils I have on hand, are correct.

Graphite powder seems iffy to me in an electric motor too, but if'n I knew fer sure, I wouldn't need to ask. Oh, I might have some 10W30 around here someplace, if that would work as a substitute for the correct stuff. Or is this one of those "once it starts squeaking and rattling and stalling it's time to throw it away" situations? It would run normally for a few seconds, squeak really loudly for a few minutes, then slow and stop. I cleaned a decade or two worth of accumulated dust and hair out of it but it's still not happy.

Also, when I switch it on and it locks itself up -- an electromagnetic situation, not friction -- does that mean it's not worth trying to lubricate it to eke a little more life out of it? It doesn't lock up every time, just once in a while.

It's pretty cool how neatly it comes apart and goes back together anyhow. I hadn't taken apart this type of motor before. Getting the fan blades off the shaft was the hardest part.

Hmm. I shoulda seen that coming ...

When mirroring an LJ entry to GJ and IJ, if the entry contains an <lj ...> tag (e.g. <lj user="examplename"> => examplename), the IJ and GJ copies of the entry will interpret the same tag as linking to a user of the same name on those sites. And similarly, if I try to link to a GJ member, the LJ copy of the entry will try to point to an LJ user with that name. Oof.

Unless TPTB decide to extend the <lj user=""> tag to have syntax that can point to other LJ-codebase sites (or maybe other OpenID sites in general), and all three sites install that feature, I've got a problem. Maybe I should just forego the convenience of <lj user=""> and <lj comm=""> and use longhand HTML to produce the same effect instead (maybe stick a macro in my .exrc file to help) ... but I think that would break the "automatically fix all your old tags to point to someone's new name when they rename their journal" feature, wouldn't it?

Thoughts? Suggestions? Obvious workarounds I've overlooked?

I'm glad I noticed this before getting around to copying all of my old LJ entries over to IJ and GJ.

[Edit: a test of doing things longhand (though I might simplify this a bit from the copy/paste version here, in practive):

• dglenn
• dglenn
• dglenn

Hmm. IJ doesn't link from the icon, and links the username to the profile instead of the recent-entries page -- odd. Bug?]

Hmm. I shoulda seen that coming ...

When mirroring an LJ entry to GJ and IJ, if the entry contains an <lj ...> tag (e.g. <lj user="examplename"> => examplename), the IJ and GJ copies of the entry will interpret the same tag as linking to a user of the same name on those sites. And similarly, if I try to link to a GJ member, the LJ copy of the entry will try to point to an LJ user with that name. Oof.

Unless TPTB decide to extend the <lj user=""> tag to have syntax that can point to other LJ-codebase sites (or maybe other OpenID sites in general), and all three sites install that feature, I've got a problem. Maybe I should just forego the convenience of <lj user=""> and <lj comm=""> and use longhand HTML to produce the same effect instead (maybe stick a macro in my .exrc file to help) ... but I think that would break the "automatically fix all your old tags to point to someone's new name when they rename their journal" feature, wouldn't it?

Thoughts? Suggestions? Obvious workarounds I've overlooked?

I'm glad I noticed this before getting around to copying all of my old LJ entries over to IJ and GJ.

[Edit: a test of doing things longhand (though I might simplify this a bit from the copy/paste version here, in practive):

• dglenn
• dglenn
• dglenn

Hmm. IJ doesn't link from the icon, and links the username to the profile instead of the recent-entries page -- odd. Bug?]

"In fact when I first got my Apple II the first thing I did was turn it on and off, on and off, just because I had the power to do so, which I'd never had on a computer before." -- Andy Hertzfeld

Swiped from Bash.org:

<Jay> Did you hear about the Linux-car finishing last in the indy500?
<MrBeek> I did now ;-)
<MrBeek> Not surprised though... You know how impossible it is to find a decent driver for linux hardware?

[I personally have not had hardware driver problems in Linux, but I've heard enough old horror stories to be amused anyhow, -- DGA]

"Now, this pulls against the cardinal virtue of ease of use. But ease of use is wrong. Ease of use is the wrong way to look at the situation, because you've got the Necker cube flipped in the wrong direction. The user of social software is the group, not the individual.

"I think we've all been to meetings where everyone had a really good time, we're all talking to one another and telling jokes and laughing, and it was a great meeting, except we got nothing done. Everyone was amusing themselves so much that the group's goal was defeated by the individual interventions.

"The user of social software is the group, and ease of use should be for the group. If the ease of use is only calculated from the user's point of view, it will be difficult to defend the group from the 'group is its own worst enemy' style attacks from within."

-- Clay Shirky, "A Group Is Its Own Worst Enemy", 2003-04-24

Yak-shaving. *grrr* At least the dialogue box that told me to install that library should have warned me yak-shaving would be required, even if, "for patent reasons", it couldn't provide a way to sidestep the yak-shaving entirely.

I guess some folks will consider it a silver lining of sorts that I have gotten annoyed enough to think I should switch to an unencumbered codec on principle even though I did manage to get what I was looking for; at the moment I'm just sleepy and cranky and completely distracted from the project I started this whole excercise for in the first place.

(Oh yeah, lots of clues on the web, uh huh. But most of them only made any sense after I'd found The Rest Of The Story elsewhere on the web to provide the necessary context, a few hours into the mess.)

Okay, the cable is connected ... 'lsusb' shows that the computer can see the printer ... I click "Print Test Page" in 'Gnome-CUPS-Manager' for that printer ... I get a message saying the test page has been sent to the printer ... the status window for that printer in thr CUPS-manager shows my print job with a status of "Printing" ... and then it changes to "job stopped" and the "resume" option in the pop-up window is greyed out. Logged in as myself, this happens. Logged in a root, this still happens.

What ought-to-be-obvious (but isn't obvious from the 'man' page) detail am I most likely overlooking here? The printer is on-line as far as I can tell (it's powered up but doesn't have a separate "on-line" or "ready" LED; the LCD says "plain paper / normal / 100%" as expected, and the "error" LED is dark). At first I thought I had a bad cable, but 'lsusb' does report a device of the right brand plugged in.

The last time I managed a printer from *nix was back in the 'lpd' days, using RS232 ports and parallel ports, not CUPS nor USB.

"Hey, what's the name of that tune ... uh, the one that goes da Da dee deedle dum, da Dee dee Dada diddle Dum?"

There exist search engines for melodies; you enter, by various means, a bit of music, and the search engine tries to find a known tune that contains that phrase. This came up in conversation sometime in the last couple of weekends, and I don't remember whom I was talking to that seemed surprised and/or skeptical that such a thing exists; so, in the hope that it was someone who reads this, and also with the idea that there are some folks here who missed it the last time folks were saying "oh cool, lookit this" or had simply forgotten about it since then (or forgotten the location anyhow) ... I figured it was time to mention such things again. It's pretty cool, though I expect that ten years from now we'll have better tools and a more complete index and this will look clunky in comparison.

Musipedia has such a search engine, and there is another (or maybe another instance of the same one?) at Melodyhound.

Here's the thing: when I say "enter a bit of music", you have a bunch of ways to do so:

• clicking on a drawing of a piano keyboard (or here)
• typing in Parsons code (or here)
• singing, whistling, or humming into a microphone -- requires Java -- (or here)
• tapping out a rhythm on your keyboard (or here)

Note that Parsons Code only requires that you be able to give a series of "it goes higher / it goes lower / the note is repeated" indications; you don't need to actually know the intervals or a starting pitch or be able to notate the rhythm.

I think Melodyhound and Musipedia only search Musipedia's own music database, not every notation site on the web. I'm not sure whether there exists a whole-web (or even "registered with the big-name tune index sites" portion of the web) search engine optimized for finding tunes based on ABC notation JC's ABC Tune Match seems pretty effective at searching the web for ABC files based on titles -- it found The Armigerous Goose on my own site -- and it looks like it's supposed to also search on tune-fragments, but I haven't been able to make it do so), but I have noticed a bunch of ABC applications that'll search your own locally-stored ABC archive. If anyone knows of more general melody search tools than these, please link to them in a comment!

But as long as I'm posting links to musical tools on the web that I may have last mentioned long enough ago to warrant a re-mention, I'll also point out a couple of ABC resources on the web:

• If you haven't gotten around to installing any ABC programs and find yourself holding a chunk of ABC notation, ABC Convert-A-Matic is a web site that lets you paste an ABC tune into an input box, and generates GIF and PDF sheet music for you in standard notation, plus MIDI.
• Similarly, if you don't already have a program that transposes ABC files for you, Jens Wollschläger's ABC Transposer transposes an ABC tune up or down via a web interface.

(So, is it late enough in the day for those of you at work to start whistling into your computers yet?)

When re-starting a web browser, it is rather nice, in the event that something as un-nice as an unexpected termination (crash) has already occurred, to be presented the choice of starting fresh or picking up where you left off and reloading all the windows and tabs one had previously had open.

'Twould be even nicer to be given the option of reloading/restoring all previously open web pages except that one over there that I think was causing the problem that led to the crash in the first place. Admittedly, scrolling through as ginormous a list of open pages as I would usually face could get a little tedious, but watching the browser start up and hang in the same place o'er and o'er, or to have it crash each time I try to close the suspected page before it finishes loading, is somewhat more unpleasant.

(In this case, I'm using Firefox (well, "Iceweasel", since it's Debian), but this goes on my wishlist for Opera as well, doggone it.)

"There are only two kinds of programming languages: those people always bitch about and those nobody uses." -- Bjarne Stroustrup

[And a happy Beltane to my friends celebrating it!]

So here's the video I was trying to turn into an animated LJ icon:

The icon version is smaller than the full 100x100 pixels allowed, with fewer frames than I would have liked for a smooth animation, and I had to trim a lot of frames before and after the yawn so it loses the "oh, just a picture of a cat -- wait, it's moving" effect at the start and turns into one of those annoying constant-motion images. If I can figure out how to set different delay times for different frames, I'll try to put the long pause back in at the start that way. [Thanks, redaxe.]

So I guess the image embedded in this entry is subject to improvement and may no longer match this text if I'm successful. And, uh, if any of y'all with real Photoshop/ImageMagick/GIMP/etc. chops wanna show off how much better at this you are than I am, I'll happily make the original 3GP file or the separated JPEG frames available for you to play with -- all the more so if you think you can describe what you do well enough for me to learn from.

Last night on the way to anniemal's after rehearsal, I started thinking about camera-phone video shot from a bumpy, moving car. I was wondering how easily the bumps and jitters could be removed from the video by splitting it into individual frames, re-registering each frame so that a feature of the car -- say a mark on the dashboard -- was at the same location within each frame, cropping each frame to the area common to all the tranlated frames, and reassembling them into a video file.

So, since I was sitting in the passenger seat, I shot a bunch of video clips to play with later, either by trying to figure out how to automatically indentify the reference feature in each frame and feed the location to ImageMagick, or by Googling for existing software already designed to de-shake phone video.

This morning I considered my to-do list and asked, "What was I thinking? When am I going to get around to that?" Whoops.

But I still need to figure out how to split a 3GPP MPEG file into individual frames (or convert it directly to an abnimated GIF) for a completely separate, much smaller project. ImageMagick, at least the versions I've got handy, doesn't seem to like 3GPP video. And a shareware tool advertised as doing what I want insisted that I needed to install a new codec even though the Windows machine I was running it on could display the video already (implying that the right codec was already present on the machine) and even after I tried downloading another codec from the web site mentioned in the error message. (I'm trying to compile another tool now.)

[Update: the tool I said I was trying to compile failed to compile under Cygwin/WinXP but did compile -- and do exactly what I had hoped it would do -- under Mac OS X. I hope it also runs at my house in case I need it again in the future -- my guess is that its working correctly on the first try under Linux is more likely than not. Now to re-read the ImageMagick 'convert' man page...]

Dear Perrine,

I know it says "press F1 for help", but falling asleep with your chin on the F1 key and launching a couple hundred copies of Gnome Help Viewer is not, in fact, helpful. I'm still trying to close all these windows, and the machine has slowed to a crawl.

Also, oh mighty huntress, I see that rolls of toilet paper everywhere need fear thy mighty fangs and powerful claws. Those are some impressively deep bite marks. But look, sometimes I do have to sleep when you want to play, and I'd appreciate your picking something already designated as a cat toy to beat up on when you're bored.

-- the one who holds the grooming brush for you to rub your face against

"I don't get or expect much of a 'wow' from any operating system interface anymore [...]. But my UI research in '81 and '82 taught me the difference between the sizzle and the steak. The OS's job really is to act as an interface between applications and the hardware, and then to provide an 'environment' (for both programmers and users) that enables and defines how they go about their business. As a user, I am not fond of flashy graphics for their own sake. For me, the XP interface is usually more than graphical enough [...]" -- Chris Rutkowski, in "360 Degrees of Vista: The InformationWeek Windows Vista Roundtable", InformationWeek, 2007-03-05 (issue #1128), pp. 39-40)

In all fairness, Perrine probably only deserves the blame for the first chunk of this, not the whole thing, unless she's learned my root password and has been very naughty in the past few days ...

Okay, so earlier Perrine managed to do something screwy when she stepped on my keyboard. I started writing a plea for help, before it dawned on me that I need to be able to type '-' to post using the LJ client I normally use ... and then my ride to 3LF rehearsal showed up before I managed to post using the web interface so y'all got the very abbreviated version via SMS. There were more details in ( the version I typed in but didn't manage to post earlier )

Well, I've got the keyboard working normally again, but other bits of the Debian box are acting messed up...

When I got home, I tried unplugging the keyboard and plugging it back in, thinking the problem might be a keyboard firmware command rather than an OS or WM toggle. No change. ( other stuff I tried )

Then something I did finally effected a change!

The keyboards stopped working entirely. :-( Both of them.

Finally, I decided it was time to bite the bullet, close the gobs of windows open on three of the four virtual desktops, and shut down the several projects (and, of course, scattered web surfing) that I'd been in the middle of and hadn't wanted to lose my place in. Maybe logging out would take care of it?

Well no, I couldn't type anything in the username box on the X login screen, either. Nor, at that point, could I even shut the machine down cleanly. Finally, exasperated, I used the BRS shutdown option (actually it's a wee translucent pushbutton with a green LED behind it, but I don't think "WGGB" -- wee green glowing button -- has quite the same degree of recognition as "BRS") and counted on EXT3 to save me from an uncomfortable fsck later.

Well here's where this turns from a too-long-for-its-triviality "I cycled power and it's okay" story, into a full blown "WTF? + Why Me? + IsThisGoingToTurnIntoYakShaving?" type of adventure. You see, after I rebooted, I could not log in.

And so began the shaving of the yak. It turned out that I could log in as root, so I reset my password just in case it had gotten munged somehow. Nope, still couldn't log in as myself: enter username, get password prompt, enter password, get initial login screen again. Feh. So I tried a bogus password just to check, and waitaminute, that got me an error message instead of just #blink#ing back to the splash screen. So, something screwed up in one of those dot-files in $HOME controlling X sessions, that I'd never bothered to look at or learn anything about? Well logging in as root again and running "ls -a ~dglenn | more" did turn up an interesting-looking filename, .xsession-errors, with a relevant-seeming timestamp.

It was complaining that the X startup script in /etc/gdm/Xsession (oh! I guess that means I'm using Gnome then -- I'd wondered^[1]) couldn't access /dev/null. WTF?! "Can't access /dev/null?? That doesn't sound at all reasonable..." Sho' 'nuff, /dev/null was 660, which Looked Wrong to my eye. Very wrong. In case fatigue was clouding my judgement, I telnetted to a Linux box downstairs to compare: as I'd thought I'd remembered (and as really only makes any sense at all), /dev/null over there was 666. Chmod 'dat, try again. Different glitch, and no login screen; power cycle the puppy agin' and try once more ... and /dev/null is 660 again ...???? Log in as root, chmod the damned null device, log in as myself, and fail because there's no entropy collector. Ask 'man' about "random" and see that /dev/random and /dev/urandom are supposed to be 644 but they're 660 too. Chmod those, and /dev/hw_random as well on a hunch, and try <voice tone="whine" pitch-modifier="nasal" syllable-multiplier="3">again</voice> to log in as myself ...

*Whew!* A Gnome desktop. BUT...

... but when the very first thing I tried to do -- start a terminal window -- only half-worked, I feared I might wind up taking tweezers to the damned yak before the night ended. The window popped open, but so did an alert box telling me it couldn't start the child process for the window. I don't remember where I saw some error message refer to a pty, but (Wait, wait, can you guess what the Next Problem turned out to be? Sure you can; it fits the theme. Yeah, that's right, you got it ...) sure enough, when I looked at /dev/pty*, the permissions were fucked up there as well.

I wonder: are all these permissions going to be munged again the next time I reboot? Or just /dev/null? More importantly, what have I not found yet that's going to bite my ankles repeatedly in the next several days? [Later Edit: SSH is broken -- when I try to connect to the Suse box as myself, I get "ssh_askpass: exec(/usr/bin/ssh-askpass): No such file or directory", and if I try to do it from a root prompt it asks me for a password but no password I enter works. SSH was working perfectly fifteen hours ago!] And what screwed this up in the first place?

It's conceivable that some malicious cracker has broken into this computer -- security on my house LAN is not very strong, nowhere near as strong as I really know it really ought to be (and really will be, real soon now, when I get around to it) -- but seems pretty darned unlikely since it's behind a masquerading router. A worm or virus is much more likely considering that even though I am immune to malware that propogates via email, I have been surfing the web with Javascript enabled and could have picked up malware from a compromised web site.

But I suspect -- purely on a hunch, mind you, with no more evidence than a nagging distrust left over from the time it overwrote my kernel -- I suspect that some recent update via aptitude/apt-get/etc. mangled the permissions on a bunch of system device files that some upgraded package claimed dominion over. 'Cause yeah, I did run a couple of updates during the, uh, [pause to look it up] month + 1.5 days since the last time I logged off of this computer (or the last time the power went out or something ... I don't log off of the bedroom machines often).

Anyone with more current Linux sysadmin skillz and/or who has had something like this happen to them, care to provide additional clues? Evaluate my guesses as to the likelihood of the cause? Suggest where else I should look for related problems first (before trying to sift the entire file system)?

And are there any more ideas on what magic chord Perrine struck that triggered the odd keyboard behaviour^[2] that started this exercise in the first place, and how to undo it without the involvement of hirsute bovines if she manages to hit the same magic combination again?

All in all, I'd have to call the yak only half-shaved considering that I didn't wind up nested three topics deep in a Google search, or have to suddenly learn All The Details of X customization and management that I've been putting off reading up on for too long, but it was still a rather longer detour than I'd been prepared for when I decided to reboot the machine. (Keep saying it: "I love my cat. She didn't mean to do it. I love my cat. She didn't mean to do it. I love my cat ...")

And in the morning, I get to start figuring out where I was in various things that got closed when I rebooted. Whee.

[1] I knew I was most likely using either KDE or Gnome, but a) I hadn't gotten around to looking closely at both to see what the pros and cons of each are, b) I got this machine already configured, so all I did was change the root password and add an account for myself when I got it, not select all those install-time choices that would've tipped me off, and c) whenever I try to install something that says it's a Gnome app it just works, and whenever I try to install something that says it's a KDE app it just works, so that didn't seem like much of a clue either. The result was that yes, I've been using a modern Linux GUI for something like a year and a half without having gotten around to figuring out whether I was using Gnome or KDE until now. When things basically just work and the interface doesn't piss me off, details like that seem like things I should know but don't seem urgent[3], so the pertinent Round Tuit wasn't being stalked very aggressively.

[2] If I know how to reliably turn this mode on and off, it would actually be kind of useful. I've been saying for years that I wanted a SHIFT-LOCK button -- like on the manual typewriters I used to use -- on my computer keyboards somewhere in the same general area as the CAPS-LOCK key. Caps-lock is usually more useful than shift-lock, so I wouldn't want to get rid of it, but I find myself wanting shift-lock just often enough to fondly remember the dumb terminal I used a long time ago that had both. (Hey dmk, it was one of the ones we used at UD, wasn't it? Do you remember? I think it was one of the CRTs, but I'm not 100% certain that it wasn't the DecWriter.) Thing is, I want it just seldom enough to keep forgetting to look up whether it's easy to switch on and off on various modern operating systems that might have it as an accessibility feature. Now that I've been told two invocations I'll play with that under Windows, but neither seems to have any effect on this Linux box.

[3] Which brings up a topic for a separate entry of its own, when I get around to that.