[Most Recent Entries]
Below are 20 journal entries, after skipping by the 20 most recent ones recorded in
[ << Previous 20 -- Next 20 >> ]
[ << Previous 20 -- Next 20 >> ]
|Wednesday, May 6th, 2009|
This journal has been deleted and purged.
I'm going to miss you. Sometimes cheerful, sometimes angry, sometimes miserable, but always insightful.
If you ever see this, please let me know how you're doing. Private message, or email, or look me up in the Montreal phone book. My profile has my real addresses, and I use my real name online.
-- hendrik Current Mood: melancholy
|Friday, March 6th, 2009|
|Tuesday, January 13th, 2009|
|Writer's Block: Shops Gone By
Woolworths shut its doors in the U.K. last week, sending many into a frenzy of nostalgia and bargain shopping. What now-closed store or chain do you wish was still open?
Marks & Spencer. I know it's still around in the UK, but it's gone from Canada.
|Friday, December 5th, 2008|
|Writer's Block: Gone but Not Forgotten
Many beloved television shows are no longer with us, like Buffy the Vampire Slayer, Six Feet Under, and Mystery Science Theater 3000. What defunct television show do you miss the most?
Crusade, the sequel to Babylon 5.
|Thursday, November 20th, 2008|
|Writer's Block: Smoked Out
Beer and cigarettes once went together like bread and butter, but now smoking in bars is banned in many cities. When you see smokers standing outside bars in the cold and rain, what is your first reaction? Walk on by, join them, or scorn them?
I feel sorry for them.
|Wednesday, October 29th, 2008|
Here's a meme that's going around, and one I can get behind:Copy this sentence into your livejournal if you're in a heterosexual marriage/relationship (or if you think you might be someday), and you don't want it "protected" by the bigots who think that gay marriage hurts it somehow.
|Monday, October 20th, 2008|
|New year's resolution.
Of all the new Year's resolutions I've ever made, this year's has been the hardest to keep.The things that have to be done a little at a time on a regular basis, like ten minutes a day -- Do them.
What kinds of things? Things like reviewing Japanese vocabulary. Ten minutes a day is enough to keep one learning. Spending five hours in a single block memorizing is not as effective as if those five hours are spread out ten minutes a day over a month.
I'm just failing completely on this one. I'll may have to reresolve it next year. I think it's important.
|Friday, September 19th, 2008|
|The day the world was not destroyed
about a decade or so ago, someone in the Russian defense system decided that although their electronics were indicating a massive nuclear attack by the Americans, he reported to his superiors that there was nothing special going on. And so the world was spared.
Now this was a specific person on a specific day, and it is now celebrated as a day named after him.
Does anyone remember the actual date and who this guy is? I seem to remember it's at the end of September of the start of October.
|Monday, August 25th, 2008|
Let me quote Hermann Weyl:The continuum of real numbers has retained its ancient prerogative in physics for the expression of physical measurements, but it can justly be maintained that the essence of the new Heisenberg-Schrödinger-Dirac quantum mechanics is to be found in the fact that there is associated with each physical system a set of quantities, constituting a non-commutative algebra in the technical mathematical sense, the elements of which are the physical quantities themselves.
Now as far as I can tell, this is the idea behind the noncommutative geometries the mathematicians among quantum-mechanical researchers have been excited about recently.
What puzzles me is that it seems to have taken seven or eight decades to get to the point that people are taking this seriously. Could it be that operators like -i hbar d/dx don't fit physicists' physical intuition, but real numbers too?
I've had an idea like that myself, too, about ten years ago, but have not the technical facility and background knowledge to develop it -- that the operator representing momentum that is the real physical quantity, and that the real-numbers eigenvalues that we "measure" are just momentary glimpses into a deeper reality.
Isn't the idea that real numbers are the results of measurements also an experimental falsehood? All the experimental results I've ever seen have been rational numbers. Why assume that the actual physical values we can only measure approximately are real? Why not assume they're members of a noncommutative algebra?
|Friday, August 15th, 2008|
|Day 15: clover.
To get something to grow in my front lawn so that the rain doesn't wash all the ground away, I planted clover a week or two ago. Now there are little green dots showing up in the bare patches. They're pretty. They're growing where the grass won't. But within some of the bare patches, the clover seeds have been washed to one side by the rain before they took root. I may have to reseed those spots and hope.
Last year, the rain here has occasionally been intense enough to wash some of the grass out of the ground. I'm hoping clover will reduce the impact of the drops a little better, and that it has better roots. We will see.
|Tuesday, August 12th, 2008|
|use of Janus (August writing, day 11)
I was asked, what am I doing with Janus?
It's the intermediate code generated by my Algol 68 compiler. There's a SNOBOL program that doesn't quite translate it into 360 assembler (I say "doesn't quite" because I'm sure there are still bugs to get out. The compiler only correctly ran a part of the Algol 68 test suite, and I would be very surprised if none of the deficiencies were in the Janus translator.)
So I have two approaches
-- translate Janus to machine code (probably a new translator). or
-- bypass Janus and translate the Algol 68 parse tree to machine code (what I was doing originally before compilation limits in the compiler I was using to compile Algol 68 H forced me to break things up and use and intermediate code.)
In combination with that choice, there's a choice of a local code generator -- there are a few now that I could use, I'm thinking of LLVM and C--, but there are probably more. There I have another choice -- to use one or the other, or neither and, say, generate assembler myself.
The questions are: which is likely to be more work, and which is likely to generate better code.
Since Janus can be compiled on a line-by-line basis with very little context (after all, it can be done by the STAGE2 macro processor), it is feasible to compile straight to assembler, the usual process for compilers on Unix. But debugging assembler is *hard*. (thought: it might not be so bad now that the world has usable debuggers). LLVM provides a fair amount of (optional) syntactic and semantic checking on the intermediate code, which I suspect will deal with most of the idiot-level bugs (and mst bugs are at that level.)
I could: parse JANUS and do pattern-matching on the parse tree to recognize stuff and generate code accordingly. I could do that parsing and matching easily enough in C, or in Scheme.
I perverted one thing in Janus -- the nesting of procedures. Janus doesn't do it. The code I generate does. The reason is that when I reach a procedure body in the normal tree-walk of code generation, the internal data structure in the compiler is just right for processing the nested procedure. To move the generated code out of the enclosing procedure would have been awkward, given the poor facilities available at the time for managing large text buffers. 400K memories were rare, and hard to get even if available; putting them in temporary files to be reprocessed later would have been awkward too, given OS/360's style of file access. But this is no problem now. Gigabyte memories are becoming the norm. What a difference a few decades has made! So it's perfectly feasible to implement text buffers in C, and let Algol W generate code into them. I've already got most of the API for that, except that it writes it all into a file instead. If I can manage to do something with the new varying-length strings Glyn has put into his Algol W compiler, it'll become even cleaner (the existing way I handle strings to be written to object code is to enclose the actual text in another inner layer of quotes to avoid the fixed-length-string restrictions of Algol W). If I use Janus as intermediate code, even if I choose to translate it into LLVM, I won't have a problem with interfacing Algol W with C++, which might or might not go smoothly.
Instead of assembler, I might be able to access a kind of low-level code generator I threw together a few years ago to generate code directly into memory for immediate execution. Leave that for later, if ever -- gdb doesn't understand this kind of code.
Does LLVM even have the kind of data structures Pascal and Janus use, with variants and such? If it does, are the ways of initializing them well-defined? Or is it just a matter of hoping future changes wont break what happens to have been implemented?
|August writing, day 10
Joseph was not an ordinary warthog. He was a smooth warthog. The other warthogs were snooty and complained about his smooth skin. ou are a smooth warthog, they woauls say, whereas I am a warty warthog. But Jseph was not this is really stupid, says my internal edotor. This is the kind af crap only a nanowrimoer could write. And it's not even Novembers. Go away, go away, dratted editor, I cry aloud. Your place it not here! Thiw is freewriting. Whatever drivel I produce is the right drivel. Nothing else will do. Begone, to the pit of profanity where you can bubble and blaspheme all you want. You will be recalled when needed. But the Editor would not leavel. He complained about toe L at the end of the worl "leave", and about the e on the to. I grabbed my head and pulled at my hair. This would go nowoere, the edotor was now complaining about the metarecursion in my story. Ant that's still more of ot! he cried with clee. How can you bring yourself to put down page after page of this crap! I said, it's only ten minutes, not pages and pages. But it would be if you were a decent typist. Then at least I could complain about style and characterisation, instead of all these stupid typos. Hey, I said, Typos are not the issue. THe ate , I repeat, not the issue. So why are you bothiering with this? Your time is to come, when I have the next draft of my novel written! Go off and complain to schoolchildren somewhere that they are -- oh I don't care. Jsut leave me alone.
I picked up my pen,. arising from the keyboard, and sprayed corrosive ink on his cut-and-paste sword. He parried, and I ended up just covering the tiger lilies in black spots. Aha! he exclaimed sarcastically, YNow you'r eon the right track! Those ink blots fit natirally with the spots on the lilies! Why can't you do somehting like that in prose. You're only a second-rate artist, but that's a lot better than you are a writer. How can you call yourself a writer when you can't even get a cleanly spelled forst fraft?
I was now completely distracted from what I had inteded to write -- I didn't even fell like I was freewriting, which had been the original ecercisel I was running around the back lawn, where I had written so peacefully only a week or so ago, trying to conquer the infernal editor. And, by the way trying not to do physical damage to the lawn while ai was at it.
end of freewrite. I can't even edit out the typos, because they'rre part of the plot!
|August writing, day 9
Nothing written yesterday, the lucky day chosen by the Chinese for the start of the Olympics. I have great trouble doing *anything* ten minutes a day on a regular basis, whether it's practicing katakana, exercising, doing a small increment of writing, taking a shower, brushing my teeth, doing exercises, clearing away junk in the basement, well, anything. Or it doesn't have to be ten minutes. It could be twenty, or forty minutes. So that includes things that are strictly for pleasure, such as watching episodes of a favorite TV show. Something always come up. Spending big blocks of time on something, that's easy. It just has to be something that snags my curiosity, and suddenly I'm spending a day or three researching it. But a lot of things aren't do-it-all-at-once things. They are skills, and unless the same skills show up in a variety of all-at-once researches, I don't get to get good at them. Mind you, there are subjects to which I return time and again, with gaps of weeks or months. But big blocks of intensity aren't the best way to learn some things. Better are many short efforts, ten or twenty minutes a day.
|August writing, day 6
My dentist has a slide show in his waiting room. Lots of travel photos, and a few of his family. But most of them are interesting. He seems to have done a good job of editing them. Some of them, he told me once, are there because they contain a visual pattern that resembles a 'Y', which it the first initial of his name. it could be a tree, a branching river, cracks in rocks, but there's a clear "Y".
There was more I wanted to write, when I was brushing my teeth this morning, but I've completely forgotten it.
|August writing, day 5
At the park on Ile Bizard.
There's a belvedere, at least that's what they call it. You go up there and see out over the river. it reminds me that I've always wanted a camera that could take panoramas -- about one or two degrees high and 180 or more decrees wide. I'm told there is software on some cameras that allows you to zoom into the distance, then turn the camera 180 degrees or more, and it tracks the movement of the image in order to integrate it into one very wide image. I think us hackers could have a lot of fun with a consumer-grade camera whose ROM was reflashable. Especially if it came with programming specs and (for safety) it wasn't necessary for the flashed ROM to contain the flashing software. Or maybe just if it could be made to boot from the same removable SD card it stores pictures on. If only the N800 had a better camera it could be used for this.
A conversation once turned to digital cameras, and I expressed a wish that the camera could use logarithmic coding for intensity, and a photographer there said, "No you don't." I wanted to ask him why, but conversation turned to other subjects and I couldn't get a work in edgewise before he left. There are algorithms involving random numbers that make this possible without much loss of fidelity or storage space. Something I'd like to program if I had access to the raw photoram data as it came in.
|Monday, August 4th, 2008|
|asthma, humidity, and LAN trouble.
Another Monday on the back porch. This time it's a cool sunny day. But I'm sitting in the shade, so the sun doesn't bother me overmuch. I don't need my hat or sunglasses, which is a good thing because I've lost my sunglasses about a month or two ago, and haven't been willing to spring the money for new ones if I might still find the old ones.
I've been thinking of the asthma thing. Hot humidity I can't stand. If I'm in a hot shower for too long so that the air in the bathroom gets thoroughly saturated and the mirrors mist over, I start to breathe heavily, as if it's an effort to get enough air in with all that inhaled water. But cool humidity is great. In fact, it's the recommended treatment for croup. When my kids had it (is seems to be common in the young) the doctor said to set them outside on the porch on a rainy day. (the porch has a roof -- he wasn't recommending that we drench them). Failing that, turn the shower on cold, and have them sit in the bathroom inhaling the cold mist -- again, don't put them *in* the shower -- that would be cruel. The emergency paramedic we once called when my son had trouble breathing said often just putting the child out in the cool night air is enough of a remedy so that when they arrive, the child, waiting on the porch, is already smiling.
So the temperature of the humidity is important. It may well be that the difference is that cool humidity has less a water in it than hot humidity, though how that would be better for croup I don't know. Or maybe hot humidity breeds allergens.
I spent the weekend fussing with our LAN. Suddenly on Saturday morning, our internet connection was down, and, independently, my daughter's computer couldn't connect the the file server where most of her files were stored. The internet problem turned out to be that Primus's server in Toronto (where they route me through even though I'm in Montreal) was dead. but try as I might, I couldn't get the network file system on out LAN working for her.
Next day it all just worked. I suspect that the network file system may have needed an internet connection to work properly, though I can't imagine why. Other NFS connections I set up in an attempt to isolate the problem worked just fine.
Anyway, that's what I've been doing instead of writing recently.
|Friday, August 1st, 2008|
It's August. A mild humid chill in the morning air, blessed relief from the hot weather only a few days ago. The clouds are out, a reminder of the thunderstorm we were promised yesterday. A peal of thunder -- or is it some other kind of noise. Impossible to tell in the city unless it repeats.
There's some lovely orange flowers at the back of the back yard, grown as tall as the peonies we had some time ago. I think these may be lilies of some sort. Gwen has gone to work, and my children aren't up yet. There's some noise, as if someone is tapping on wooden blocks, or bumping them together.
I seem to want to sit and look around. Well, why not do that? The not-quite chill humidity is my ideal weather; the grey clouds above are a delight. I've always preferred this kind of weather. Hot sunny days? No. let other people have them, I hate the heat; the sun is too bright. Give me air a man can breathe in comfort. There. Another gust of mild breeze just wafted over me. Heaven on Earth.
|Tuesday, July 29th, 2008|
Unitarians are often accused of having no definable faith, or of worshiping God, except for leaving God out of the service, and so they can't really be said to be worshiping either.
But really, belief in God, or disbelief is not what we are about.
As a group, we believe in religious freedom, not in a particular version of religion. It would seem that this is a motherhood issue these days, but it is not so.
My heart goes out the members of the Tennessee Valley Unitarian-Universalist congregation. Recent events there
make it plain that supporting religious freedom still costs lives.
|Thursday, July 10th, 2008|
Take a tetrahedron. No, it doesn't have to be regular. Call it tetrahedron ABCD. Pick one of its edges, maybe the longest one. Let's say that's edge CD. Now pick a point E on CD. Maybe near the middle is a good choice. Now divide ABCD into two, and get get ABCE and ABED.
Keep doing this, and get more and more tetrahedra. If you'd like the tetrahedra to line up, you could have the convention that each edge is divided only once, no matter how many tetrahedrons is it on -- all those tetrahedra will have to agree on where they are divided. Or one could systematically, each time divide the longest edge in the whole shebang.
Now while creating this whole shebang of nested, more-or-less edge-matched tetrahedra, you could apply various attributes to the tetrahedron (such as what materials it's mostly made of), and modify them when progressing from a tetrahedron to its components. Or when dividing an edge in two, you could pick a point slightly off the edge, breaking it up into two edges that aren't extensions of each other.
Would this make a usable space division for, say, a mining video game? The nice thing about it is that the division of tetrahedrons in a volume can be compatible with a division into triangles of the surface. Triangle meshes are well-known in computer graphics. You could generate a whole approximate planet by starting with an icosahedron and dividing its volume into twenty tetrahedra that all meet at the centre.
In fact, this generalizes nicely to multidimensional hyperspace divisions, in case anyone should ever want to make that kind of a simulation game. Or physics simulation. Maybe evaluating some kind of Hamiltonian or Lagrangian and using that to affect subdivision statistics would yield something interesting. A multiscale differential equation solver on lumpy space, anyone?
|Monday, February 18th, 2008|
|Yellow, Blue and Green
My favourite colour is yellow. It used to be blue. A few years ago I started using bright yellow filing folders because it made the act of filing more pleasant. In retrospect, those filing folders were probably the first hint that my favourite colour was changing.
My favourite colour had been blue for a long time, ever since I was little, for as long back as I can remember. Even when I was a child my favourite colour was blue, although at the time I had thought it was green.
Funny how one can be mistaken about something like that.