DIARY 2003















Mon, Jul 21

<00:59> Saw The Bicentennial Man. A nice take on the SF story whose basic premise we've already seen in Pinocchio and its later forms like A.I. and The Animatrix. I did have my doubts of Chris Columbus as a SF director, but the end result nevertheless had a little of the oldskool SF atmosphere that makes justice to Asimov's original story. I'm pretty sure that part of this SF atmosphere was due to the actor Sam Neill whom I'm used to seeing in SF and fantasy movies and series.

I'm left particularly intrigued by the issue of immortality that the story poses. It is argued that immortality is not compatible with being human — though one given explanation is that only one person having it would create massive social imbalance, and I agree with that. However, supposing we had the technology to suppress aging (this was the point in the movie as well, not true immortality really) I would like to argue in favour of it.

First of all, the reaction that it's against the 'order of nature' is hopelessy naïve; the same idea could be used against any technology. Taken further, however, this argument says that evolution and reproduction have an important role which would not work with immortal beings. But consider the facts that human reproduction is heavily controlled already, and evolution by natural selection is pretty much screwed up in our modern societies. It's not like the smartest and fittest are out-spawning the dumb.

The possible problems would be quite subtle though. For instance, I'm fascinated by the possibilities for learning and experiencing much more, having the room to make more mistakes. But human brain in its present form is not infinitely flexible; the benefit of having new generations with unspoiled, open minds is quite remarkable. Still, I believe these mechanisms in the brain are heavily linked to aging in general, so the problem of mental stagnation might not exist in the ideal case of suppressed aging. Of course this means more challenges for those developing the necessary technology, but it has little to do with the ethics of the matter.

<01:58> After reading a bit more about Sam Neill, it turns out he played Dr. Alan Grant in the Jurassic Park movies. For me this means a massive piece of synchronicity which I'll outline below. It's not connected to this moment though. I would have noticed it back in 2001, but apparently I'd forgotten the name of that character, having read the first and second novels some years ago.

In 2001 I was in CERN and Jurassic Park III was just being released, though I never watched it. It crossed my mind that it would be cool to watch it just there, next to the Jura mountain range which had given the name to that geological/paleontological period of prehistory, through dinosaur fossiles found there. I'm sure it was much more fun when I climbed to the top of Jura (as mentioned here). Anyway, for the real coincidence here... my supervisor at CERN was one Dr. Alan Grant. 8-/

Sun, Jul 20

<20:51> Weather's getting a wee bit more tolerable, with the occasional rain and thunder cooling and clearing things up. On the other hand my great job-seeking quest isn't going anywhere still. It seems these things are always quiet in the summer, especially July. But a major nail in my career coffin is a study that was mentioned in Helsingin Sanomat yesterday. It said that about 60% of new jobs are never advertised, but the positions are filled via personal relations. Like I have any of those that could be of professional use.

Sat, Jul 19

<18:33> More on the kernel problem: the init glitch is probably due to my updated gcc, which is a development version (3.3.1 preview). I noticed this as I had to recompile a 2.4 kernel and it panicked at boot. Of course I have a few old, reliable kernels available. Besides, Prkl has been using Grub for a week or so, meaning it's easier to select non-predefined kernels at boot. I just updated the gcc which seems to have a kernel-compiling bug fixed. I guess the bleeding edginess of Mandrake isn't always so convenient, but it's certainly fun :-).

Fri, Jul 18

<19:59> The Teknohog Whatsup Paradox: the more things I do, the less time I have to wrote about them. It's painfully obvious, but it's a definite problem in the nature of this diary.

On Tuesday as my parents were away on a brief trip, I spent some quality time boozing and discussing with Antti. I had planned to talk him about the basic of op-amps, but we ended up in the age-old topic of Fourier Transforms as well, though the two are certainly not exclusive. It's still almost debilitatingly hot and we decided to minimize the action and stay home, armed with various drinks and some Chinese food. Can't say we made any huge philosophical or scientific progress but the night was mutually fun and informative anyways.

Wednesday night I decided to give Linux-2.6.0-test1 a try. It worked quite well despite a problem with the USB cable modem: the driver didn't give it a device registration (i.e. eth0). I have already looked into it and will report if the fix works.

Thursday was a full working day at the summer place. In the evening dad and I went fishing, and I caught two nicely sized fish, both with a little pink jig. It was pretty cool: after the first cast with that lure I noticed it isn't swimming too well, so on the second run I changed my style and applied tiny, frequent movements which made it go more or less like a worm. That worked instantly and a pike was hooked. A few casts later I got a bream (Finnish: lahna) of similar proportions. No more guessing what would be for dinner that night.

<23:58> About the usbnet problem: I believe I found the correction, both in the ac2 patch and here. Unfortunately, the kernel now refuses to cooperate with init. It may be a case of non-kernel software then, though I've already updated SysVinit. Mandrake does have module-init-tools (replaces modutils) and other goodies necessary for these beta tests, but there will probably be more similar updates until it works smoothly. Currently I don't like booting up Prkl in this inferno so it'll have to wait :-).

Tue, Jul 15

<03:18> Been infernally hot for the past couple of days. Something around 28°C and humid, with nearby thunderstorms. Just when I commented on Slashdot that Linux is exceptionally cool because it originated in a country where it's bloody cold during most of the year. Apparently these days aren't exactly most of the year.

Trying to find something cool and hackable to do is hard. I have a few PCBs to de-solder, leftovers from old hardware with useful components on them. But that would be just overly hot, considering I'm sweating on the keyboard right now. So, for a non-solder disassembly I decided to take apart AST once again.

I tried to look at the display in more detail than before, to find out if I could connect it to my real computers. Turns out it's a NEC NL6448AC30. 640x480 TFT with 4 bits per subpixel (i.e. 12bpp in total, though the machine only uses 8bpp). That would certainly be nice for many tasks, if not the most graphical ones.

The interface is certainly interesting compared to what I've met thus far. It has HSYNC, VSYNC, R, G, B like any VGA — except the R, G and B are digital, four lines each. So, in principle the signals already are there in most graphics cards. Problem is, they are not accessible. In principle I could re-digitize the RGB components, and with 12 bpp the quality would not likely be a problem. I cannot easily remove the graphics controller chip from AST. There are actually separate PCI cards for these displays, but buying one would defeat the whole idea of hacking.

So, I've now put it back together. I had this cool zen-like feeling as I mounted all the screws in their places while listening to Grieg's melodies. It could have been a scene from a coffee commercial, if you (a Finn) know what I mean.

As I'm currently having a mild interest in the C64 scene and SID music, I felt like retroing around a bit in my own way, so I booted AST into FreeDOS and hooked up a Covox, a simple DAC made of resistors on the parellel port. Played some old STM modules as the old player (DMP 2.78, vintage 1993) messed up XMs (the newer mod format). Certainly brought back memories, and the inevitable glitches (not the least from the hack DAC) were more apparent than ever, playing in all their 8-bit glory through GainHog.

Sat, Jul 12

<18:20> Watched Nausicaa of the Valley of Wind last night, it was only the third Miyazaki film I've seen besides Princess Mononoke and Porco Rosso. Unfortunately the DivX had fairly bad quality in both sound and picture. It was perhaps not the best option to see this after Mononokehime and other more recent anime. In many ways Nausicaa seemed like the first iteration of ideas that were better developed in later works. Especially the trippy moments, where Miyazaki's portrayals of pantheism and grand unity are explicitly condensed, are too short and thin IMHO. It's certainly not bad overall though.

Fri, Jul 11

<20:06> It's been rather quiet with my book-reading experiences lately. There are numerous reasons, one being that there's plenty to read online as well. Yesterday I got through Albert Hofmann's LSD - my problem child. An exceptional mixture of scientific objectivity and mystical experiences. Not necessarily much new information, considering sources like E2 and Lycaeum, but this book brings many facets together quite well.

<23:27> Some days ago, Mr. Palomäki told me about having access to two ISPs at once, sometime in the future. While there would be another roommate using the net, he said it would be convenient to load balance the two connections, so that the full bandwidth could be used by whoever is using them at a time. I didn't have any advice for him then, but the scenario sounded interesting enough that I started to look for the info.

And now I got it working! Actually I don't have two ISPs, but I'm testing it with my internal network which comprises Willow and Prkl with the cable modem. Here's the topology:

              Ethernet                     TV cable
    |                           |
         PLIP            USB

Of course the ISP limits the total bandwidth of the cable so I'm not increasing my Internet capacity very much. But at least I can see that both routes are being used. And I'm getting slightly more capacity between Willow and Prkl ;-). Here are the instructions.

Wed, Jul 9

<04:45> There are some down-to-earth but nevertheless quite complex issues with my audio and computer gear. I really mean earth, which is missing from the mains sockets at that part of the room where Prkl and GainHog are located. Sometimes I get tiny sparks when connecting two devices, because their dangling grounds have accumulated strange potentials. Of course there isn't much current there, but static voltages can be nasty anyway. On a related note, my ionizer gives about 6kV which I was able to measure with a DMM, so its static voltage is probably even higher.

I got a reminder of the earth problem when the cable net was down today, and I went through my side of the cabling, looking for visible problems. When connecting the cable back into the modem, there was another of those tiny sparks. Then another spark (conceptually, I assure) lit up in my head: the cable modem's ground is fixed to Prkl via the USB cable. Therefore I'm getting the devices grounded via the TV cable, which I believe is a proper 'source of ground'. :-)

On the other hand, what if I have some other device, grounded via the mains plug, connected to these? I assume the mains ground meets the cable ground somewhere underground. That way I'd just get a HUGE ground loop, connected to my precious sound system.

Mon, Jul 7

<00:49> As the GainHog measurements gave me some interesting information on Prkl's soundcard itself, I thought it would be interesting to run it on Willow too. Unfortunately its ESS Maestro doesn't handle full duplex, not even with ALSA.

There's a lot I could do better with the GainHog measurement as well. There are issues like windowing choices I haven't considered to their fullest extent, however the soundcard usage in the first place is not very professional and I'm happy with the current results. I would need a scope to fully examine impulse response, which I very much would like to see.

<02:04> For those of you who read about the frequency measurements already, note that the graph was upside down (i.e. the difference was the wrong way :). I noticed this while generalizing the Python script I'd used to handle the data, as it might be useful for other projects. Should be OK now, at least the infrasounds are closer to theory now :-)

<15:00> Having mentioned the soundcard frequency responses before, here are the actual results. I finally got some results from Willow as I switched back to the OSS driver. It seems the ALSA driver is simply broken, and its theoretical superiority doesn't help.

The graph only shows variations as the absolute gains are not fixed in any way, but it shows clearly how Willow is worse at high frequencies. Overall the difference isn't clear, but I'd say that Prkl's SB 128 is better. If only for the experience that it's much less noisy — being mounted in a rather well-shielded location, as opposed to the tightly crammed onboard sound of Willow. The latter even picks up interesting noises like beeps during hard drive transfers :-). You can see this as the peak around 9500 Hz.

Sun, Jul 6

<02:00> After some serious work and play at the summer cottage, I managed to do a long-planned rehaul of my photo gallery. In fact the change is not finished yet, but the basic point is to have the rather general Flora and Fauna sections. I had too many place-specific nature galleries there previously, in fact I still have Nature-Finland but that will probably be changed as well. I thought it's more natural (pun intended) to group them by theme, and I've added quite a few pics from previously uncharted locations. The problem remains though that many 'nature' photos don't fit into either of the two categories. Of course there's only one Nature which even encompasses our technology, so these divisions are quite arbitrary. Just go there and explore!

One new observation of the new amplifier: the channel separation is much better than what the Radiotehnika had. It's quite obvious if you compare the wirings inside, but I didn't expect the difference to be that clearly audible. This will be a good thing if I try Orthoperspecta again some day, as its channels are, in practice, more independent than those of vanilla stereo.

<21:42> Did some measurements of the GainHog.

Thu, Jul 3

<16:31> "Uh-oh.. Rome does not look good." For a moment I thought the thunderstorm that just went past nearby was a sign of increased electromagnetic activity, due to glitches in the rotation of Earth's Core.

As I'd found out before, this movie was a nice example of oldskool SF, in fact I would say pre-classical SF: something in between the sensationalist 'amazing stories' and real, scientifically credible fiction like Clarke. The copy I had was a rather low-quality screener with the most geometric distortions I'd ever seen, but my eyes adjusted after a while. It would definitely be interesting to see this properly on big screen. You could tell that the plot is probably taken from a 1940s/1950s story and souped up for the modern world with elements such as cracking the net. IMHO the pieces fit quite well together, even if both the pseudo-science and the drama had their problems. For instance, I'm still wondering how the terranauts could communicate with the surface base — perhaps via some form of sound waves? Nevertheless it's a nice piece of action with plenty of science and geek humour, not a bad way to spend two hours.

Wed, Jul 2

<00:33> "Maybe there's hope." These last words of Special Agent Fox Mulder will remain my pons asinorum to the memory of probably the greatest thing on TV ever.

Tue, Jul 1

<17:04> Yesterday I finished the first 'milestone build' of the amplifier. All units such as the transformer are now steadily mounted in the case. The top plate is now isolated from the ICs and grounded, which has reduced the mains noise to practically nothing. There are still future improvements planned as I've listed here, but the system is now fully functional and usable.

I even built the external attenuator to be used with my CD player. However, I still prefer the software controls with my computers, as it makes the component count lower. For this reason I've now added volume control to comms, after noticing how much I use it with XMMS and MPlayer (too lazy to fire up alsamixer :-).

<19:48> Since the amp is working well, I have considered adding my existing surround speakers. However, both of the times I've been about to do this, I've noticed how well the soundscape is being distributed as it is. It relates to a previous observation of a more accurate spatial picture, and reinforces my notion that for people with two ears, two channels should give enough information, even for sounds appearing from the back. Besides, for acoustic recordings the surround system will completely screw the actual soundscape.

I have also conducted some actual measurements of the case temperature. The amp has now been on for a few hours, listening now and then, and the inside air is only at 33 °C. The trafo does heat up more than the ICs, but still not too much.

Risto A. Paju