More news from the SF-come-true department:
Researchers at the Princeton Engineering Anomalies Research program, or Pear, have been attempting to measure the effect of human consciousness on machines since 1979. They are trying to determine if the output of machines could be slightly but quantifiably affected by people's thoughts. The experiments involve the use of random event generators (REGs) as participants concentrate on controlling the machines' output; initial tests employed REGs that produced high-frequency random noise that was converted into ones and zeroes, while subjects focused on first generating more ones, then more zeroes, and then nothing.
Read more on this Wired article.
I am subscribed to quite a few mailing lists on programming (either directly or via gmane's gateway), but if i had to take just one with me to a desert island, i'd have no doubts: ll-discuss. Originated in the MIT's Lightweight Languages workshops, it discusses programming at large. There you will find great names in the programming languages community (do Paul Graham, Guy Steele or Peter van Roy ring a bell?). The list has a schizo behaviour: it works by fits of high traffic interleaved with long periods of inactivity. Were're right now in one of the former, including a very interesting thread on Guido's bias against functional programming (one of the reasons that makes me stay away from Python, by the way).
And, if you're looking for a news site, Lambda the Ultimate is, of course, the one.
Time machines are, or have been, limited to the realm of SF. Theoretically, there is the possibility of travelling to the future via the twin paradox, and unreal models such as Godel's Universes have what is called closed time-like curves in the jargon: paths along which one could travel to her own past. Amos Ori has just proposed a doughnut-shaped model with such curves. As Ori himself admits, the model is "wild speculation", but fun anyway.
If you feel intrigued, Paul Davies' How to Build a Time Machine may interest you.
You may be wondering why we need yet another build system. I have outlined my reasons in the draft Conjure users manual, but there are other opinions that point to similar directions, like Adrian Neagu's in this recent article. The system closer to Conjure mentioned in there is SCons: makefiles are replaced by Python scripts, greatly improving their expressivity and means of abstraction. Conjure starts with similar goals, but will be based on Scheme.
Recently, make replacement proposals abound (see for instance OMake, based on OCaml, or bake, which also uses Python), leaving me with mixed feelings. On the one hand, it seems clear that the need for a better build system is real. On the other hand, I wonder if, in view of such proliferation, Conjure is really worth the effort: should I focus my time-limited efforts in some other free software need?
Before leaving a couple of months ago, my research at the university was centered on mobile agents, and i was playing with the idea of implementing a distributed system based on Scheme, and, specially, continuations. As models, i had previous systems like D. Halls' The Tube, and Kali (based on Scheme48). Now, the Gambit guys at the University of Montreal are developing Termite, a Lisp for distributed computing that is almost exactly the kind of thing i had in mind, only better: it's based in Erlang's distributed concurrency model, for a start.
Definitely worth a look.
I have just finished Anarchism: from theory to practice, a fine book by Daniel Guerin (with a preface by Chomsky) on libertarian ideologies. As I suspected, the ideas of Proudhon, Bakunin, Voline or Kropotkin, to name a few, articulate quite nicely some half-baked ones I have had for a long time. A libertarian society is, from my point of view, a very desirable one, provided that Hobbes was wrong and we humans are, essentially, good (a point so easy to put into question). Noam Chomsky has described quite accurately the kind of ideas I am talking about.
Over the years, I've developed a dockapp addiction. These are the ones currently running on one of my ion3 frames:
From left to right:
wmclockmon -sc -ca
wmcpuload
wmmemload -b -c
wmtop
wmnetload -b
wmibam
wmmaiload
wmdiskmon -m -p /dev/hda3 -p /dev/sda1 -s
And yes, there's an underlying theme in there. To ensure that they are
launched in the proper order, I use this scsh script in my .xinitrc, like
this:
exec wm-with-dockapps ion3
I am using a new color theme, with slightly lighter colors (still dark background). I dubbed it, rather unimaginatively, bgg. It looks nice with my current ion3 style, by Matthiew Moy, who also has written emacs-like keybindings for ion.
Although my favourite Lisp dialect is Scheme, I keep an eye on other variants, like Elisp or Common Lisp. When you want to get serious about any programming language, it's always a good idea learning about its roots. This is even truer in this case, since Lisp has been around for a while. Paul Mc Jones has set up an excellent History of Lisp page, full of interesting contents, including early Lisp books. You can read an overview of that contents in this recent blog entry by Glenn Ehrlich. Finally, there are additional bits of Lisp history in this interesting page.
Playing with e-w-j again: this time, i just set up icons for the
entry categories... as a matter of fact, this is a fake post to test
them ;)