There seems to be a meme running through the Lisp community lately: movies. To be fair, there's tradition: every lisper has seen the Lisp Machine videos, specially those too young to have seen the Genera interface in the flesh.
I already mentioned the DSLs in Lisp video. Marc Baringer is joining the fun with an Uncommon Web hello world (UCW is a continuation based webapp framework) and a wonderful Slime tutorial. Slime is a powerful Lisp environment for Emacs, and even if you're not into Lisp, the video will give you a taste of the kind of things that make working with Emacs such an awesome experience.
So, let's go to the movies!
Japan has announced plans to construct an amazingly powerful supercomputer for modeling climate change, the formation of galaxies, and new drug behavior. The machine is expected to boast a peak performance of 10 petaflops: that means it will be five times switfter than then 500 fastest systems on the planet today... combined. The expected cost is almost one thousand million (and no, that's not a billion by here) dollars.
Right now, the fastest computer in the world is IBM's Blue Gene.
As i mentioned the other day, we were trying darcs at work as our SCM. But, as our needs have become clearer, we have decided to return to our original plans and use GNU Arch. Darcs is a pretty nice little toy, a pleasure to use, but too limited for our needs. For instance, we need to handle projects consisting of several "configuration items" (subprojects) handling versioning in a rigorous manner. We need to maintain different repositories with different roles (development, master, archive) seamlessly integrated regarding version control. Arch, via build configs and tagged branches, offers all of this out of the box. With darcs, we would need a series of hand-made scripts and written conventions to do the same, but not quite. Also, having symbolic links under version control, or commit hooks is, if not a strict requirement, at least a great plus.
Moreover, we are Emacs guys, and there's nothing like xtla in darcsland. Combined with the bazaar front-end, we have a great head start in alleviating Arch's steeper learning curve (and we are willing to tackle more complexity in exchange of more power). Finally, we don't mind Arch's rigid (some people would even say fascist) naming conventions: as a matter of fact, in our project they're a plus, since it is in the aerospatial field, with lots of formal process and rules and standards and, oh well, fascist quality assurance practices.
So there we go!
I am always shifing window managers. Over the years, i've tried many, from owlwm to metacity, but i've stabilized in two types: the WindowMaker mileage (Afterstep, Blackbox, Openbox, Pekwm) and ion3, a mileage on its own (together with Ratpoison, but i've never used that). I've never been fond of bloated desktops.
During the last months, ion3 was my default, but a coworker returned to Blackbox last week, and i tried it again. Having non-tiling windows felt good, so, with the new laptop, i thought of turning back to my favourite of old: pekwm. This is how it looks:
Isn't it pretty?
New job, new laptop, courtesy of my current employer. Saturday I bought a Samsung X50 notebook. The nearest contender was a Powerbook, but since I do not intend to use Mac OS X on this computer, I thought twice.
I am very happy with the new toy: 1Gb RAM, 1.6 Ghz Centrino CPU, 60 Gb HD, and a shinny 15.4" panoramic display with WXGA driven by an ATI Radeon 128M (probably my favourite one). All in 2.4 Kg. Not bad. And GNU/Linux support is quite good: I've devoted the entire disk to a Debian Sid. Wireless and hibernation are working, as well as X.org (the Debian installer figured out the correct config without a glitch, including a 1200x800 modeline).
I'll be reporting on how to configure this little beast. I guess i am nearly completely unswitched by now.
Qi 6.2 is released. This language, which brings static type checking and pattern matching to the Lisp family, is the base of the IFPE project. This Qi release conforms to the text Functional Programming in Qi and contains some patches to 6.1. and minor corrections to the HTML documentation.
For download and more details see the downloads page.
Abelson and Sussman's Structure and Interpretation of Computer Programs is the best book I've read about computer science. And its video lectures are a pleasure worth the huge download, if only to know why CS is neither about computers nor science. Now, they are available as a torrent. Don't miss them!
The idea of enriching web contents with semantic data is almost as old as the WWW, actually pioneered by Tim Berners Lee as the Semantic Web. Based on metadata (in the form of RFD) attached to documents, it aims at leveraging the huge knowledge base that the Internet has become.
Problem is, that metadata is invisible for users. Only after computer processing does it become useful. Contrast that to links, or the vCard and iCalendar services. There, you can actually see the extra info. In addition, no new languages or standards are needed: simple extensions to XHTML suffice. The Microformats inititiave relies on these points to propose a new, small and simple way of taking the Web one step further. Tantek Celik, one of its proponents, gave a presentation on microformats at Supernova 2005. I think they may well be on the right track.
You're probably familiar with Mendeleev's periodic table of elements, invented (or was it discovered?) in 1869. Oxford ecologist Philip Stewart has designed a new periodic table of the elements that shows the elements as a galaxy of elements. Sciscoop is running a history about this new perspective.
In a recent post, Brian Mastenbrook observes how hard is to find good tutorials for Lisp, and thinks it's time to rectify and let Google know about good tutorials about Common Lisp, especially good tutorials about Common Lisp available for free... Won't you help us?
