posts | images | bookmarks

Ted Nelson

By anders pearson 28 Feb 2004

today i got to go to meet <a href=”http://ted.hyperland.com/“>Ted Nelson</a>, the inventor of <a href=”http://dictionary.reference.com/search?q=hypertext”>hypertext</a> and visionary behind <a href=”http://xanadu.com/“>project xanadu</a>, the “<a href=”http://www.wired.com/wired/archive/3.06/xanadu.html”>longest-running vaporware project in the history of computing</a>.”

he was at columbia giving a workshop on <a href=”http://xanadu.com/zigzag/ZZdnld/“>ZigZag</a>, which is… well… after three hours of listening to Ted and playing with it, i’m still not sure exactly what to call it. it’s sort of an advanced data structure that Ted believes can be used to hold any kind of structure whatsoever. sort of a relational database exploded into multiple dimensions. but also a basis for a document model, or possibly a filesystem.

the man is clearly brilliant, if somewhat out there. he literally sees software design as a branch of filmmaking. much of his work is motivated by the extremely non-linear way that he thinks and his hatred of hierarchies (i never did get the chance to ask him what he thought about BeFS and <a href=”http://www.namesys.com/whitepaper.html”>ReiserFS</a>) and the limitations of paper (and computer representations of paper). unfortunately, he isn’t a programmer himself and he seems to have had personality conflicts with many of the programmers that he’s worked with, resulting in a long string of half-finished projects. the personality conflicts i think were because he views himself as an artist and, like Orson Wells or Frank Lloyd Wright, insists on having complete control over every single minute detail of his creations.

i don’t really have a point. while he may never get around to actually producing anything directly usable, he has some really fascinating ideas, so i think he’s worth paying attention to.

introducing cgi_app

By anders pearson 10 Feb 2004

i spent many years writing CGI applications in perl. lately, i’ve been writing more and more code in python.

even perl's detractors have to admit that perl has an amazing collection of high quality modules available for just about anything you want to do. as much as i've been liking python, when it comes to CGI programming, there were some niceties i missed.

in particular, the powerful combination of CGI, CGI::Application, and HTML::Template made writing well organized web applications with display logic and business logic cleanly seperated a breeze.

so, scratching an itch, i've ported CGI::Application to python, incorporating some basic CGI.pm like functionality and integrating with Tomas Styblo's HTML::Template clone.

if you haven't moved on to something more sophisticated like Zope or mod_python and just want to write simple CGI applications in python without a lot of hassle, check out my new cgi_app module.

orkut

By anders pearson 05 Feb 2004

let me know if you want an <a href=”http://www.orkut.com/“>orkut</a> invite. i’m only going to invite people that ask.

from what i’ve seen so far, it’s basically <a href=”http://www.friendster.com/“>friendster</a> and <a href=”http://www.linkedin.com/“>linkedin</a> combined and implemented by some programmers who know a little more about writing scalable, usable web applications than the people at friendster.

i’d also like to point out, for the record, that i built a web-based social networking tool for a columbia business school class in 2000 (that’s 3 whole years before friendster). it was basically a mini-friendster where the students in the class would enter their contacts (only among other students in the class) and they could then see their network visualized (using a nifty java applet), and it would show them various statistics about their network and about the aggregate class network. the prof would then use these networks to demonstrate various concepts.

i just wasn’t smart enough to get venture capital, dammit.

update (2004-06-12): i thought it would be obvious, but apparently it isn’t. the offer for orkut invites was intended for people that i actually know at least in some small way. the complete strangers who find this entry on google need to stop asking me.

924

By anders pearson 28 Jan 2004

i’m actually kind of excited about the impending snowstorm that’s supposed to be hitting nyc for the next couple days.

my poor old combat boots, which i’ve basically worn every single day for about 5 years (and they were used when i purchased them), are finally starting to show their age. the sole has started to wear through. if i’d taken much better care of the uppers it might be worth getting them re-soled, but i didn’t. plus, i only paid $30 for them, so it would probably end up costing me more to get them fixed than they are worth.

so i decided to retire them and ordered some <a href=”http://www.batesfootwear.com/Catalog/buydetail.asp?ProductID=8424”>Bates E00924 Boots</a> as replacements, which came in the mail today. they’ve gotten fantastic reviews and are apparently the boots issued to SEALs for training in.

so the snowstorm gives me a good excuse to really abuse them and see what kind of traction they’ve got.

they aren’t properly broken in yet, so i can’t say much about them. but they are amazingly light (i think my dress shoes weigh more) and the DuraShock soles aren’t as squishy as i’d feared (my old boots just had hard rubber between my feet and the pavement and i liked it that way). they kind of squeak when i walk on tile floor, but i’m hoping that’s just a new boot thing and they’ll quiet down. i’m also a little worried that the treads will track mud.

anyway, nothing like some sexy new combat boots to make a snowstorm more enjoyable.

wiki

By anders pearson 23 Jan 2004

we’ve been using <a href=”http://en.wikipedia.org/wiki/WikiWiki”>wiki</a>’s at <a href=”http://ccnmtl.columbia.edu/“>work</a> for a while now with some real success.

last night, the power supply in my main machine at home blew, so that will set back the development of the new thraxil engine somewhat. to make up for delays, and keep people entertained while i work on it, i decided it would be interesting to set up a wiki for thraxil and see what people did with it.

so <a href=”http://thraxil.org/wiki/“>here it is</a>. go nuts. just don’t abuse it or i’ll take it down.

python meetup

By anders pearson 10 Jan 2004

went to the new york <a href=”http://python.meetup.com/“>python meetup</a> last night. it was pretty good. plenty of smart people.

one guy there was a developer on <a href=”http://www.stackless.com/“>stackless python</a> and <a href=”http://www.twistedmatrix.com/products/twisted”>twisted</a> so we got a pretty good overview of what’s going on with those projects. other topics of conversation included <a href=”http://psyco.sourceforge.net/“>psyco</a>, <a href=”http://openmosix.sourceforge.net/“>openMosix</a>, <a href=”http://www.pfdubois.com/numpy/“>Numeric</a>, <a href=”http://c2.com/cgi/wiki?InfernoOs”>inferno</a>, and the trials and tribulations of porting C code between architectures (to OS X in particular).

other random observations: all 6 people who showed up were employed (it seems uncommon; every other programmer gathering i’ve been to lately has always been half unemployed people hoping to make a contact to find a job). none of us were windows users; all Linux or OS X users (and all of us owned at least one piece of Apple hardware, whether it ran OS X or not). i didn’t ask, but based on the conversations about type inference mechanisms, closures, continuations, compiler and runtime optimizations for dynamic languages, and methods of proving program correctness, i’d guess that everyone there had at least some formal CS background.

semiconsciousness

By anders pearson 09 Jan 2004

reminder to self: when planning to take a nap while listening to music, take a minute or two to put together a playlist; don't just rely on shuffle mode over the entire collection. i just have too many weird, creepy ambient tracks in there and run a significant risk of dozing in and out of a 20 minute long track that consists of howling wind noises, distant buzzing sounds, screwed up tape loops and ominous whispers with a title something like 'never answer the telephone'.

this is particularly confusing and unnerving when combined with my tendency towards sleep paralysis when i sleep at odd times.

black chocolate stout

By anders pearson 09 Dec 2003

Gerard

it took me about 2 hours to drink a single bottle. it’s extremely thick and dense; it noticably bends light in its vicinity. the chocolate taste is very subtle and mostly present in the aftertaste. if you like stouts, i recommend it if you can find it. if you don’t like stouts, well, there’s no hope for you…

bounties

By anders pearson 03 Dec 2003

i'm really fascinated with the recent gnome and mozilla bounties.

i think the idea of offering bounties for development of open source code has a lot of potential. if they opened things up so other users could contribute to the bounties, it might work even better. by myself i probably don’t have enough money to really make it worth anyone hacking in a feature that i really want, but if there are others like me, each willing to contribute to a bounty, we could probably come up with a decent amount of cash.

one very apt criticism of open source software is that the programmers tend to work on things that programmers care about or are fun to work on. end users’ needs are often secondary. a nice bounty system could really improve this situation, letting actual users of the software vote with cash on what features are actually important to them. plus it helps programmers who want to work on open source code generate some income without having to sell out or get a day job.

the whole free-market approach might also help to shake off the open source community’s ‘commie’ image in the eyes of the hardcore capitalists who insist that proprietary software must be inherently better because it has monetary incentive. but it does it without actually giving up any of the advantages of open source.

of course there are all kinds of practical problems to be overcome to make it actually work, but i’m confident that it’s possible.

even without an infrastructure in place i still think that large companies and governments should look into directly funding open source projects. the US government must spend millions each year on windows licenses. if they came up with a list of issues that are actually preventing them from switching to linux and presented it to the open source community saying "we're about to give microsoft $X million dollars. if you can fix these issues, we'll instead give you that $X/2 million dollars." they would save a lot of money, and the whole world would benefit from the improved software. since microsoft operates on 85% profit margins, it shouldn't be hard for the government and maybe a few big corporations to come up with enough to fund the actual development of what they need.

honestly, i think that these kinds of models, if implemented well, have a greater long-term viability than the current commercial software industry. open source projects don’t need the overhead of an advertising department or overpaid executives.