Up: Making Sense of This Site | [Related] «^» «T» |
Monday, June 11, 2001
Context and Web-memory
By Paul Ford
This is a summary of my technical background, and if you don't care about things like databases and XML, skip it.
Around 1995, in college, I began writing Web pages using flat HTML files, entering links by hand. I used some “server-side includes” and processing tools to make it all come together, to avoid updating every page.
In particular, I found a groovy pre-processor called htp, which I used to build the culturefront Web site for the New York Council for the Humanities (which only looks a little like my original design, now). Htp was my first programming language, and NYCH was my first client.
I gave up HTP because it's a Unix tool, and I had to use a Macintosh. I gave up the NYCH in 1996, because I'd found full-time work.
In 1997, the Subway Diary was born, written in a text file with little stubs in it; MacPerl sniffed around in the file and generated a bunch of web pages.
1998 Ftrain was born, and like an ugly adolescent, it was caught between worlds; while its data was represented in XML, it was still coded in Perl, in code so atrocious as to make a monitor melt and good programmers rip out their own eyes.
By mid-1998, I'd given up on XML, and Ftrain was published to a database on a Web server, with inter-linked categories and automatic listings and so forth. I posted the entries directly into the database through a web form. They were pulled out on the fly by a program that ran on a web server. This was fine, except the hosting provider put too many people on each machine, and their server often crashed. The database did the sorting and processing for me; all I did was write prose and make choices.
In 1999 I took some time off, but in 2000 I returned to XML, and the functional programming language XSLT. Until computers exhibit human-like intelligence, this is the framework I'll be using.
Ha, ha, you say. No, I'm not fibbing; document trees, which can be traversed by functional scripts, is about as far as things can go; past this, we get into some really blurry linguistic and cognitive territory where mortals dare not go. The next step is to create an artificial intelligence that can help you connect your ideas together. We have a little while.