Prentiss Riddle: Toys

aprendiz de todo, maestro de nada

Prentiss Riddle
aprendizdetodo.com
riddle@io.com

 
home art austin books
causes chuckles garden
kids language movies
music time toys travel
 
Search this site

Archive by date
Archive by title
RSS/XML

XML::Weblog->Hello("World");

Unlike some blogniks, I don't really like to natter too often about tech here. But I've passed a few small milestones that may warrant a recap.

Having been unhappy with performance at blogger, I started out rolling my own blog in raw HTML using SSI to glue the pieces together. When the cut-paste-link-test cycle got tiresome, I hacked a bit of PHP to dynamically assemble page views from what are still underlying chunks of raw HTML (I just can't give up final control :-) ). Once I got into PHP, it was simple to add a discussion feature with Dotcomments.

Then I found myself wanting to play with XML and RSS/RDF syndication. The first thing I did was to use XML::RSS to write my own little aggregator in perl, the result of which you see in the "In other news" sidebar on my top page. It grabs a bunch of RSS feeds from various places (including a couple I generate myself by screen-scraping someone's HTML) and combines the results with various crude weighting and randomization tricks to try to produce a list of headlines I'll find interesting. I'm still playing with the details -- it works pretty well except that certain persistent links tend to float to the top and stay there. I also found that the XML:RSS parser is way too finicky about errors in incoming XML, so I reimplemented it with the more forgiving RSSLite.

Since I'm accepting other people's XML I also wanted to produce my own. Again it seemed that XML::RSS was overkill, so at the moment I'm just scraping my own screen (so to speak) and producing an RSS 0.91 feed which you are welcome to grab here. XML/RSS/RDF I've registered it with UserLand, as far as I know the only general-topic aggregator still out there after the demise of my.netscape.com. It's nice to see my name up in lights but I don't know how that translates into traffic.

In addition I registered at Weblogs.com, which just lists recently updated weblogs rather than aggregating them item-by-item. The technique for letting Weblogs.com know about updates is really simple: despite a lot of blah blah about XML-RPC and SOAP, you can do it in one line of a shell script by invoking lynx against their submission form with a couple of arguments supplied after a question mark.

That's more detail than most of you wanted to hear, but possibly someone may Google their way into this page and have questions. If so, just holler. I'm happy to share.

toys 2001.11.09 link