Wednesday, 21 August 2002

Distributed reality Blog syndication clarification. [...] But there would still be a server-side part of it. It would store the list of blogs, the list of keywords, and the list of categories. A category is simply a list of keywords and other categories. I.e. to get the feeds in a category you run a search against Lucene with the keywords in it, where the set of keywords is the transitive closure of categories (since they can reference one another). The client can then update its list of blogs,keywrods and categories periodically. This will be a reasonable load on the server, so wouldn't need that much in terms of hardware. There's no web content generation involved either, just SOAP calls. [...] [Random thoughts] The problem with requiring a server side component in anything is that said server forms a single point of failure. Given what you appear to require of this server is just a way to collate distributed state information (collection key word suggestions, blog address, ...), all of which can be handled by defining xml file formats and/or publishing on home pages akin to blogrolls. So really, there is no *need* for a centralized server, and I really have no wish to hand over authority to someone else. 5:45:49 PM

IBM GSA Who is using Tomcat or Jetty in production. Interesting stuff. Lots of companies seem to be moving away from IBM and BEA to the open source alternatives. [James Strachan's Radio Weblog] Back when I worked for a new media offshoot of an old media company, we got quoted something in the order of 2 million bucks (aus dollars admittedly) for an install of webfear. The quote came from GSA tho ... 3:46:18 PM

© Copyright 2002 Brett Morgan.

[Macro error: Can't call the script because the name "nextDayLink" hasn't been defined.]