permalink.gif 2004-09-27

permalink.gif Models of interest

Mon Sep 27 21:54:58 BST 2004  Permalink 

I'm trying to come up with a mathematic model that allows me to expresses interests, their strengths, and how they change over time. My original thinking was simply to model what is current as most interesting and anything prior to that as less interesting on some scale. But I started asking myself questions about how interests are shaped and changed.

If I've been talking about nothing but, say, XML Topic Maps for a year and then I spend a week talking about my recent vacation to Hawaii. Well probably XTM is still the more important topic. So the model has to encompass ideas like continuity of interest.  In the more realistic world where we talk about different things at once and leave and come back to topics as our interests wax and wane, well lets just say I don't have a clear shape for this in my head yet.

permalink.gif Half way between a SAX and a DOM

Mon Sep 27 16:03:40 BST 2004  Permalink 

Ian E. Gorman has released GXParse 1.5, a free (LGPL) Java library that sits on top of a SAX parser and provides semi-random access to the XML document.. Ian E. Gorman has released GXParse 1.5, a free (LGPL) Java library that sits on top of a SAX parser and provides semi-random access to the XML document. The documentation isn't very clear, but as near as I can tell, it buffers various constructs like elements until their end is seen, rather than dumping pieces on you immediately like SAX does. This release completes namespace support is complete, eases exception handling, and adds a few operators to CurrentElement. [Cafe con Leche XML News and Resources]

This looks like a really useful library for XML parsing.  Quite often I envisage potentially big data sources which rules out DOM tree construction and then, as I am buried in SAX, begin to regret that 4GB of memory is not the default these days.  GXParse could be a very useful half way house and I may have need of something like it soon.

permalink.gif Aging gracefully

Mon Sep 27 08:02:53 BST 2004  Permalink 

I'm looking for algorithms that will model the aging of data.  I'd be really grateful if anyone could help me out with some pointers to existing algorithms and such.

What I am trying to do is come up with a good system for weighting information such that more recent information is considered more relevant than older information.  But I want it to degrade nicely and I want it to be flexible since I think I'll have to tinker a lot.

I'm playing with simple logarithmic functions but not convinced.  Has anyone trodden here recently?