State of play...

posted on Monday, February 14, 2005 1:57 PM

Well, sorry I've been light on the posting for a couple of weeks - been a bit caught up in some stuff going on (my uncle died) as well as starting a new project at work. I haven't been entirely away from blogs though, here's a few articles I've been reading recently:

• XML Performance Checklist, and some issues on XPath evaluation
• Neural Networks in .NET
• Customize User interfaces and Pass User Information to Installer Classes
• Customizing your .NET deployment projects

Also been buying a couple of decent books (and a couple of crap ones which I won't mention)

• ASP.NET Cookbook-Pretty decent book, lots of little ASP.NET snippets, not for pure beginners but has some interesting solutions to stuff which can be confusing.
• Head First Design Patterns - Certainly provides a novel way to examine the use and implementation of many of the GoF patterns, unfortunately it's in Java so yopu have to do a bit of manipulation to get the examples working in C# (also you have to know when C# implementations are different than Java, e.g. delegates and attributes)
• The Definitive Guide to Windows Installer - Basically covers how to use Windows Installer, good book but I'd have like to have seen more involved coverage of using VS.NET - covers it but kind of lacks the depth I wanted.
• Artificial Intelligence : A Guide To Intelligent Systems - Not actually got this yet, my current minor obsession (influenced by this video on Channel9)
• Introduction To Machine Learning- As above...looks like a pretty decent book.

Apart from that, I'm currently looking at using the XPathNavigator for use in a path-sensitive URL rewriting system for a new application I'm building at work- so urls like http://me/you/other/thisitem.aspx can be mapped dynamically on to a 'sitemap-like' navigation structure. Trying to avoid using RegEx based rewriting ala .TEXT - in mine the path defines the context so I don't need to carry object identity information in the URL e.g., http://www.mostlylucid.co.uk/archive/2005/02/04/1649.aspx - in mine I would have the path info as well as the object identity statically mapped in XML (well, editable but rarely updated...). Any opinions / great implementations you've seen of this sort of thing (information on most efficient in-memory storage using binary-search-trees (e.g., Scott Mitchell's series on data structures) etc...would be greatly appreciated)