RSS Bandit

I tried Aggie again wand it worked like a charm with nary a thrown exception. This meant I didn't have to take Ziv or Joe up on their nice offers to help me troubleshoot my issues. However I didn't like the fact that Aggie uses a Web interface + native GUI instead of a native GUI with an embedded HTML viewer like FeedReader. So I decided to hack out my own FeedReader clone sans crashes.

This is the first Windows or .NET GUI application I've ever written so the Windows Forms FAQ and Google Groups were a big help. Especially when I started chasing down some extremely cool yet blind leads with regards to COM interop.

I've created the basic functionality besides the ability to track state between runs of the application and allowing all feeds to be centrally managed from a single menu. It was rather surprising that FeedReader didn't have a simple way for me to specify that all my feeds should only fetch feeds once an hour instead of once every fifteen minutes.

In writing this app I've already added a bunch of things to the reason why I despise the lack of checked exceptions in the CLR. Two of these things are the fact that the documentation for both the XmlSerializer.Serialize() and the XmlConvert.ToDateTime() methods do not mention that any exceptions are thrown. I find it quite hard to believe that no exceptions are thrown if say I tried to write to a full hard drive or attempted to convert an invalid date. I miss the compiler making up for crappy documentation. I guess I could go grep around the source and see if it actually is the case that no exceptions get thrown in both cases before making judgements. Then again, it would be scary if no exceptions are actually thrown in both cases.

Once I'm done refactoring my hastily written spaghetti and added the features I like I'll probably turn the app into an example for an Extreme XML column since it does use XML and that would make up for me being late to work. :)


Slashbot Xenophobia

Recent articles on Slashdot have played to the unemployed masses and played up the "Indian H1Bs stealing jobs out of the mouths of deserving Americans" threat with the expected demagoguery erupting. I can't help but chuckle at the hypocrisy of the Slashbots who justify copyright infringement on Kazaa and Napster by calling for the RIAA to learn to adapt the technology instead of trying to preserve a metaphorical horse-and-buggy business model in an automobile world who now want IT unions to protect them from the threat of a global marketplace.

This thread captures the spirit beautifully. Those who want to preserve their HTML jockey/Javascript code monkey jobs by any means necessary and those who point out that even if you could win the fight to keep foreign IT workers out of the US, the jobs would come to meet them.

Globalization in action.


Tools Will Save Us

Mark Pilgrim's brilliant series of posts about the Semantic Web have been interesting, educational and quite fun to read. I especially like learning about the Garden path model which helped crystalize some of my thoughts around the problems facing people working on AI and Natural Language Processing. Programming is about feeding simple, step by step instructions to a very stupid device that is very fast and performing this simple tasks repeatedly. The problem with AI and Natural Language processing is that we have no idea what the simple, step by step instructions for how to "read a sentence and understand what it means" because we don't know how we do it. Yes, I realize this is a "DUH moment".

I particularly like the scolding the guy who wrote that the complexity of RDF/XML doesn't matter since most people editing it will use tools or will be experts.

One of my co-workers likes pointing out the major two flaws behind the W3C's most maligned technology. The first is that half way through they decided not to create a type system but instead focused on making it a validation language and the second was thinking that the format could be ugly & obtuse as hell because "no one will ever write it by hand". Well, guess what? Most people still write it by hand because it so has been too complex for tools vendors to support in anything beyond a cursory manner.

The person being scolded failed to realize why his post was so ironic and deserving of the scolding he got from Mark. I can't believe he failed to realize that the point Mark was making was that if we (i.e. he) can't even create properly machine readable versions of documents for simple formats like HTML and RSS, yet believs that we will somehow magically do this with more complex and confusing technologies like RDF/XML, DAML+OIL, OWL, etc. <chuckle />


Get yourself a News Aggregator and subscribe to my RSSfeed

Disclaimer: The above comments do not represent the thoughts, intentions, plans or strategies of my employer. They are solely my opinion, thanks for playing.