This post originated from an RSS feed registered with Agile Buzz
by James Robertson.
Original Post: Code ain't unstructured data
Feed Title: Avi Bryant
Feed URL: http://smallthought.com/avi/?feed=rss2
Feed Description: HREF Considered Harmful
The flexibility afforded by text files outweigh the need for structured data stores. The fallacy one needs to recognize is that perfect a-priori domain modeling is a effort in futility.
and then
Eclipse is another good example, with appropiate use of fast intelligent parsers, manipulation of text (i.e. code) has never been so easy.
I'm with him in general - I'm a big fan of systems, from wikis to address books, that deal intelligently with loosely structured input. But those arguments simply don't apply to code. IDEs and source code databases don't impose any structure on code that the language semantics aren't already enforcing; things like method and class boundaries aren't emergent properties of the code, they have to be clearly and explicitly put there by the programmer. What magic increased flexibility am I getting by marking these boundaries with curly braces in a text file? Does Eclipse somehow infer more structure from the text file than VAJ did with its (far superior) method-level source code management? I don't think so.
Ok, that's all the troll feeding I can handle. Back to letting Patrick and James defend the faith.
Previous Topic
