|
From: Uche Ogbuji [mailto:uche@ogbuji.net]
------------------- I hold the complete opposite opinion. I know of no other existing technology (primarily including the entire ecosystem and stack) which I can know ahead of time has a very high chance of being able to reasonable
encode, transmit, use, and reuse *almost any kind of data* reasonably well. Given a specific datum, I can come up with a better representation. Given a set of known (or more often unknown) set of data I know of no better technology to use that I can count on to make work. To my mind, XML is a near perfect "hub" or "bucket" where I can throw nearly anything into it and use the same tools to manage, query, store, retrieve, transform and otherwise muck with. There are a few outliers
such as "pure binary" where its pointless to try to munge it into XML ... and perhaps "pure text" where its very slightly uglier to add a <doc> wrapper then to not. But other than that, there is no other single technolgy+ecosystem today that I know of and can count on to make work no matter what I throw at it. Others of course have different views, but mine is *exactly contrary*
to XML being a "bad, bad idea" for "pure data" as a general rule. -David |