Jonathan Peterson wrote: > I often hear that databases are 'bad' because a > change to the taxonomy requires much work re-writing > SQL queries, whilst in XML, a change merely requires > updating a schema. I can't say agree. All data > processing makes assumptions about the data.
Well, I suppose "all data processing makes assumptions about the data" is true, but only in the most mind- numbing (and least useful) way. Rather, the I see the *real* issue is that Planet Earth is accustomed to processing data that is fully or partially structured, mostly because vendors offer products for that and that's what's familiar to users. By structuring data, you impose form, which constrains options for data representation, which makes parsing and processing easier (more bounded). It's the structuring of data that causes "re-work" to occur when you change your mind as to what the structure is supposed to be. It is also what causes problems when the beginning of a large data representation effort demands some content manager make guesses as to what the future holds (exactly what data will exist, and how that data will be processed). The ever-present "Databases V. XML: Battle To The Death" comes from the fact that (relational) databases are *highly* structured, and XML offers more degrees of freedom from *highly* to *partially* structured. If your data defies invariant description, XML has more wiggle room (but still requires definition of structure to a large degree). Yes, it is *entirely* possible to have "fully unstructured data". This places no assumptions as to the specific type, volume, or mix of content exists for a given "well-known" entity. However, there are few technologies that pursue this seriously (it's a hard problem). Nonetheless, there is recognized need in many markets, which is why relational databases are being wrapped with XML interfaces (getting more wiggle room, but it's still a bandaid at best). > It matters little whether you are re-writing your > SQL phrasebook or the XSLT sheets, either way you > are re-writing. Agreed, both require rewriting of processing logic, and probably import/export (or conversion) of existing data from the "old" form to the new. > My final thought is that much information doesn't > fit into either XML or relational systems, simply > because it does not fit into a rigid taxonomy. *Completely* agree. That's the "structured" v. "partially structured" v. "unstructured" fight. I predict the rise of more unstructured data technologies since this is a huge need with lots of user dollars behind it. Yes, HTML was kind of like that, and XHTML is kind of like that too. But, no cigar because of parse inefficiences, platform technology assumptions, and the way it's a terrible storage format. --charley [EMAIL PROTECTED] __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com -- http://cms-list.org/ more signal, less noise.