On Friday 06 May 2005 12:19, Dave Rolsky wrote: > On Fri, 6 May 2005, Simon Perreault wrote: > > I don't agree with that. What's wrong with the correctness of data using > > XML? XML is a way to enforce correctness in data. Sure, a DBMS also > > enforce correctness in data, but if that's all you need then you'll > > probably be better off using XML. It provides everything you said, plus > > it's easier to work with. > > Err, no.
Yes, really! ;) > You are simply wrong. Do a little bit of research into the history of > the relational model. Read Codd's original papers. I think you're emphasizing the R part of RDBMS while I'm emphasizing the DB part. But you still haven't told me why the DB server should be executing the recurrence function instead of the application server. > Because if you want the fastest data store possible, and you don't mind > biasing it for a specific set of queries, you can probably hand-code a > faster data store yourself. The whole point of a SQL DBMS is that you > can design your model so that it allows for as-yet-unforeseen queries. Again, I'm telling you that the DBMS is optimized for quick retrieval in big data sets, not for executing code. That's all I'm saying. I'm not saying a DBMS is the fastest way to query data, just that it shouldn't be encumbered by recurrence function executing. Anyway, you do as you wish, obviously. My DB stores and retrieves data, and does not execute application code. SELECT answer FROM generate_answer_to_the_question; +----+ | 42 | +----+ 1 row in set (408949123019587 secs)
