Re: Development Trends in Web and Oracle

Hexathioorthooxalate apparently said,on my timestamp of 13/03/2005 8:04 PM:

> it is just the data in and data out can be a bit more verbose. But you
> cannot keep dismissing all the other benefits of XML, the other benefits I
> have reiterated over and over again in this thread. Over and above
> everything, an XML schema offers a very easy and efficient mechanism to
> validate data, and validate it one place, the XML schema.

What these fools consistently refuse to acknowledge is that a relational database is a mechanism equally easy and a lot more efficient at validating data, also in one place: the dictionary or schema.

Mostly because they have NO CLUE what a database does. The clincher is when they mention the word "persistence": 100% assurance the punter thinks a database is a flat file and has no clue of the efficiency of a relational database.

> can be in or out of the database, or both. Unlike the relational model,

Whoop-tee-doo. Like: there is a BIG advantage in this...

> dealing with XML doesn't force the database to perform data validity checks
> over and over again via stored procedures, triggers etc before, for example
> inserting a new row into a table. This is the overhead you won't accept as
> part of your "overhead" argument.

As usual: no clue. The validation is NOT done with triggers. Only in substandard "databases" does that happen. And it is as efficient and safe as can be: if the validation is associated with the storage of the data, there is NO WAY that it can be subverted. As soon as it is moved somewhere else under false pretense of "efficiency", subversion is the result. Like I said: these folks have NO CLUE why databases do what they do and WHY they evolved to do precisely that.

-- 
Cheers
Nuno Souto
in sunny Sydney, Australia
wizofoz2k_at_yahoo.com.au.nospam