Lists Home |
Date Index |
Bullard, Claude L (Len) <email@example.com> writes:
> Sounds like an excellent paper topic for one of the extreme
> conferences or a paid-for article at xml.com.
> I've crunched docs down into relational entities, and yes,
> it isn't impossible, but it isn't much fun. I'd rather do
> a straightforward relational design and map out to the document
> type if I can, keeping the mapping in the SQL and code.
> Joins don't scare me. :-)
> Oh... you want this to be a real time dynamic system and not
> a batch system?
He, he, yes, with sub-second response time.
Joins do eventually become an issue, a couple of weeks ago, in a note to
myself about optimization I wrote:
"The joins across the system for the constant resolution of type data
are killing us. [Big RDBMS vendors product] just can't keep the
precedence of the relationship structure straight WRT to the originating
query focus. Why should it? How should it know what we don't already
know? If we knew the relationships in advance the schema would be fixed
instead of dynamic!"
That (and some discussions on xml-dev) is partly what lead me to a
triples model in a relational database: we were joining for relationship
type instead of explicitly modeling them in the first place.
<snip>Danny's responses -- see my other reply to him</snip>