RE: ACTION: elaborate on 4.4

I'm not entirely clear as to the intent of the discussion around the
HTTP "Accept:" header.

I've gone on record claiming that it makes a lot of sense to try to keep
the query language and the protocol as independent as possible, and this
thread seems to be going in just the other direction. It's one thing to
say that a query language will (logically) generate some results in a
well-defined format, and then a protocol may transform those results in
some way (like applying compression), but it's quite another to consider
some kind of format specifier being passed through the protocol into the
query language proper. Even if protocol and language are fully separate
standards, there is nothing which prevents this being done as an
optimization, but it seems bizarre to be requiring such functionality in
query processors.

Received on Thursday, 24 June 2004 12:46:17 UTC