Hello, Sean!

Tuesday, October 5, 2004, 3:42:07 AM, you wrote:

LS> But I need to point out: The problem you describe is with the API/wire
LS> protocol! Not with the manner in which the data is stored in the
LS> database!!

LS> If a new wire-protocol allowed for the return of Blobs with the row data
LS> (perhaps to a defined size) then there would be no problem with the
LS> current storage approach.

This is nothing with storage approach. I'm talking about blob separate
from record storage, Roman tells about reading blobs stored with
record. This is completely different. First needs to add one tiny
flag to header page to allow ajustment of server performance, second
needs to rethink or rewrite API or extend data types with new type.

The only responces against my suggestion are "it works OK for me"
or "it was coded OK". Yes, indeed in both cases. But production
databases are not the same, they have their own specific.
And big thanks to Jim that he coded blob storage that way, now
we can just add flags to maintain this behavior manually.

LS> The big negative for a new type is the inevitable user confusion -- what
LS> is the difference between BLOB and BINARY???

Currenlty not all knows difference between char and varchar :-)

--
Dmitri Kouzmenko