At 07:56 PM 1/25/2003 +0100, Pavel Cisar wrote:

>
>It would be really cool to implement in-memory versioning in Firebird
>(3.0 ?), but that would mean to focus on SuperServer and discard Classic,
>right ? That would not please the Classic-centric wing of project :-) You
>must be a very happy men when you don't need to make such technical and
>political compromises.
>

Hey, if I thought there was any chance that Firebird could implement
this stuff I'd be a great deal more reticent about talking about it.

But seriously, I do have to make technical compromises all the time --
that's the stuff of engineering. But I don't have to convince anybody
but myself that I'm right. Which is good, because I'm a lousy
salesman.

The political compromises, however, are the Archilles heel of the
open source movement. About the only way a herd of ego driven
anarchists can agree is with an external standard, and the SQL
standard is a piece of [explicit deleted]. No leadership is possible,
and those people capable of innovation are stiffled. There isn't
the slightest doubt in my mind that I couldn't have developed
JRD/Interbase as part of the Firebird project. Cramming the concept of
blobs down DEC's throat was hard, but I got it done (the most
cogent argument was that they had never lost a sale because they
didn't have 'em). But getting the Firebird folks to pay attention
to unsolved problems of 2003 when they want to refight the battles
of 1990 is impossible.

So, yeah, I'm a happy man. I don't have to support MPE/XL,
SCO, MACs, and other unfortunate historical accidents. I
get to write squeeky clean code (mostly) devoid of conditional
compilation. I have build procedures that are nothing but
straight makefiles. I follow the SQL non-standard when
convenient so I don't have to argue about it, and extend it
any time the whim hits. Live free and code!

Jim Starkey