paulruizendaal wrote:

>
> The blr intermediate code in dynamic SQL processing could be eliminated
> without legacy issues and the blr for triggers and stored procs could
> be replaced by a new byte code (just like I had to do for PL/SQL).

Rewriting gbak, of course, and throwing out qli - or better yet
replacing it. And there are still preprocessed applications -
I'm maintaining a couple on Firebird 1.5. Just ripping out and
replacing all the DSQL intermediates is a non-trivial task.

>
> Debates around roadmap, copyrights, etc. would be harder to solve than
> the technical issues.

There I completely agree.

But beyond that there's the memory usage model - Netfrastructure
is far less frugal than Firebird. And recovery - Netfrastructure
has a log and requires a recovery process after a crash. I'm not
saying its impossible, just that it would be a longer process than
the Vulcan plus Vulcan merge, and that's already been most of two
years.

Regards,


Ann