> >
> >I think the confusion here is that the old Superserver vs Classic
> >document listed Classic as experimental on Windows. From my
> >understanding, it is no longer considered experimental.
>
> Au contraire, it appears to have more to do with people (notably,

native

> English speakers) believing that "experimental" means "unstable".

I don't see experimental as unstable. Experimental is synonymous with
unproven and if you read it in that way, it is inappropriate choice
for a production environment. I am pleased to see that the document
has been changed to reflect that no-one really considers it to be
experimental anymore. Obviously those who read that document a few
months back may from time to time make off the cuff questions about
the stability considering it is experimental.

>When Fb
> 1.5 was released, Classic was stable. If it were not, it wouldn't

have

> been in a release.
>
> However, with nothing useful from field-testers regarding its

performance,

> etc. on Windows, with real-life loads, it was released marked
> "experimental". Two years and three sub-releases later, with lots

of high

> volume deployments, it's probably even past "experimental" now. :-

))

Definately agree. I would really have to explain myself to my
employer if I deployed an "experimental" RDBMS to a live/production
database. Perhaps something like "not yet benchmarked" would have
better captured the gist of the release.

Adam