[jdev] Jabber Certification Program
temas at box5.net
Fri Jun 18 10:04:40 CDT 2004
On Jun 18, 2004, at 5:21 AM, Justin Karneges wrote:
> I may have had a cynical tone in my post, but I wasn't rehashing an
> Standards, procedures, and policies are important. Maybe I wasn't
> happy with
> what happened back then, but if you re-read my text above you'll see
> that I'm
> actually "on your side" now.
> - Don't implement Experimental JEPs.
> - Do implement Draft JEPs.
> - Too bad if you implemented something that got superceded.
> Those were the words of the council, and these are my words now.
> Rachel was suggesting that the certification program might want to
> experimental JEPs, and I'm just trying to explain that this is a bad
> I can't imagine the JSF have a certification program involving
> JEPs, when you consider the nice red warning text that is currently at
> top of all of them. Left hand, meet right hand?
"my side" does not state don't implement experimental JEPs. It
advocates being responsible if you choose to do so, but implementing
should be a fundamental part of the JEP development process, as I
stated in the email, and we'll get to again in a sec.
Perhaps non draft JEPs are not fully certified, but I don't see why
they couldn't be ear marked into a certain certification level. This
would cause people to look at impls or start them, and then clean up
the JEP more quickly. Once it's approved it could become an official
part of the certification.
> Maybe times have changed, because back then you (read: council) made a
> deal about not implementing Experimental JEPs. I mean it was a really
> deal, like a "don't even touch it" kind of thing. But it has been
> awhile, so
> who knows anymore... but I think this led to the adding of the red
> text in the JEPs.
> But heck, by all means, change this policy if you want. To me, it
> matter, as long as we are all on the same page. However, if the new
> is that implementations of Experimental JEPs are to be encouraged,
> then the
> red text probably shouldn't contain "not recommended" (which, to put
> in perspective, is also in the Deferred and Retracted text).
OK, I just reread all the appropriate threads to make sure my
statements aren't off base. First, it wasn't a don't touch it kind of
thing. It's pretty much always been, don't put it in a fully
production system or something that you are unwilling to change. The
red warning on the JEPs even clearly allows for impls in an
"exploratory fashion". We don't need that debate again though, we just
need some clear rules about how they fit into the design of a
certification process, perhaps my idea above works?
As to policy changing, I would say it's loosened ever so slightly, but
is still the same as it was back then.
p.s. - Maybe I'm misreading your tone in this email, where it seems you
are for the implementation of experimental JEPs on a wider scale, back
then you had the "tough cookies" sentiment ;-)
More information about the JDev