I'd agree.  The first release should be a mechanism to get feedback
from a wider audience that doesn't want to build from CVS.  Heck, I've
worked on the build, but I've barely tried out the API yet.  And
probably won't have a chance to finish my "playing with it" phase for
a month...

Another requirements for a release: a bug-tracking system must be
available & setup (categories/modules/versions defined, etc.) to
receive & track the feedback...

-Stephen


On 6/20/06, Elias Torres <[EMAIL PROTECTED]> wrote:


Garrett Rooney wrote:
> On 6/20/06, James M Snell <[EMAIL PROTECTED]> wrote:
>>
>>
>> Garrett Rooney wrote:
>> > On 6/20/06, James M Snell <[EMAIL PROTECTED]> wrote:
>> >> I'd say the goals for M1 are pretty simple...
>> >
>> >> produce a stable, reasonably functional impl for folks to start
>> working
>> >> with.  Where "reasonbly functional" means it implements the RFC4287
>> with
>> >> no obvious gaps.
>> >
>> > That makes sense to me, although one question to think about here is
>> > how stable the parser/object model API will be at this point.
>> >
>>
>> stable == the svn will be tagged at a specific point.
>
> Well, obviously, any release gets tagged at a specific point, what I'm
> getting at is are we going to say "these APIs will only evolve in a
> backwards compatible manner going forward", or is this a pre-1.0
> liable to change, use at your own risk kind of release.

I'd rather this be a pre-1.0. We need more feedback from the community
and road tests for the API.

-Elias



--
Stephen Duncan Jr
www.stephenduncanjr.com

Reply via email to