4.6 release plan

[Steve Lime]

Yes, with GEOS I just need to write the polygon translator but that's
essentially the same as
what's done to write GML so it's just a matter of doing it. I submitted
a paper on this to force
the issue. Beyond that exposing GEOS processing via MapScript is
trivial. I've already got simple
line buffering and convex hulls working (e.g. $buffered_shape =
$shape->buffer($distance);).
A full integration will take a long time and may span several releases,
this is just the first step and
it doesn't affect the core code.

The GML3L0 writer works, as do the transformations. It's the XML
headers and stuff that will need
work, plus schema production.

Steve

>>> Daniel Morissette <dmorissette at DMSOLUTIONS.CA> 4/19/2005 7:41:48 PM
>>>
Steve Lime wrote:
> As long as you consider a partially implemented feature (like some
of
> the application specific schema stuff) as being legit to work on
post
> feature freeze, then I can live with this. I have two siginificant
> pieces going on: the GML3/application schema stuff and GEOS
integration.
> The latter is almost finished, but the former is going to take some
> time.  I'll port my changes for WFS GML to the WMS writer and
re-initiate
> some discussion on a couple of topics:
>
>   - controlling WFS output type (GML2/GML3/????)
>   - WFS schema creation
>   - changes to WFS queries brought about by finer control of GML
> output
>


Well, the idea of the feature freeze is that after that date nothing
should happen that would significantly break things so that we can
reach
our objective of fixing all critical bugs and having release quality
by
the planned release date.

I have entered those two items in the Wiki page.

The GEOS thing is the one that would scare me the most but you say
that
it's mostly done.

With respect to the GML thing, it affects only WFS/WMS so it's not too
bad if it's not finished yet, but would have to be completed in the
next
2-3 weeks or otherwise this part may not have been properly tested for
the release.

BTW, I have found that the 6 weeks release cycle with 3 betas has
worked
really well in the past to produce a quality release in time, but if
you
or others think that this is too long then we could discuss other
ideas.

Daniel
--
------------------------------------------------------------
  Daniel Morissette               dmorissette at dmsolutions.ca
  DM Solutions Group              http://www.dmsolutions.ca/
------------------------------------------------------------