[OpenLayers-Dev] Are there enough OpenLayers committers? (Was Re: Different lon and lat map resolution (and other issues))

Schuyler Erle sderle at metacarta.com
Sun Nov 4 16:22:54 EST 2007


* On  2-Nov-2007 at  2:40PM PST, Jachym Cepicky said:
> 
> thanks for this comprehensive explanation. I would like only add this
> small comment: In GRASS, more then 30 developers do have cvs write
> access. It works. People are respecting each other's work, if somebody
> breaks something, the change can be reverted (but this does not happen
> too often). It just speeds up the development.

Jachym, I agree with you completely in principle. I have always been
in favor of allowing more people to commit to trunk. However,
OpenLayers differs from GRASS significantly in two subtle ways that
are not obvious but have served to slow the extension of trunk commit
access:

(1) MetaCarta, the copyright owner of OpenLayers (and by far its
biggest supporter in terms of resources committed to the project), is
also the holder of some patents related to geographic software. The
PSC has tried to be sensitive to MetaCarta's concerns about infringing
their patents, to retain their ongoing support, and so in turn we have
had to ensure that people who could commit to trunk understood the
issues involved. This point is no longer a problem since the change to
a new BSD license that explicitly excludes grant of *anyone's* patent
rights. However, it did make us conservative about granting trunk
commit access.

(2) Where GRASS is an application, OpenLayers is an API, and is
designed to have other applications *built* on it. This means that
anything that goes into trunk has to be supported for the remainder of
the major version of OpenLayers thereafter -- this is our stated
commitment to help reassure developers that we will not capriciously
break their existing code on the next upgrade. (Though it does happen
that some users' code has broken, typically it has been because they
relied on some internal component to OpenLayers, rather than the
published API.) The result is that any time we make a new release of
OL with something in it that sucks in some way, we are stuck with it
that way until version 3. As such, we need to be more careful about
what goes into trunk that the GRASS developers do, I think.

That said, I think we would still benefit by lowering the bar to
inclusion.

SDE



More information about the Dev mailing list