Re: [Libxmlplusplus-general] resume of one day's work

[Stefan Seefeld <seefeld sympatico ca>]

Murray Cumming wrote:

> We have stated several times on this list that we want to aim for a
> stable 1.0 API for the sake of stability while continuing with an
> unstable API. Unfortunately the archives might not be reliable, but I'm
> fairly sure that we've mentioned this in emails to you.

yeah, the sf mailing list archives are totally buggy...:-(

Indeed, I'm aware that you are aiming at a stable release. But I'm
not aware of any concrete time frame (a week ? a month ?). I'm also
aware that, while I'm obviously very confident concerning my own
changes, others may just see the number of changes and panic,
anticipating a lot of work to stabilize all that.

We should create a little unit test suite to formalize these tests,
so 'stable' becomes a more measurable entity.

[...]

> It's OK to be wrong. We've sorted it out. But if I need to spend time
> explaining the process to you repeatedly then I have less time to
> actually consider your technical suggestions.

fair enough. Having a document (online or included in cvs) discussing
this kind of stuff would certainly help. Notably with a more measurable
timeframe than 'soon'. In fact, I was totally fine with all changes to
go into the development branch, but at some point you or Christoph
suggested it may go into 1.0 if it turned out to be managable. That
may have been a bit ambiguous.


> We want your improvements but they must be discussed and managed. That's
> how good APIs are created. If you feel that you _must_ have full
> unlimited power over the project then you are free to fork. I hope that
> you will choose to work with us instead.

If you think I'm looking for 'unlimited power' you are wrong. I'm not.
I'm enjoying to collaborate, and my only worries are that the
development process is as efficient as possible.

Stefan