Re: Bidi-less Applications Patching Policy

[Behdad Esfahbod <behdad at cs dot toronto dot edu>]

On Tue, 1 Jun 2004, Mohammed Elzubeir wrote:

> On Tue, Jun 01, 2004 at 02:40:34PM +0100, Muhammad Alkarouri wrote:
> > > > if both would be 100% compliant and the minibidi license is more
> > > > permissible?!!!
> >
> > Well, they don't have to.
>
> They don't have to what? I realize this is a touchy issue and we have
> tried to resolve this in the best way possible. However, like I have
> said before, our options are to either do miniBidi or re-write those
> non-GPL friendly applications. Which one would you pick?

NO NO NO NO.  People here always confuse LGPL with GPL.  The way
to go is to convince people to use an external library
(compile-time optional, or even run-time optional).  So, you go
and give them an immediate option called miniBidi, so why should
they ever accept another option?  What you are missing is the
hassle to update all those embedded minibidi instances in
zillions of applications later.  Shared libraries have been born
for a reason...

> > > > Behdad, Ahmed, do you think we can go like this?
> > >
> > > We already are.
> >
> > Munzir, you know better. This is a 'core' level decision.
> > The magic word 'core' seems to be appearing more and more these days.
> > In another topic which I didn't want to comment about, something would have to
> > be revoked by core not QAC. There was no reason to comment after that.
>
> Oh that hurts ;) Come on now. These are two different issues and mixing
> them together is just not fair. QAC is there to set the guidelines and
> "assure quality".. not change Arabeyes general policies. That's what
> core is for. The issue you are referring to (in the 'doc' list) touches
> on the general Arabeyes policy.
>
> Now, as for the 'core' level decision as you put it -- it was. However,
> no core decision is engraved in stone if the majority (or even a
> size-able number of people) disagree with it. You know better than that.
> Our decision to forge ahead with miniBiDi was only questioned by
> yourself and Munzir. No one else has said anything about it. The core
> team (4 members) + miniBidi's author (1 member) want to go ahead with
> the decision where as 2 do not.

You know my vote of course.  But I'm not in Arabeyes really.  And
of course Roozbeh just copies mine, if that matters.

> All I have heard so far is discontent with the resolution, yet I do not
> see any alternative plans. Please, if you do have a better plan, share!
> If you think it would help, we could setup a meeting on irc where all
> concerned parties can show up and throw ideas around till we find
> something we are all comfortable with.

I'm up to clear my relation to Arabeyes sometime next week.
It's more about the policies.  Mine is to contribute to Free
Software (as defined by http://www.fsf.org/), while Arabeyes' is
apparently not.

> > I am thinking of redoing unicode coding (already bought my Unicode 4 standard
> > book:) in order to learn. Naturally, this will be the umpteenth implementation.
> > As they say, the best things in standards is that there are so many to choose
> > from;)
> >
> > I expect it to take 18 months to complete, so make use of what is available for
> > the time being. Targets are the small devices, and I may make a Python module.
>
> Cool ;)
>
> > It will be under the MIT license (http://www.x.org/Downloads_terms.html) which
> > by the way includes the sentence:
>
> Even cooler ;) You know I love MIT/BSD style licenses ;)

That's the root of our disagreement ;).

> > [..]
> >
> > While no plan currently exists for the unification with FriBiDi (license
> > issues) or miniBiDi (I want to learn on my own), future coordination is
> > welcome. In particular, a unified interface for all of them will be for the
> > benefit of all (well, not all).
[snip]
>
> Regards,

Cheers,
--behdad
  behdad.org