[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: pngcrush and bugs 331068, 338659, 355693, 352177, 330026


On 3/25/06, Thijs Kinkhorst <kink@squirrelmail.org> wrote:
> On Fri, 2006-03-24 at 18:20 -0300, Nelson A. de Oliveira wrote:
> > I still don't see good reasons to merge the 4 programs :-/
> It's of course always a tradeoff between the extra work of merging
> things and the possible benefits. I can see two major benefits:

Extra work for me is not the problem. I was thinking about the extra
work that the upstream authors will have.

> If you can convince upstream to join forces and integrate the work into
> one tool, this is an improvement for the quality of the code. In my
> opinion this is the best way to go: there's only one copy of the
> infrastructure (reading files, user interaction, error handling and
> recovery...) with more eyes on it, and only the algorithms differ. This
> should improve the quality of the general program.

I can try to do this.
I will contact the authors.

> Another major benefit is for the user: they want a program to 'crush'
> pngs. Speaking for myself, when I was looking for a program to do that
> task, I wouln't want to install four different Debian packages and
> evaluate them. Each of the programs would probably have different
> options and a different modus operandi. I would want to install just one
> package with either four programs I can try, or better, one program that
> has a switch for the algorithm to use. It has the additional benefit of
> having one Debian package to maintain.


> I can't imagine that the size overhead would be that big to be
> significant. If you can convince upstream to go along, you would satisfy
> both benefits, but if you don't, I still think the second benefit is
> worthwhile enough to implement it.

Just in the case that none of the 4 authors want to merge the
programs. We will create one Debian package, containing the 4
programs, with a wrapper (a "pngoptimizer", for example), that will
select one of the programs, based on the options that the user

Best regards,

Reply to: