J bz said :
> "So who packages pd for pure:dyne then? I'd be curious to know the reasons
> for not using pd-extended instead of pd... "

Hurray! Ye good ol' vanilla vs extended thread ;)

It's true that when Puredyne started, extended was very young (someone
would have to check but I also think that it was first introduced for
win and macos and was only made available to Linux at a much later
stage).

Next to that there are some personal choices in using vanilla and a
reduced set of externals:

- vanilla is the reference -> decreases chances to shoot yourself in the
  foot by making patches which features relies on a specific
distribution of Pd. For example recently hc announced that he would stop
maintaining some externals in extended, which I understand perfectly
given the insane task it must be to distribute everything related to Pd
in a single convenient package for 3 different platforms.
http://puredata.info/docs/LibrariesInPdExtended/

- less is more -> we prefer working with fewer sets of externals that
  have been around for a while and well-tested. It's a choice that
influences creativity as it is a constrained process that forces us to
really explore what Pd and a few "low level" externals have to offer
instead of approaching patching from the "one feature == one external"
mindset. It makes maintaining patches easier also, etc.

- we package only stuff that we use -> so we have a direct interest in
  having it working well, which benefits to the user ultimately. Instead
of providing everything Pd related and not have time to make sure
everything works.

- we package everything in a modular way to distribute the workload and
  maintain components individually (useful if version 2 of
[cheeseburger~] is broken but version 1 is bug free which is a more
difficult thing to solve in a monolithic distribution such as extended).

Of course all of these points are personal choices, we don't hold the
truth in that regard, but it is just what works for us, both as artists
and Puredyne developers, and it is the type of "philosophy" that made
Puredyne a reliable production environment over the past years. YMMV ;)

That said, there have been some attempts in the past to coordinate
efforts and convince other Pd dev to join Puredyne Pd packaging in a
modular approach on Debian based systems.
http://lists.puredata.info/pipermail/pd-dev/2009-04/013415.html

Sadly enough it did not happen, despite our call. I can't really explain
why. We were too early? Both side were a bit territorial in the way
things should be led? Does not matter. Everyone moves on.

The good news though is that a few Pd dev have very recently started to
package externals as part of the Debian multimedia team community in
such a modular way and hopefully it will allow the user choose what
she/he wants on her/his system while providing higher quality binaries
of externals. Everyone will greatly benefit from it and will make our
isolated work redundant, which is a good thing as duplicating efforts is
just nonsense in such small communities.

Concerning the documentation side of Pd extended, I can't really tell as
most of the Pd documentation needs of Puredyne were driven by GOTO10
workshops (mostly around 2006-2008) and we had our own tutorials for
that. For instance: https://code.goto10.org/hg/index.cgi/canvas/file/

About the interface changes, this is my *own* opinion: I don't care
about fancy looking objects, custom fonts and all kind of visual
improvements. I remember Frank Barknecht, years ago, saying that it
forces you to patch well and not dive into sloppy spaghetti patching,
relying on visual feedback to keep track of what is going on. I stand by
this argument big time. For those who wanted how to play with the
interface it was also covered by checking the different hacks from
Carmen.

Of course there are some improvements to be made though on the keyboard
side, I love what Chun and Matju did on the dd branch, a long time ago,
to allow complete keyboard driven patching. I hope that Yvan Volochine
can port some of this stuff in his plugin, this is a *real* feature, and
would be a shame to have this work lost. Unfortunately it is unlikely
that we provide 0.43 for Gazpacho, so no plugin for now (to reply to
another mail).


> I think it's Claude Heiland-Allen and Chun Lee.

Yes, initially it was all of us, but as the development of Puredyne
increased in complexity, things got more specialized and at the moment
Claude and Chun are doing all the work.

a.
--
http://su.kuri.mu

---
[email protected]
http://identi.ca/group/puredyne
irc://irc.goto10.org/puredyne

Reply via email to