[PD] ehu abstractions to be released

[Chris McCormick]

Yo,

On Tue, Dec 16, 2008 at 06:43:07PM -0500, Hans-Christoph Steiner wrote:
> There is a third option: people can release libraries like 'cyclone'  
> as libdirs.  Then you could just download the right one and drop it  
> into place to install it.

Cool, I will check out libdirs and see if I can package my
s-abstractions stuff like that. That said, I wouldn't be dropping a
binary version of [cyclone/prepend] into my system any time soon.
Libdirs don't solve what is a fundementally bad idea - using
[cyclone/prepend] in place of [list prepend].

> This reminds me of a similar discussion that happens in Java-land.   
> Lots of people still swear by Java 1.0.  Sure, you can do what you  
> need, but newer versions of Java are widespread and have more  
> shoulders of giants included to stand on.

I like to think that I'm not a luddite, and that I embrace progress as
much as any technophile, but to cast pd-extended in the role of "next
best thing to happen to Pd" is arguable. I for one definately don't see
pd-extended as the "newest version of Pd". I see it as an alternative
distribution which provides lots of extra functionality.

I am a big fan of Python's "batteries included" philosophy, which I
think is what you are advocating with pd-extended, but it should be
noted that the Python maintainers draw the line somewhere. They don't
compile every single possible library out there into a giant monolithic
release.

Since you mention Java let me frame my argument in terms of that
language. If I was writing some code in Java I would always prefer using
classes from the core of Java, over using 3rd party libraries, wherever
possible. I would do this because if I want to port my Java program to a
webserver, or embedded system, or mobile phone, I am much more likely to
be able to easily port it if I use the core Java language. Additionally,
if I release my source code, others will be able to use it far more
easily if its only dependency is the core of Java. The less dependencies
there are, the more robust my library becomes in terms of places where
it can be ported. To me this is the identical situation to what we've
been talking about. If users start using [cyclone/prepend] in their
patches instead of Pd's built-in [list prepend], then they immediately
make their patches less portable to other Pds. That is all I am saying:
favour Pd builtin's wherever possible so that your patches will be as
widely distributable as possible.

Sorry to bang on about this, but I am still arguing this point because I
want to make sure that your bad advice doesn't influence other Pd users
into making bad decisions and making everyone's life more difficult. I
am not trying to be harsh here, but rather I am trying to be honest
about my motivation for persisting with this thread.

Thanks again for all your hard work with pd-extended, though! Definately
appreciated (even though I don't use it, just like you don't use
RedHat).

Best,

Chris.

-------------------
http://mccormick.cx