We haven't really seriously considered it, no. One stumbling block is that Chimera really isn't completely open source yet. We distribute the source to the Python part of Chimera but not the C++ part. We plan on distributing the C++ part once we have organized it enough that outside developers would have a prayer of compiling it, but we haven't put in the work to do that yet -- always so many things to do! The other problem is that we need to track our downloads so that we can report those statistics to our funding agency (NIH). Perhaps fink allows for that, I don't know.

Eric Pettersen UCSF Computer Graphics Lab pett@cgl.ucsf.edu http://www.cgl.ucsf.edu

On Sep 13, 2004, at 1:29 PM, Ersatz Sophist wrote:

Hi,

Have you considered distributing chimera through fink?

Payam

"...who search the reason of things Are those who bring the most sorrow on themselves." --Euripides, The Medea

---

"...who search the reason of things Are those who bring the most sorrow on themselves." --Euripides, The Medea

---

Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://www.cgl.ucsf.edu/mailman/listinfo/chimera-users

On Sep 14, 2004, at 4:42 PM, Grant Jacobs wrote:

FWIW, I prefer the .dmg style installer, or at least single package, one click installers. Its also useful to have the application relocatable to any folder within /Applications.

Certainly if you're _not_ using fink, the .dmg is almost certainly what you want. I find it highly convenient too. If you're using fink, perhaps you'd like to be able to do everything through that mechanism, but as you point out sometimes even fink users prefer the .dmg install. And given the work we would need to do before we could even try to have Chimera available via fink, it won't be happening soon.

...

We haven't really seriously considered it, no. One stumbling block is that Chimera really isn't completely open source yet. We distribute the source to the Python part of Chimera but not the C++ part. We plan on distributing the C++ part once we have organized it enough that outside developers would have a prayer of compiling it,

:-)

For my money I'd be more concerned about good documentation of the APIs (etc) so that people can build on what already exists without having to decipher someone else's code base. Just my 2c - my pet peeve about open source is that poor or non-existant documentation forces everyone up the same time-wasting learning curves.

...

but we haven't put in the work to do that yet -- always so many things to do!

Yes, better documentation would be part of "organizing" the C++ part. Right now, the only comprehensive API documentation we have is the programmer's reference guide, which amounts to simply a dump of all APIs with some doc strings thrown in -- barely more useful that using 'dir' on Python objects of interest. The best documentation we have is the programmer's examples, which do a good job of explaining how to do various commonly-desired tasks in Chimera, such as getting your extension into the menus, using molecular data, and saving into sessions. Once you want to go beyond the realm of the examples, we trust that people will look at other Chimera extensions that do similar things to what they want to do and figure things out. It's the best we have until we get that "organizing" done. Of course, we do answer any development-related questions we get on chimera-dev@cgl.ucsf.edu.

The other thing about a documented API is that you can expect that it won't disappear from one release to the next -- at least not without a migration path being provided.

Eric Pettersen UCSF Computer Graphics Lab pett@cgl.ucsf.edu http://www.cgl.ucsf.edu