Hi everyone, my name's Lorelei Kelly, and I'm interested in applying for a GIMP summer of code project.

I'm a senior computer science and psychology double major at Grinnell College, and I plan to enter a masters in human-computer interaction program in the fall (in the midst of determining which program right now). The projects I'm most interested in are creating a unified UI for scripting and improving python scripting. I'm also interested in making GIMP resources taggable, but I know less about what might be involved in such a project.

Last summer I participated in a project for the Grinnell CS department involving the GIMP. The department is converting the introductory curriculum (taught in Scheme using DrScheme) to a media computation approach, and they wanted to take advantage of the GIMP's Scheme scripting to teach media computation in Scheme, but were unhappy with the script-fu console interface. My group set up a system that allows DrScheme to interact with the GIMP via the script-fu server, so students write code using the DrScheme interface, which gets executed on the script-fu server. For the past two semesters, this system has been used by the intro CS classes. Most of our work was done in PLT Scheme and MrEd, a Scheme-based GUI toolkit, but we also spent some time learning about the internals of the script-fu server, and working with tinyscheme and the scheme procedural database.

I've worked mostly in Scheme for the past year or so, but I've also had two semesters of C and three of Java, and I've been learning Python and GTK+.

If anyone is interested in mentoring me for one of these projects, let me know, and I'll answer any questions. For my part I'd like to know more about what I might need to know or do in preparation for these projects, and what expectations you have for my application.

Thanks, Lorelei

Thanks for the response. I'd certainly like to see our work, which we called DrFu, become more widely used, but I'm not sure what it would take. Right now setting it up involves having both the GIMP and DrScheme installed separately, getting our files into the right directories, and then using a script that opens both programs and gets them communicating.

My first thought is that MrEd would need to be distributed with the Gimp (from what I can tell, it's currently only available bundled with DrScheme) and then I could write an interface specific to this system rather than relying on the existing DrScheme program, which does a lot we don't need. That would fit with the work I've already done, as creating a new interface is something I looked into last summer but didn't have time to finish. I'd then want to convert the system from relying on the script-fu server to directly interacting with the procedure database. The server was pretty clearly not designed to do some of the things we want from it, and I think this would help make it less fragile.

The concern I have with this plan is I don't know how reasonable it would be to add MrEd as a dependency. Maybe there are other strategies?

Lorelei

On Sat, Mar 22, 2008 at 5:05 PM, Bill Skaggs wrote:

Lorelei wrote:

Hi everyone, my name's Lorelei Kelly, and I'm interested in applying for

a

GIMP summer of code project. [...]

Hi Lorelei, thanks for the information. The ideas on the Wiki page are only
suggestions, and you shouldn't take them as the only possibilities. My immediate thought is that, from a purely selfish gimp-o-centric point of view, the big win for us would be to leverage your impressive scheme background, rather than getting you into a project where most of it would be new and you'd have a hard time completing it. I wonder whether there is any possibility of doing something like packaging the interface your group developed to make it broadly available?

Anyway, Kevin is our Scheme expert, and Yosh is our Python expert, so I leave it to them to respond more specifically.

(If you are interested in batting around ideas, I invite you to visit the gimp irc channel, #gimp on irc.gimp.org.)

-- Bill