Friday, December 16, 2005
December - January DUIM tasks
Spurred on a little by the activity on the gardener's list (see previous post for link) I've decided to juggle my plan of work with regards the DUIM port a little. I had planned to get a decent amount of unit test coverage in before starting work on any back ends; now I think I will follow a different schedule:-
1. Transliterate the vanilla port code and fix up the existing unit tests
2. Write a simple back end (e.g. not necessarily fully functional) -- chances are I'll use Peter Herth's LTK library for this
3. Write unit / regression tests for anything that's obviously broken during writing the back end
4. Put together a 'first release' (maybe register the project on common-lisp.net)
5. Maybe write a simple UI on top of RT (there's one already for McCLIM that I might modify, or a new one) and continue with unit test coverage.
6. Rewrite the API to be more Lisp-like (copying the CLIM API to a great extent for this)
7. Put together an advertised release
8. Eat my own dog food ;-)
1. Transliterate the vanilla port code and fix up the existing unit tests
2. Write a simple back end (e.g. not necessarily fully functional) -- chances are I'll use Peter Herth's LTK library for this
3. Write unit / regression tests for anything that's obviously broken during writing the back end
4. Put together a 'first release' (maybe register the project on common-lisp.net)
5. Maybe write a simple UI on top of RT (there's one already for McCLIM that I might modify, or a new one) and continue with unit test coverage.
6. Rewrite the API to be more Lisp-like (copying the CLIM API to a great extent for this)
7. Put together an advertised release
8. Eat my own dog food ;-)
Thursday, December 15, 2005
Useful DUIM links
I mentioned on the cl-gardeners list that I'd provide a couple of DUIM links here today (there's quite an active discussion going on there regarding GUIs at the moment) so here they are:
Functional Object's DUIM Library Reference
Functional Object's DUIM Tutorial
Functional Object's Dylan Extensions Reference (pretty useful if you want to know what some of the stuff in the DUIM code was intended to do)
cll post from SWM that made me say "why not do just that?". That whole thread is pretty interesting I think...
cll post from SWM regarding the problems wrt implementing native look and feel in CLIM.
cll post from SWM hypothesising about UI design in general.
Gwydion Dylan homepage; the original DUIM code is in this repository, along with the Functional Object's Dylan IDE.
The FunO code is released under the LGPL which is handy.
Functional Object's DUIM Library Reference
Functional Object's DUIM Tutorial
Functional Object's Dylan Extensions Reference (pretty useful if you want to know what some of the stuff in the DUIM code was intended to do)
cll post from SWM that made me say "why not do just that?". That whole thread is pretty interesting I think...
cll post from SWM regarding the problems wrt implementing native look and feel in CLIM.
cll post from SWM hypothesising about UI design in general.
Gwydion Dylan homepage; the original DUIM code is in this repository, along with the Functional Object's Dylan IDE.
The FunO code is released under the LGPL which is handy.
Saturday, December 10, 2005
DUIM code conversion 'complete'
Subtitle: completeness is in the eye of the beholder :-)
I'm pretty much done converting over the Dylan code to Common Lisp. I've revisited the macros and think that pretty much all of the core code is converted (with one exeption). I've decided to use RT as the testing framework.
The next steps are to write a bunch of tests, and to transliterate what's left of the code into Lisp (all that remains of the core DUIM code is multi-processing related).
Then it's on to providing a back end, writing documentation and putting together an alpha release. I don't like to make timescale estimates, but I'm hoping I will be able to start on a back end in earnest by the end of January. We shall see...
I'm pretty much done converting over the Dylan code to Common Lisp. I've revisited the macros and think that pretty much all of the core code is converted (with one exeption). I've decided to use RT as the testing framework.
The next steps are to write a bunch of tests, and to transliterate what's left of the code into Lisp (all that remains of the core DUIM code is multi-processing related).
Then it's on to providing a back end, writing documentation and putting together an alpha release. I don't like to make timescale estimates, but I'm hoping I will be able to start on a back end in earnest by the end of January. We shall see...
