Dear Lachlan,
Thanks for this note. It might be worth looking at (again). This stuff
appears to be Tcl/Tk (tickle!). My experience with it is very limited but
it does look attractive. Brian Toby has written a GSAS front end GUI for
unix in Tcl/Tk which looks a little like my "PC-GSAS" front end for Win95.
However, in trying to port it to MS Windows I found that it doesn't
interact with "DOS" programs very well. Since all of GSAS consists of DOS
executables with a DOS memory extender it doesn't work - it can't start any
GSAS routine. You might be interested that the PGPLOT graphics package in
GSAS can also support Tcl/Tk in some way; I havn't explored that at all.
Bob
At 02:48 PM 2/15/98 +0000, Alan wrote:
>Modern programs tend to use command-line input - questions
>and answers, with the next question often depending on the previous
>answer. This makes it much more difficult to construct a graphic
>user interface, especially if you want it to be system independent.
>(because you cannot readily seperate the interface from the rest).
>
Alan then listed some possible GUI languages. I'd add Java to that list of
"structured" languages. It looks a lot like C++ but it is far more
consistent and a lot less dangerous to use (no pointers!). I have explored
Java a bit (even wrote the GSAS space group interpreter in Java) to see how
it works. Not bad, but I have some questions about "serious" math
performance. Can anyone imagine GENLES in Java? The other thing that has
put me off so far is that Java is evolving very rapidly as the big software
companies take it up and make extensions for it. So stability of the
language is an issue.
I'd add one other point and would like some feed back on this. My problem
with making a GUI for GSAS is finding a way to "better" the current menu
system in GSAS with its error checking, type ahead & help lists. Any ideas?
Bob Von Dreele