[Ktechlab-devel] Kde4

[Alan Grimes]

I took a look at porting to kde4 today and pretty much gave up in
frustration.

You see, there are certain parts of ktechlab that I've been studying
carefully for many years and can explain in pretty good detail including
problems that it solves and fails to solve.

However, there are other parts of the program that I've never even used,
code I've never looked at, and things that I just don't understand. =\

Software development these days really must be a team effort because the
 sheer volume of knowledge required to put together an application on
the scale, and scope of ktechlab is far beyond what any one programmer
has time for. What am I good for? thermionic valves, and little else. =P

A major issue is that there is a vast chasm between kde3 and kde4. They
are almost completely different systems. You can't just load ktechlab
into the new kdevelop because the build systems are completely different
and kdevelop 3 doesn't support the new one, and kdevelop 4 doesn't
support the old one!!! The howto said I needed am2cmake but I don't have
it. (maybe someone can patch svn and unilaterally upgrade it to cmake
for us).

Then we have two major chunks of code that we've stolen from other
projects:

canvas (stolen from some old version of qcanvas for the sake of having
custom geometry calculations that allow for negative coordinants.

Then we have stolen the MDI framework from some ancient version of kate
(see katemdi.h/cpp). I would imagine the state of the art in MDI's has
advanced considerably and that we should move to one of the newer
external frameworks, thereby improving our features and lightening our
codebase.

These are very sophisticated issues that would require the mastery of a
whole bookcase worth of material. =( I'm just not the man to do this. My
main programming interest is in AI, I am seriously not in the mood to
become a desktop applications expert just to get a vacuum tube
simulator. =(

I beg anyone who has done some kde4 work for ktechlab to spill it, mail
it to Zoltan or lainwir3d who is in the IRC channel in the afternoons
France time.

#### project of the day:
When saving icn documents (circuits and flowparts), normalize the
coordinants of the circuit by taking the average of the positions of all
parts and then subtracting it from all parts.
###

-- 
New president: Here we go again...
Chemistry.com: A total rip-off.
Powers are not rights.