[long+short] autorouting patch status?

[Thomas Harding <tom thomas-harding name>]

Hello,

[I'm French, and English grammar (especially terse) and syntax are 
difficult to me while writing]

I wonder of "autorouting" patch status.

It has been posted here a while ago (3 years?) by someone as a part of 
his graduate thesis.

Feature was: shapes was automatically dispatched regularly while line 
and connection points
remained attached to them, then you could move the whole just by click 
and point one shape
-- and the batch were moved together then re-dispatch, or manually place 
one shape same
as Dia habits.

*[in short until end of post: this was a cool option and it would be 
nice if someone would fix it]*




I just played with it once or too, as it was not well optimized (memory, 
speed, response
time).

Such "autorouting" could be useful, especially when using SQL2dia 
(postgres2Dia ?)
or other code/data to Dia canvases programs: I sometimes used for 
SQL2dia, all objects
were same place and it has been a tedious task to dispatch the shapes 
(then do or redo
all connections, I can't remain if they has been already done by SQL2Dia).

Note I don't know how good SQL2Dia works currently (it was in early 
development).

There are recent posts here which points Dia is a software with a 
volunteers development
model [*]:

I counted in too, years ago, with my little skills, on very little parts 
(French translation
of the manual which nobody have improved yet (Debian Wheezy/Dia v0.97.2 
here), then doc
build tidying and some foreign docs integration to the manual). In fact 
dozens of hours.

I'm still very proud to have done that effort (especially in a phase of 
depression), and was
very surprised to have been especially granted for that in the (0.96?) 
release! ;).

So, here is just an idea:

Depending on what that patch already is and what it should or be planned 
to do, to integrate,
enhance and perfect it /could/ involve heavy mathematics and programming 
knowledge (and
I can ensure you I have not).

So, depending on needed levels, this would be possibly a good challenge 
to an  engineer to
graduate; a Master or PhD thesis part. Likely as a project team 
(eventually Internet
created and linked).

Creating such a team (supervision, design, Q/I, ...) from scratch and 
integrate it to Dia
development would be a good part of that challenge, as it could seems 
quite innovative to a
supervisor and a thesis committee (while it is an habit to /join/ one 
for internships) :P

In such case I just remain VideoLan started from and has been for long a 
scholar project
of via-ECP  (École Centrale de Paris/Central School of Paris), also the 
Cyrus project is still
hosted by Carnegie Melon University, and I expect some parts of Linux 
and other kernels
too (I knew for an engineer having that kind of thing for graduate 
project)...

That one, if even started by any interested[1], will obviously be 
largely smaller and short term.

Personally I don't really need that feature: I can spend for one hour or 
two to place manually
shapes next time.

It's just a cool thing I remember: I'm  currently authoring a training 
course for the LUG I belong to,
then used "distribute h/v" and... Hey! that "auto dispatch" menu could 
really help someone on
some tasks! :)


Best regards to all list readers,
TSFH.

[1] You, student, who already uses Dia or currently search for a good 
diagrams authoring program!

[*] In fact the guy who posted for selling copies searched for another 
program "Dia something":

     Nothing to do with diagrams, "dia" is a French prefix for ~"dual", 
although the word for diagram
     in French is "diagramme", and we have also "dualité" and 
"dia-publication".

     He would find first last (a proprietary one) program in any 
Internet search engine by typing
     the two words together :D