Is it a library like OpenCascade/OpenDesign?
Or is it a application like AutoCAD/MicroStation?
Or is it a solid modeler like Inventor?
I like to see it to become a library that every one can use to build
its own application on, from simple 2D sketching to 3D modeling.
How do other members see this project?
What are the skills of the individual members and what role do they
want?
Gertwin
For a general-purpose-cadTool would be necessary:
- lib's for: import of cad-formats (change into native format)
- lib for: change native fomat into native objects
- lib for manipulating the native objects (geom-libs)
- lib for storing the native objects (database)
- libs' for exporting to other cad-formats (the objects or the native format)
- lib for viewing the native objects (OpenGL-driver)
- a cad-interface (the application)
- a plugin-interface for using all the libs/functions above.
All this libs must fit together, but they should also be independent to use.
The cad-native-format is one part, but internally the progams work with binary
objects; eg. a point has three floting-point values; but the decription of a
line or a circular arc is different in most systems/libraries.
And here comes the more difficult part to define the native-objects.
For the collaboration of programmers they need a description of the objects to
create/manipulate. I do not know how to describe this objects without a
languange-binding since we always use Ansi-C-format.
Many people use java (collabcad, gvsig ..) many use C++ (opencascade) ..
What do programmers out there prefer ?
ad:
> German is a difficult language for us,
> so post this document if you want.
The complete documentation and all object-specifications and descriptions are
included in the devel-package of gcad.
Franz
How does this relate to existing projects? brlcad is e.g. quite exactly
what you describe, such a set of more or less independent libs and
tools. They lack a cad-interface that one could really call general
purpose, and opengl is not supported as well as needed at the moment
(would need some meshing work afaik to generate triangles from csg). Do
you plan to include existing libs, or do you want to rewrite?
CU Lars.
I think the most important it´s get a file and have the posibility to
work in Catia-Autocad-Microstation with this file, but in a future we
can share code with opencascade and BRL-CAD people to get a complete
platform. If people work together is easy and people like our format
before DWG/DXF.
the great final challenge is get a Catia and a Autocad... but this is
in a long way...
Opencascade is a company and it´s difficult but BRL-CAD is an open
group. we can talk with this people in their forum and Sean Morrison
is inside this list although he don´t talk by now.
I can help in surveying and promoting the iniciative in Spanish
countries (spain, Argentina, Venezuela, ecuador, cuba,
uruguay,chile...)
In venezuela is obligatory use GNU/linux in 2 years...
I want free software to surveying and i need a CAD,... I don´t want
build a CAD but it´s neccesary.
regards.
On Oct 19, 9:30 am, Franz Reiter <franz.rei...@cadcam.co.at> wrote:
>
> For a general-purpose-cadTool would be necessary:
> - lib's for: import of cad-formats (change into native format)
> - lib for: change native fomat into native objects
> - lib for manipulating the native objects (geom-libs)
> - lib for storing the native objects (database)
> - libs' for exporting to other cad-formats (the objects or the native format)
> - lib for viewing the native objects (OpenGL-driver)
> - a cad-interface (the application)
> - a plugin-interface for using all the libs/functions above.
> All this libs must fit together, but they should also be independent to use.
>
> The cad-native-format is one part, but internally the progams work with binary
> objects; eg. a point has three floting-point values; but the decription of a
> line or a circular arc is different in most systems/libraries.
>
this is great Franz.
On Oct 19, 11:09 am, "Lars O. Grobe" <gr...@gmx.net> wrote:
> How does this relate to existing projects? brlcad is e.g. quite exactly
> what you describe, such a set of more or less independent libs and
> tools. They lack a cad-interface that one could really call general
> purpose, and opengl is not supported as well as needed at the moment
> (would need some meshing work afaik to generate triangles from csg). Do
> you plan to include existing libs, or do you want to rewrite?
>
> CU Lars.
It´s important reuse all code in BRL-CAD or opencascade. We don´t need
reinvent the wheel. From BRL-CAD we can get help, but from opencascade
we don´t. We must learn opencascade by ourself.
JAVA or not JAVA. GvSIG people want work in JAVA but they need c++
bindings to use openscenegraph or gdal,...
JAVA people can use bindings or build libraris from scratch...
I believe that C is more compatible for all people, and not very
dificult to implement. We have Gobject:
http://library.gnome.org/devel/gobject/unstable/index.html
Using C is easy build CADs in Qt, GTK, FLK, Wxwidgets, TK, VDK...
We have here a war about C vs C++ between linus torlvarls and others :
http://thread.gmane.org/gmane.comp.version-control.git/57643/focus=57918
C and C++ are both good choices for all.
Regards.
Well, than actually we would not have to talk too much about data types,
they are all defined. If this is the way to go, we need to write
triangulation code for brlcad (enabling opengl-display) and than start a
nice GUI - more or less all geometry is implemented there.
> JAVA or not JAVA. GvSIG people want work in JAVA but they need c++
> bindings to use openscenegraph or gdal,...
I think the question was more related on how to store objects in memory.
An application today almost always consists of parts written in
different languages. The day it comes to a dscussion about the UI I will
argue for Java, as I never found any GUI toolkit being really
cross-platform ;-)
CU Lars.
On Oct 19, 1:47 pm, "Lars O. Grobe" <gr...@gmx.net> wrote:
>
> > JAVA or not JAVA. GvSIG people want work in JAVA but they need c++
> > bindings to use openscenegraph or gdal,...
>
> I think the question was more related on how to store objects in memory.
> An application today almost always consists of parts written in
> different languages. The day it comes to a dscussion about the UI I will
> argue for Java, as I never found any GUI toolkit being really
> cross-platform ;-)
>
I tell this about JAVA because Franz said this:
"""Many people use java (collabcad, gvsig ..) many use C++
(opencascade) ..
What do programmers out there prefer ? """"
Qt is 100% cross-platform. Java is joined to Trolltrech to improve the
GUis:
http://trolltech.com/products/qt/jambi
GTK is 100% cross-platform thanks to Ridley project:
http://live.gnome.org/ProjectRidley
In the future all devices are based in qtopia(opie) or in
GMAE(maemo,openmoko,gpe,...)
We have c/c++/python in GMAE and JAVA it´s not sure in GMAE.
http://www.gnome.org/mobile/
Nokia n800 and 770 use GMAE:
http://www.nokia.es/A4323131
In qtopia(opie) we have C++/JAVA.
http://trolltech.com/products/qtopia/qtopia_platform/platform
Sony amylo uses Qtopia:
http://www.learningcenter.sony.us/assets/itpd/mylo/prod/index.html
In c/c++ commmunity the important thing is separate APIs:
XML comes from w3c standard library (libXML)
http://xmlsoft.org/
Graphics from freedesktop:
http://www.freedesktop.org/wiki/Software
text procesing by standard libraries like bison or PCRE. PCRE is used
by PHP,python,Apache, mysql,... I use PCRE and it´s great. It´s a
complete PERL pattern matching in C and C++ bindings.
Regards.
>>Do you plan to include existing libs, or do you want to rewrite?
there are many libs and how-to's in the net; the least effort is to use the
most complete library (if the license makes this possible ..) and rewrite all
missing functions ..
> I believe that C is more compatible for all people, and not very
> dificult to implement.
we have chosen c because of the speed; c seems to be much faster than all
others ..
>>the great final challenge is get a Catia and a Autocad... but this is
in the "Linux Format" - the #1source for Linux (LXF June 2007) on page 21 Mr.
Paul Hudson writes in the article "Clone wars":
.. - and let's face it, any attempt to clone AutoCAD will take several
thousand men years - ..
Franz
I really did not want to start a which-GUI-toolkit-is-better discussion.
I just remembered that I still have some software I would love to use on
my Powerbook that does not build (or is not useable) written using qt on
a Mac. I am sure you can fix it, but it causes lots of trouble.
CU Lars.
On Friday 19 October 2007 12:07, martin_gnu wrote:On Oct 19, 11:09 am, "Lars O. Grobe" <gr...@gmx.net> wrote:How does this relate to existing projects? brlcad is e.g. quite exactly what you describe, such a set of more or less independent libs and tools. They lack a cad-interface that one could really call generalso this could be the primary source of object-definitions and functions .. ? seems to be C-code; how about the licenses ? Could it be used for commercial programs also ? Why are the first words of the brlcad.org: Army .. strong Defense ?
Most of the funding is from the US army, they use it as a solid modeler
for simulations.
> I am not an expert in licenses, but BRL-CAD is a mix of modules of BSD,
> LGPL and GPL.
I think all potential problems arise not before you distribute all
together. So if you have a brlcad installation as a prerequisite but not
distribute it with your whatever-licensed application, there should be
no problem if your application calls functions, right?
If necessary I can try to find out by contacting some debian-folks among
my friends who know about these issues much better then me ;-)
CU Lars.