yade-users team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

:
> > The idea behind my proposal is to design a new structure for yade
> > documentation.
> > I know also that the current wiki contains a large quantity of
> > documentation thanks to contributors.
> > But honestly it is not very easy to find some information (this is NOT
> > a criticism because I am very pleased to be able to access the
> > informations contained in the wiki).
> >     
Yes, I more or less wrote a similar message recently. The presentation 
on wikia is not optimal, and there are many adds, which is always a bit 
strange for presenting an academic code (not criticism either, I'm not 
less responsible than others of the current situation anyway). I recall 
cloning the content to a site hosted on a Grenoble University server was 
in my plans. Perhaps merge both plans?

> > So I began this adventure in the hope that WE will find a nice design.
> > Here is a draft proposal of the doc structure: http://yade.wikidot.com
> >     
>
>   
Very good!
However, is there any guarantee that wikidot will not inlude adds at a 
point? IIRC, wikia added more and more with time, and changed the pages 
presentation too.

Vaçlav said :
> Thanks for your proposition. I would prefer to have some content ready
> before designing. There has been too much designing (in yade) of things
> that never see the light or world (like new class names, coding
> conventions, fem models) in yade and I am a bit suspicious about that.
>   
This is true.
> Are there some obstacles for people adding pages now? Likewise, we have
> bug tracker at launchpad, how many people use it?
>   

Not much I guess... The problem as I see it here in Grenoble, is that we 
have many users who are not very experienced in sophisticated coding 
tools (me included). They are happy enough when they can install linux, 
download/compile Yade, handle c++ syntax and understand inheritance. At 
this point, they quit following the incredibly steep learning curve and 
focus on their PhD subject. Most of them won't even commit anything in 
their last PhD year if the superviser is not requesting so.

It is in fact hard to start using the advanced development tools when 
you don't know in what way they could help, or even you don't know they 
exist (see Vincent who is a skilled developper but keeps using cerr/cout 
instead of log4cxx :-P ).

Also, some crucial information about Yade is not in the wiki, nor in 
doxygen doc : it is in this mailing list. Which means new users will 
probably have hard time to find it.

What Jerome Duriez did once before was a really good idea : give a 
starting point for using Yade.
Now, it would be good a time for a "level 2" guide for Yade 
usage/development, which would summarize the advanced use of Yade. 
Perhaps not a big text, but at least a general view with pointers to 
relevant pages.

Emanuele Catalano will start working with Yade very soon (he used Yade 
last year during his master and he is now in 1st year of PhD), and he 
agreed to take notes and prepare this "lvl2" guide while he learns the 
advanced features himself.

I'm thinking about the following possible topics, feel free to add more 
if some come to mind :

- Downloading/compiling Yade
- Documentation (J. Duriez's begginers guide, wiki, doxygen + doxygen 
commenting convention) and how everybody can improve it.
- Creation of new classes and usage of SVN (I see people modifying 
existing classes to do what they want, best way to stop updating/commiting).
- Python console and scripts
- The logic of Sconscripts
- Kdevelop/CTags and perhaps more tricks to navigate in the code
- debuging tools
- usage of log4cxx
- bug tracking (?)
- Parallelization
- Profiling

I know something exists already on most aspects. Again, the idea is more 
to give links than to write a new doc from scratch.
Perhaps also update some of the existing-but-old wiki pages.

Bruno


--

_______________
Chareyre Bruno
Maitre de conference

Grenoble INP
Laboratoire 3SR - bureau E145
BP 53 - 38041, Grenoble cedex 9 - France
Tél : 33 4 56 52 86 21
Fax : 33 4 76 82 70 43
________________