[CMake-Promote] CelEngine

[Brandon J. Van Every]

Brandon J. Van Every wrote:
>
> As it turns out, CelEngine is not trying to be yet-another-3D-engine.  
> It's trying to encapsulate content, using any old 3D engine as a 
> plugin!  Rather ambitious goal.  I would go so far as to say 
> foolhardy, and lacking a business model to justify the commercial part 
> of the license, but we'll see what they do.  Ogre3D, a fairly popular 
> open source 3D engine, is their initial plugin target.  I've raised 
> some strategic questions in the CelEngine forum, like whether it would 
> be better to just write a content system on top of Ogre and call it 
> good.  Maybe they'll think about it, and having thought of it, maybe 
> they'll become a noteworthy project someday.

As it turns out, the words of 1 respondent in the CelEngine forum have 
misled me. 
http://forums.celengine.org/viewtopic.php?t=12
Other respondents have said that CelEngine is in fact an Ogre-centric 
project, and all the stuff about plugins is only to give them options 
later.  They are focused on Ogre for now, which from an engineering 
standpoint is sane.  So it seems I haven't offered any kind of 
guidance.  They already have the right idea, they just aren't 
communicating it on their homepage.  So, now they've gotten my feedback, 
that it's worth having these things clearly spelled out on their 
homepage, so that other developers will take their project seriously.

As for licensing, I've been arguing the merits of the LGPL over "GPL + 
commercial," their current scheme.  I do feel that more people will be 
interested in their project if they go with the LGPL, especially given 
that Ogre itself is LGPL.  Of course their licensing is up to them.  
Sometimes a licensing decision is driven by firm beliefs about how 
software should be distributed, like the recent arguments on the main 
CMake list about dual GPL + BSD licenses, for instance.

More often though, I find that when a project is just getting started, 
the project proponents really haven't thought through the consequences 
of various licenses.  They tend to grab the GPL simply because it's a 
widely disseminated license that a lot of people feel they understand.  
Emphasis on *feel* they understand, because most people haven't gone 
over the GPL with a fine toothed comb to understand what they really can 
and can't do with it.  For instance, I explained to them how someone can 
make $1 million off the sale of GPLed software in a vertical market.  
The GPL is not about preventing developers from making money; rather, it 
enforces a particular business model.  In practice, software development 
has to be sold as a service, since the code itself is easily reproduced 
and disseminated once sold.

So now it's wait-and-see if they can grow their project, stay organized, 
hit milestones, garner interest, get momentum, etc.  I don't feel like 
stepping in there personally to do their CMake build scripts, and it may 
not be necessary anyways as they seem to have a CMake expert in the 
house.  But I will keep track of them, especially as to how their work 
is valued by the Ogre community.  I could get more people to say, "Hey 
how about the LGPL" for instance, and thereby accelerate their growth.  
If *Ogre* went CMake that would be *very* valuable.  So proxy exposure 
to CMake is important; it's worth offering simple advice on what may 
make CelEngine more popular in the Ogre community.

Meanwhile I'm going to see if G3D http://g3d-cpp.sourceforge.net/ is 
buildworthy on Windows.  Then I'll look at what can be accomplished with 
it already.  If it passes muster, then I'll see if it's easily 
Chickenizeable.  If it is, then I'll help write their CMake scripts.  
I'm going to be careful though, because the last time I took this sort 
of thing on, it cost me 9 months!  We'll be shipping a unified Chicken 
tarball in the next release, supporting both the old ./configure and the 
new CMake build in the same source tree.  That took some doing.  I am 
hoping that G3D is a lot more stable on Windows already than Chicken 
was.  With Chicken, there was tons of work to get the MinGW build up to 
snuff, and a fair chunk for MSVC also.  It just didn't have a lot of use 
or testing on Windows.  I don't want to support that kind of half-broken 
project again.


Cheers,
Brandon Van Every