Hi Robert,
You're absolutely right about the fact that OSG isn't a game engine. I
personnaly think OSG should be considered as a good layer to create a game
engine on. That explains what I said about "quick and dirty":
1. People code "quick and dirty" things in games (as often).
2. Some of this code is refactored to be included in the game engine (beceause
general purpose enough for one or two games).
3. Some of this refactored code can be re-refactored to be included in OSG
(beceause general purpose enough for OSG community, in the case OSG is a layer
under a game engine).
The code can thus "get down and be filtered to OSG", as rain is filtered in
soil and finaly is cristal-clear and pure groundwater.
After all, that's the same with any other software. :)
I hope I'm a bit clearer.
About the "fear of OSG", I saw multiple things (mainly in face to face
meetings):
- Some fear "too complex" code ("Pointers? Oh my god!") and are still coding
the 1980's way. Nobody can do anything for them... except C++ training.
- Some don't even know what is a scene graph.
- Some don't see how they could use a scene graph in their game engine.
- Some are sure of their technical superiority. They affirm their "home made"
engine is far better suited than any other, and that building an engine on top
of an existing code is a mistake (whereas they often simply do a XYZ-like game).
- Some simply haven't got the time to create anything "low level" (= not
gameplay code) and will simply buy licences of famous game engines. This
behaviour is certainly pragmatic and well suited for some projects, IMHO.
- And finally I saw some that fear that *any* open source software will also
"open source" their code. They generally don't understand how business is
posible with open source. I even heard things like "utopia", "anarchists" and
"communism"... Consternating.
But I saw once someone that worked with Performer and that moved to OSG to
build an engine on top of it. This is exactly what you said. I worked with him
(that was OSG 0.9.x), but I must admit that before that I didn't know enough
about scene graphs and what they can bring to a game engine. So maybe some
people fear using a scene graph at all, arguing they'll have to learn to use it.
> My wife used to call it the mythical 1.0 because it took so long to
> converge to ;-)
:D
> Glad to say that we've now got systems in place to enable us push out
> the stable releases much more regularly now.
You said it! That's very pleasant to have such a reactivity.
> Understood. My role as lead is partly as stabilising influence, and a
> filter/guide, that can weather the pressures of different needs and
> guide our average contribution in a direction that encompasses as much
> our diverse needs as we can.
Understood too. And I think you're right doint this as you do.
> I think we should probably try and fit the naming of the packing into
> ones discussed a week or so back. This will mean categorising. If
> we want to discuss this further then the previous thread on this topic
> would be the one to restart rather than do it hear on what is
> effectively a different topic.
Hum... I'm sorry, but I don't understand. Do you mean create separate binaries
archives/installers, like "Core OSG", "Standard plugins", etc.?
Anyway, I'll wait for CPack results...
Sukender
PVLE - Lightweight cross-platform game engine - http://pvle.sourceforge.net/
Le Fri, 12 Dec 2008 10:36:47 +0100, Robert Osfield <robert.osfi...@gmail.com> a
écrit:
> Hi Sukender,
>
> On Thu, Dec 11, 2008 at 9:53 PM, Sukender <suky0...@free.fr> wrote:
>> Seeing your answer, it seems that I was not as clear as I should have been.
>> Well, about blogs, that's okay.
>> But about gaming industry, I was just saying that advertising/marketing
>> towards game developpers should be interesting. For now, I have the
>> impression that some of them fear OSG as if it was too complex, or not
>> enough game-oriented for them (which is wrong, IMHO).
>
> The OSG certainly isn't a game centric API, and certainly doesn't have
> the feature set to be considered a game engine, so I can certainly see
> that many might see the OSG as something not best suited for gaming.
> The OSG is in fact about as general purpose as scene graph as has ever
> been written so can certainly be used, and is used, for games, but it
> does take seeing beyond needing a full game engine that ticks all your
> boxes.
> You impression of fear is an interesting one. What influenced you to
> come to this opinion? Online discussions/blogs? Face to face
> meetings?
>
> My guess is that if one hasn't used a professional grade scene graph
> before I can certainly see that it might seem a bit daunting. What is
> pretty clear is that those who've used Performer/Inventor/Vega and
> similar API's before take to the OSG pretty easily, often it's similar
> enough to the concepts used in the other scene graph for them to grok
> how one might tackle things, and after a short while the realisation
> of extensibility, flexibility and access to source code and vibrant
> community that the OSG adds on top this prior art does seem to be
> appreciated.
>
> I suspect the much of the games industry sees scene graphs as rather
> more limited and simpler structures than we have, pushing complexity
> more into the application framework than the underlying scene graph.
>
>> When I said that "Quick-and-dirty" would happen, I was talking about what
>> some people may code (in general), but not about core-OSG code. Only clean
>> and general-purpose code should be merged, as usual: I was not saying that
>> quick-and-dirty should be merged.
>
> OK. So I how would the "Quick-and-ditry" affect us? Are you
> suggesting that you just want to encourage game developers to not fear
> using the OSG and just dive in and prototype with it? Not clear on
> what you are envisaging.
>
>> I began using OSG since 0.9.x (I remember how the wait until 1.0 was
>> terrible!):
>
> My wife used to call it the mythical 1.0 because it took so long to
> converge to ;-)
>
> Glad to say that we've now got systems in place to enable us push out
> the stable releases much more regularly now.
>
>> I of course saw what has been achieved and think (as many) that the
>> direction is the right one. I'm not trying to
>> *change* OSG, I'm just trying to open discussions about what could be
>> interesting for game devs or game engine
>> devs, and thus for OSG.
>
> Understood. My role as lead is partly as stabilising influence, and a
> filter/guide, that can weather the pressures of different needs and
> guide our average contribution in a direction that encompasses as much
> our diverse needs as we can.
>
>> And about management, I was not talking about management *of the project*
>> but of some contributions, I'm terribly sorry if I seemed unpleasant.
>
> No offence was taken.
>
>> I was simply saying that some contributors are willing to build binaries,
>> test things and do other tasks, and these short-terms tasks could be tracked
>> on a public chart (or something like that) in order to prevent doing the
>> same thing twice, and motivate some to fill the chart (well at least me!).
>> I've done a draft on the wiki's sandbox for binaries builds:
>> http://www.openscenegraph.org/projects/osg/wiki/SandBox/ContributorBuilds .
>> Useful?
>
> I think we should probably try and fit the naming of the packing into
> ones discussed a week or so back. This will mean categorising. If
> we want to discuss this further then the previous thread on this topic
> would be the one to restart rather than do it hear on what is
> effectively a different topic.
>
>> Off-topic question: is it possible to change the HTML title, META keywords
>> and descriptions in the wiki?
>
> Good question. It is possible, but I don't know enough about Tracs to
> give you an answer.
>
> Robert.
> _______________________________________________
> osg-users mailing list
> osg-users@lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
_______________________________________________
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
