View previous topic :: View next topic |
Author |
Message |
clayasaurus
Joined: 21 May 2004 Posts: 857
|
Posted: Thu May 01, 2008 8:57 am Post subject: Arc Design... Modular vs. All-in-one |
|
|
Here's an idea that I'd been thinking about in my head a while, about
which design route I should be taking ArcLib in.
1. Modular design
- Arc Core Library
- Arc Scenegraph
- Arc Particle
- ArcLib compatible Blaze
- etc.
2. All-in-one design
- The way it is now. Everything is well mixed together.
I'd rather stick with one design or the other, to have a strong design
philosophy behind the library, at least.
I've been leaning towards the all-in-one design, which is what ArcLib
is currently, for ease of use. However, to keep up with this design, I
/will/ need to integrate Blaze directly into the ArcLib code base. It
also puts the pressure on me to maintain / develop more code, and
slows down the release cycles.
Recently I've been thinking of a modular approach as well... this
approach works well for the SDL library (SDL_image, SDL_ttf, etc.) It
would allow game programmers to 'pick and choose' which components of
ArcLib they would want to use, and the core of ArcLib would become
really well polished. DSSS would remove the pain of installing all the
libraries separately. This way, different members of a future 'arclib
community' can make arclib compatible extension libs and people can
choose to use them or not, and my 'GUI' implementation would have the
possibility of 'disappearing' if a better one from the community
emerged, or certain sections of arclib can be redesigned without
really much of a problem.
Anyways, I've been struggling with these ideas for a while and kind of
want to know what others think about them.
Last edited by clayasaurus on Thu May 15, 2008 7:42 am; edited 1 time in total |
|
Back to top |
|
|
zzzzrrr
Joined: 17 Feb 2007 Posts: 139 Location: Washington, DC
|
Posted: Thu May 01, 2008 9:05 am Post subject: Re: Arc Design... Modular vs. All-in-one |
|
|
clayasaurus wrote: |
Anyways, I've been struggling with these ideas for a while and kind of
want to know what others think about them. |
Hi, Clay.
Good to hear that you've found a little time to think about ArcLib!
I can see pluses and minuses for both the all-in-one and modular design. However, I'm more included to believe the modular approach may be the best strategy to help ArcLib evolve. My reasoning:
#1) Less pressure / time commitment for you maintain
#2) Give developers more choice in package deployment
#3) Frees you up to concentrate on the ArcLib core
#4) Plug and play mods and extension support
#5) More opportunity for community support
I think ArcLib will have a better shot at continuous evolution if you take the modular approach. In any case, I would be more than happy to assist you with Blaze integration or other aspects of ArcLib development as needed.
Regards,
Mason |
|
Back to top |
|
|
Xeon06
Joined: 16 Apr 2008 Posts: 19
|
Posted: Fri May 02, 2008 6:18 am Post subject: |
|
|
I agree with the post above me, modular would definitely be the most useful for me anyways. |
|
Back to top |
|
|
ChristianK
Joined: 26 Sep 2006 Posts: 159 Location: Berlin, Germany
|
Posted: Sat May 03, 2008 12:29 pm Post subject: |
|
|
I concur that it's good to keep things as modular as possible, but also think that splitting Arc into different packages goes too far.
Why would using Blaze force us to incorporate it into our code? Users who don't use the part of Arc that uses Blaze shouldn't need to have it installed. |
|
Back to top |
|
|
clayasaurus
Joined: 21 May 2004 Posts: 857
|
Posted: Sat May 03, 2008 7:38 pm Post subject: |
|
|
True, they could not just use the feature.
However, multiple packages allow for easier concurrent development by community members or allows people to pick and choose features or allows packages to be easily replaceable by other packages.
For example, even though there is already functionality similar to Sprite, I'm thinking of reviving that as a package. I wouldn't have to quite worry as much about user submitted code or low quality packages because people wouldn't be forced to compile and link with them. More code = longer link/compile times.
I think it makes the most sense practically and for long term growth.
Arc core
--> ext1.
--> ext2.
Furthermore, dsss kind of makes it easy to use multiple packages, and packages can still rely on each other or the arc core. |
|
Back to top |
|
|
ChristianK
Joined: 26 Sep 2006 Posts: 159 Location: Berlin, Germany
|
Posted: Wed May 07, 2008 11:37 am Post subject: |
|
|
Quote: | For example, even though there is already functionality similar to Sprite, I'm thinking of reviving that as a package. |
I think that could be a good idea. It'd allow people to get started quickly.
Quote: | I think (multiple packages) make the most sense practically and for long term growth. |
I'm just a bit worried about the additional overhead, but if you think it'll be manageable, sure. Since I haven't found the motivation to contribute to Arc much over the last months, you'll probably be more or less on you own anyway... |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|