Re: [freesci-develop] Proposal for new direction of the project

[Christoph Reichenbach]

Hi,

On Sun, Nov 05, 2006 at 10:53:47PM +0000, Jordi Vilalta wrote:
> Our main idea was to try to add SCI support into ScummVM. The obvious
> easy way would be to integrate the FreeSCI code into ScummVM. We would
> like to know what do you think about it, because we don't want to fork
> from FreeSCI, but to join efforts into a larger project. We want to
> establish collaboration between both projects and to do it with your
> acceptance.

  The idea is conceptionally with quite some merit, but there are certain
technical issues to be overcome.  I recommend you to review the IRC logs.
Unfortunately, I'm a bit low on spare time ATM, so I'll just give you a quick
result of grepping for ScummVM:

http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.02/log20040215.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.03/log20040314.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.03/log20040319.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.03/log20040328.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.03/log20040329.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.03/log20040331.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.04/log20040401.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.04/log20040405.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.04/log20040406.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.04/log20040407.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.07/log20040722.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.08/log20040805.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2004.08/log20040822.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.01/log20050106.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.02/log20050210.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.09/log20050925.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.10/log20051012.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.10/log20051022.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.10/log20051030.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2005.10/log20051031.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.02/log20060211.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.02/log20060212.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.02/log20060217.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.02/log20060218.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.03/log20060317.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.03/log20060331.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.04/log20060415.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.05/log20060513.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.06/log20060610.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.08/log20060811.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.08/log20060829.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.09/log20060905.html 
http://www-plan.cs.colorado.edu/creichen/freesci-logs/2006.10/log20061002.html 

  At least the later 2006 results are likely to be relevant.

> In our opinion, this merge would benefit both projects: ScummVM would
> get support for some new games (like it did with the relatively recent
> integration of Sarien), and FreeSCI would win a lot of new testers and
> probably some new hands on your code. By the way, the new SCI engine
> would also immediately benefit from the nice features from ScummVM,
> such as great portability, graphic scalers, etc.

  Simulataneously, FreeSCI would lose some of its benefits, such as scaled
background images and pre-scaled views.  I see the benefit mostly for SCI1
games, for which ScummVM's palette emulation would be very helpful.  In
general, I see ScummVM's gfx subsystem as a possible graphics driver for
FreeSCI.  I'm less clear on sound and configuration, the other areas in which
merging would be possible.

> This code move could also benefit of adding new features. For example,
> we've done a small look at the SCI Studio code and it seems like there
> are some parts that support newer SCI versions than FreeSCI does. We
> think it could be a great opportunity to integrate such code into
> FreeSCI (or directly into the new ScummVM SCI engine).

  Before you incorporate any code into FreeSCI, please make sure that it
wasn't derived from illegal reverse engineering.  FreeSCI is a clean-room
re-implementation of SCI.

> Of course we don't want to tell you how to rule your own project, but
> we think it could be a good step, and we would help with the process.

  It's ultimately Lars' call, him being the maintainer.

> If you agree with this proposal, then we could talk about what
> codebase to take (glutton I guess), what things should be enhanced,
> etc., to make FreeSCI go further with this process.

  My _personal_ opinion/position, summarised:
  (a) I will not be able to help much in the near future, no matter which
      direction the project takes.  This is likely to be a permanent issue.
  (b) There are possible benefits for FreeSCI from re-using ScummVM code, but
      I'm not sure if our infrastructure agrees with theirs enough.  (Recall
      that the ScummVM infrastructure was developed for reverse engineered
      games which largely use primitive IBM PC-style graphics/sound
      operations.  The FreeSCI infrastructure was developed specifically for
      handling SCI.)  IMO, It would not be too hard to port FreeSCI to ScummVM
      while sacrificing the various additional features we've added
      (cf. README).  It might be very hard to do a merge without sacrificing
      anything.
  (c) The FreeSCI project should take whichever direction benefits this
      particular project most.

-- Christoph