[CF-Devel] Future crossfire changes/projects

[Mark Wedel]

 Please remove me from an explicity copy if also sending to the list - I don't
need two copies of everthing.


gros wrote:
>
     
      I think the problem is not 'too less docs' but 'badly organized docs'. The
     
     >
     
      source code itself is well commented, but it is sometimes hard to find 'where
     
     >
     
      is ... done'.
     
     
 Docs are always hard to keep up to date. 

 Certainly, there are some problems.  The client/server protocol is (IMO)
decently defined, but that file is actually in the client distribution
(Protocol).

 The other problem is just that documentation does not tend to be a high
priority issue for most developers.  Its much more fun to write up some new
feature than go back and document it.  I think the documentation has probably
slowly gotten better, but there is just a lot stuff that was done before there
was even much a push for documentation.


>
     
      There was some discussion about it one year ago (The question was to decide
     
     >
     
      if splitting the game engine and the rules was a good idea). Although I think
     
     >
     
      an object-oriented language like C++ or Java should be better for Crossfire,
     
     >
     
      porting the actual code would need massive rewriting, not only because each
     
     >
     
      function needs to be translated to another 'dialect' (not a big problem for
     
     >
     
      C++, a bit harder for Java), but also because some basics would need a
     
     >
     
      complete rethinking. Anyway, if someone gets enough time and will...
     
     
 IMO, if going to C++ was actually going to be done, it would probably be better
to do start from scratch.  Sure - look at what is currently there, but I have a
feeling that trying to take the existing C code and modify it piece by piece
into C++ probably is not a useful exercise.

 As a note, a couple years back, someone was looking to more or less do that
(wanted to give it as a class project to write an OOP game), and determined it
would be easier to start from scratch than try to redo crossfire.

 As far as the engine/rules split, that is much more a code re-orginzation than
necessarily a rewrite.  Some new code would be needed.  But a lot of the idea
would be to try and clean up a lot of the mess bettween the 'common' and
'server' directory we have right now.  A lot of the stuff in common isn't that
general.  For example, does god code really belong in the common directory? 
What is currently in the common directory isn't really that common - it is tried
pretty precisely to crossfire..

 If crossedit goes away/gets redone, such a cleanup may be more useful, as there
is also some code in there that only crossedit uses.