Networking (Re: Release Plans)

[Sascha Flohr <s.flohr@xxxxxxx>]

On Mi, 2005-01-19 at 18:47 +0000, Jens Granseuer wrote:
> On 19.01.2005 12:58, Marcus Schreiner wrote:
> > > Then we should concentrate on polishing the maps and squashing the 
> > > bugs and go for 0.5.0 with two complete campaigns, what do you say?
> 
> I'd say that's a pretty ambitious timeline unless we can recruit some
> as yet hidden army of mappers. And if more people start making maps
> I don't expect myself to have much time left for mapping. As you can
> see when looking at the current situation, a single mapper can easily
> swamp me with ideas upon ideas of nice things to add. And multiple
> mappers probably mean multiplication of suggestions ;)

Perhaps we have to take more effort to the map creation tool chain. It
is beside the problems of the "creative process" of map creation perhaps
too difficult for most users to create maps.

Currently I do not have enough insight for map creation and its
problems. Perhaps we should collect the opinions of current map creators
what they think about the editor and the tools to create maps and what
would be supportive to them.


> (In fact, right now, I'd be happy if we managed to add one map per
> release to the Yalwa campaign.)

If we manage to release in short enough terms that would be sufficient.

Currently I like the two maps of the Yalwa-Campaign. They are easy
enough for beginners. My opinion about them:

The first map is small and can be solved very fast, that is important
for beginners. I my self like simple entry.

The second map looks very complex compared to the first. Perhaps it
LOOKS a little bit too complex for beginners. Winning this map isn't as
difficult as it seems at a first glance. A pro is that the player has
the chance to learn about capabilities of the units without the danger
of loosing too frequent.

I think the only glitch (perhaps only for me) is that the same strategy
wins in both maps. Bind the computers units an then rush with the
infanterists (supported by the transporter) to the target buildings.

The third map should avoid that.

A solution could be that you have to hold buildings some rounds to win!
 

> Speaking of 0.5.0, I don't yet know what could justify the jump. Two
> completed campaigns probably would but I don't expect that to happen

One complete campaign should be the first goal ... remember the
tilesets :-)


> soon. Network play could be another major feature if Sascha really
> starts to work on it. If none of this happens we'll probably stay
> at 0.4.x for a while.

I really started hacking on this problem. First results may be expected
next week. Currently I have experimented with the network library for
SDL. Any worries about adding this dependency to another lib beside base
SDL and SDL_ttf? Using plain bsd sockets would be as easy as the lib,
but not as portable.


BTW ... what development environment are you using?

My preferred environment is surely Eclipse since my main programmming
plays the Java game. But I also like the latest milestone build of the
Eclipse CDT. Integrates very well with Linux and gcc.

Sidenote: Eclipse has a very nice analysing system (the C/C++ Indexer)
that gives many hints to code. It mentions a few glitches in the current
code, for example with the macros in "hexsup.h" used in "ai.cpp".

The macro ReverseDir() was used:

	int j = ReverseDir( (Direction)i );

But the macro is defined:

	#define ReverseDir(dir)  ((Direction)((dir + 3) % 6))

So there are two type-castings. Most compilers ignore this, but really
it is not correct. There are other place where you use the macro and did
not double the cast.

I would prefer to dismiss the usage of macros completely. That's
C-coding not C++ :-) All the uses of macros could be replaced by methods
(inline if necessary) and constants. The have the benefit of type
savety.


> all that much, yet. My preference, however, would be to build the campaign
> from beginning to end in the proper chronological order, so I'd propose
> doing the sequel to Swords & Plowshares if that suits you.

My 2ct for that!

Sascha.