[Freeciv-Dev] Re: Development Strategies [Was Documentation, Usability and Development]

["Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx>]

At 03:58 PM 01/12/10 -0500, Jason Short wrote:
>Tony Stuckey wrote:
>
>> On Sun, Dec 09, 2001 at 09:59:55AM -0500, Andrew Sutton wrote:
>> 
>>>the point wasn't necessarily to develop mutable civ rulesets for civ
games - 
>>>although that should certainly be possible (e.g. period games: civil war, 
>>>roman empire, etc.)
>>      Precedent for which are the scenarios for Commercial Civs.
>> Midgard, WW2, etc.
>>>- but to allow developers to customize the game 
>>>completely, inventing new games (like SMAC or MoM).
>>      A goal I'm not sure is a good one.
>
>The goal may be unachievable or too difficult to be worthwhile, but 
>working toward it should give good results of its own.

On the otherhand it might be surprisingly easy.

Moreover, it might make Freeciv development a lot easier which would
enable it to develop a better Freeciv, even if that is a scary thought.

>>  MoM (and MoO, and MoO2) are just plain different games.
>
>I agree with you on that one.

If you look at the rules or the end-product display. But if you start
breaking it down into basic operations, there are maps and units and
units move on the maps with only small changes in the constraints that
govern how this happens.

>>      MoM has 2 separate but overlaid maps; bizarre interconnected unit,
>> building, and faction rules; a completely different combat system with a
>> tactical sub-map; a research tree where major parts are locked out and gift
>> techs and research bonuses are chosen at game start, etc.  *SO* much in
>> both the server and client would have to change to even present this
sanely.
>
>I think many of these shouldn't be as big of obstacles as you're making 
>them out to be.

Implement the general code for small operations and a framework for tying
them together by rules, with specific elements to deal with Freeciv or
Freeciv-like game mechanics as the final subclass extension. Some 
enterprisingly MoM player will come along and add a new set of extensions
for MoM and plug it into the system.

The concepts of how to do it are quite straightforward once you shake
the idea that a monolithic Freeciv codebase that is designed for maximum
inscrutability and inflexibility is the only way to go :-).

>- The research tree where parts are locked out could be simulated by 
>providing nation-specific initial techs.
>
>- The different nation building abilities (units and buildings) can also 
>be simulated by providing nation-specific initial techs.
>
>- The different maps could be simulated easily under the general 
>topologies change I'm proposing.  The problem of moving between the maps 
>would have to be handled by special teleporting places (the towers) or a 
>teleporting unit ability that acts in a fixed way.
>
>- The different combat system (the swords/shields thing) should be easy 
>enough to set up.

These are at least some ways of starting down the generalized path by
thinking about the core concepts and specializing them only at the last
stage.

>But the problems remaining are still tremendous.  Most notably:
>
>- The combat sub-map would be very difficult, most likely impossible 
>under the real-time system FreeCiv currently uses.
>
>- The different city setup would also be a difficult problem.  Instead 
>of being assigned to tiles, workers are just made into "farmers" or 
>"workers" (or, under MoO2, "scientists"), each of which produces a 
>certain amount of food and "production".  It wouldn't be insurmountable, 
>but it's definitely a different system.
>
>- The change from "trade" (=tax+research+luxury) to a "gold"+"power" 
>(=mana+skill+research) system would also be a huge change.  Then you 
>need to start thinking about things like handling skill points, and 
>overland spellcasting.

Refactoring from a system of specifics into hierarchical layers of a
generalized approach is not necessarily obvious at first blush, but
the mark of a good system is that it needn't be required to use it.

At some point the modular simplification can just stop and treat the
underlying implementation of a particular subsystem as a monolithic 
blob. Everything will continue to work. And the next generation with
more time or insight can refactor the blob into simpler component parts
tied together by configuration and rules in ways that expose new sub
elements to any other game that wishes to incorporate them.

Or they can implement this from scratch in a component system and 
replace the blob when there is no functional reason to retain it.

GUI toolkits and Device Drivers are examples of modular components
that are typically configured into systems through well-defined 
interfaces to produce many quite different end products.

>>      Don't get me wrong -- I'd like to see MoM and XCom: Ufo Defense
>> remade with modern technology and better code.  But I don't think that the
>> Freeciv project is the place to even think about that.

MoM specific modules, probably not. But Freeciv modules sounds like an
emminently fine thing for the Freeciv project to be working on.

>Ahh, XCom.  That was a quality game.  Unlike MoM, which can still be 
>played under windows (and I hope will eventually be playable under 
>WINE), I don't think any modern system can reasonably play XCom.
>
>Well, you certainly got me started thinking about it. :-)
>
>jason

Cheers,
RossW
=====