[CF-Devel] Map building, continued

[crossfire-devel at archives.real-time.com]

>
     
       Well, testing other stuff is problematic.
     
     >
     
     
     >
     
       Probably the biggest problem arises when someone says 'I have some 
     
     >
     
      experimental <xyz>' - people get turned off the the experimental/still 
     
     >
     
      early work in progress, and think to themselves 'I'll wait for the more 
     
     >
     
      polished version to come out'.
     
     
But the other side of the problem is that it's often easier to do 
experimental stuff, ask others for advice (since they'll always pinpoint 
what you missed, and make useful suggestions, many times :)), then go on 
implementing your ideas. Doing first a total development leads to either 
discussion after, and many changes, or arbitrary commits (like 'this 
works, so i add it even if you guys don't like it')
ok, i'm caricaturing :)

>
     
       I'm personally not a big fan of in game map building.  Have people use 
     
     >
     
      the editor if they want to build maps.  I'd almost be more interested in 
     
     >
     
      some way for players to upload maps and do sanity checking to make sure 
     
     >
     
      they are OK.
     
     
But how can you 'sanity' check automatically? (let's not consider manual 
review for now). Imagine player using tricks with item builders & such 
to create items?
Also, my point of view is that doing maps in editor is less fun. 
Building your apartment ingame is much funnier, and also can be made 
really expensive, so that players need to spent some money ^_-
And don't forget you'd have to fire the editor (and so have it, as well 
as java), make changes, upload each time you wanna adjust something.
And many people probably just don't want to take the time to learn how 
the editor works in the first place.

>
     
       but there is also a difference between map building and map 
     
     >
     
      customization. The former suggests that a player can put most anything 
     
     >
     
      on the map - the later suggests they have some limited ability to change 
     
     >
     
      an existing map.  I don't have quite as much problem with the later.
     
     
What I did (currently) is maybe more map customization. Yes players can 
put walls, doors, remove walls, in maps. But this can be totally 
controlled (use the 'type' field for walls/floors, to check whether you 
can build or not, restrict to unique maps where only the player can go). 
And what they can put/remove is totally controlled, using custom items 
with fixed properties (you can't specifiy what wall archetype to use, 
and such).

So yes, a player could potentially put anything on a map, but that 
requires someone to make specific archetypes for that first.

>
     
       I'm also a bit leery of adding any new large block of code if it isn't 
     
     >
     
      really needed.  More code leads to more bugs and more code to maintain.  
     
     >
     
      But that of course has to be tempered with the usefulness/demand of the 
     
     >
     
      feature.
     
     
Hum. In my (humble) opinion, this code isn't indeed needed. But it's 
something that changes from hack & slash, and I'm pretty sure people 
would have fun decorating their apartment.

>
     
       There is no such branch.  I'm not sure how useful such a branch would 
     
     >
     
      be - if too many people used such a branch (on the idea of hey, its 
     
     >
     
      experimental, I can check this code in right now) - that would lead to 
     
     >
     
      the branch often being out of sync with your code, requiring more work 
     
     >
     
      to potentially be able to get your change up to date.
     
     >
     
     
     >
     
       Also, there is the problem that if there are several such experimental 
     
     >
     
      features in that gate, may be more reluctance for people to use it - 
     
     >
     
      feature A may seem interesting and something to try out, but feature B 
     
     >
     
      (by perhaps someone different) doesn't, but you're then stuck with both 
     
     >
     
      of them to play around with, which could be a detriment (if something is 
     
     >
     
      buggy, harder to pinpoint exactly where the problem is).
     
     
Indeed, that'd be a trouble.
Though I wonder how many feel like testing stuff but don't take time to 
apply patches. I for one have already enough CF trees around so that i 
don't feel like copying, applying a patch, testing, then trashing (or 
applying the patch, testing, reverse-applying the patch).
'pushing' patches to cvs gives a better chance of tests.
(but please lemme know if i'm wrong :) And this is not a criticsm, just 
a comment)

Nicolas 'Ryo'


_______________________________________________
crossfire-devel mailing list
     
     crossfire-devel at lists.real-time.com
     
     
     https://mailman.real-time.com/mailman/listinfo/crossfire-devel