[Freeciv-Dev] Re: gen-impr stuff

["andi payn" <paynfc@xxxxxxxxxxx>]

Per:
> On Fri, 4 Jan 2002, andi payn wrote:
> > First, I added in all of the effects that weren't handled, and
> > changed a few
> > of them around (I don't think pollution adjustments are supposed
> > to be
> > additive, for example), and that's mostly done.  I also added some
> > of the
> > missing things (like reactor meltdowns).
> Nice. But how do you handle the pollution adjustments then? Merely
> pick the higher one?
No, I multiply them. They're obviously supposed to combine in some way, but 
adding them together is not the right way. (Should a 34% Pollu_Adj and a 50% 
Pollu_Adj be equal to an 84% Pollu_Adj? Then the two buildings together are 
less effective than either one separately! Even worse, with no buildings, 
you get 100%--so with a 34% Pollu_Adj, you'd end up with 134%, which is 
worse than nothing!) Treating them as (100-amount) and adding is a little 
more sensible, but still not right (a 34% and a 50% together give you -16% 
that way).
When you multiply, everything seems reasonable (a 34% and a 50% together 
give you 17%, which sounds right). Of course I had to change the rulesets to 
change all of the -50 values to 200 (because they were there to counteract a 
50 value).
I also changed the Pollu_Set effects. They don't make sense added together 
either, and I couldn't imagine how you'd want to combine them, or even why 
you'd ever use anything besides 0, so I just pick the minimum and use that.
> This will presumably be much easier to fix after the ai cleanup crew
> have finished their work.
Yeah, probably. I think building_value as it stands needs to go away, but I 
wanted to leave the structure of the AI code alone as much as possible.
> Sure. Can you upload it as a huge patch to freeciv incoming ftp?
OK, no problem. I just built a huge patch against the 05Nov2001 CVS (which 
is what I started from). The changes include:
* the 11Dec2001 impr-gen-combined patch
* all of my impr-gen changes
* most of my other changes (I stripped out a few things that don't
 work right, like satellites and withdrawing)
* lots of other patches that I've downloaded, mostly modified to
 fit my changes (e.g., the 26Aug2001 CivWorld patch with all kinds
 of changes to support recent CVS builds, impr-gen, vet-levels,
 etc.)
* all of the standard rulesets, modified to handle the new stuff
* the Ancients and Fantasy modpacks (1.12.0-1.4.0 and 1.11.4-1,
 respectively), modified to work with my code and take advantage
 of a few changes (e.g., Ancients has custom calendar ages)
* two rulesets that I've used for testing and playtesting (civ2c
 and civx)
* changes to the hires tileset to handle some of the stuff in the
 various rulesets I've used for testing

Known problems:
* I didn't add capability strings for anything (mainly because I
 started hacking before reading the guide), so don't run my client
 against a stock server or vice-versa
* Reading unmodified rulesets gives you lots of warnings, but
 should work
* Loading existing saved games should work, but may have a few bugs
* The new calendar age code may be screwed up
* The AI misestimates some buildings' values
* The AI never assumes any unit will have a vet level about 1
* I haven't tested on any platforms but linux/x86 and linux/ppc
* Although I've changed the other GUIs (besides Gtk) where necessary,
 I haven't tested them (or even built them)
* The borders code is all screwy, because I started stripping it out,
 then changed my mind; what's left is basically the functionality
 from freeciv-ac with some bugs added...
* I haven't tested the optional diplomatic-battle code in a while
* Most of the later changes (especially the impr-gen stuff) haven't
 been tested multiplayer yet.

Other than that, it should just work.


_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com