[freeciv-ai] Re: Massive AI 5

["Per I. Mathisen" <per@xxxxxxxxxxx>]

On Mon, 11 Nov 2002, Raahul Kumar wrote:
> I'm sorry Per, but you're introducing far too many big changes in one go. The
> place well known for its hot climate is clearly freezing over, because I have
> to agree with Raimar. Please split this up.

I'll do it if you promise to review the split up patches... ;)

> Now for the relevant bits. Why is your gold reserve 5 * pplayer citizens ? Is
> this number based on something, or did you pick it because it worked?

It gives diplomats enough gold to incite cities. Number chosen by lots of
trial and error. If we make AI diplomats go for techs first, we can lower
this number, which would be a good thing because it gives it more time for
science (less tax men mode).

> Next, you need to get rid of the limited huts implementation you have in your
> massiveai patch. Greg already split that up.

Yes. I didn't remember to remove that before making the patch.

> +    } else if (city_list_size(&pplayer->cities) > 6) {
> +        /* Don't waste precious money buying settlers late game
> +         * since this raises taxes, and we want science. Adjust this
> +         * again when our tax algorithm is smarter. */
>          continue;
>
> Can you send in a patch, an independent one, that makes the easy AI not
> buy settlers?

Already been posted. Attached again as "speedai".

> And also make it build coinage all the time.

I suppose you mean "not". That patch was posted earlier and has now been
committed (received no comments on it). In case you are interested, I've
also attached it.

> I love your DEFENCE and ATTACK POWER macros by the way. Greg's influence?

Nope. All my own idea ;)

> I liked your use of bools. It should be Freeciv policy.

It is not?

> You're making simultaneous changes to Goto, unit roles, hp recovery, 
> bodyguards
> and diplomacy. I can't follow all the changes, and I've actually read the
> individual patches before. It's too much.

AFAIR I haven't made any changes to diplomacy, so you must mean diplomats.
I'm trying to get that committed first now. The unit roles changes should
now be in cvs. I can take out hp recovery to make the patch smaller, then
commit that later.

> +void ai_unit_new_activity(struct unit *punit, enum unit_activity task)
...
> Does this mean that now the AI will actually fortify properly?

Yes!

  - Per

trump.diff
Description: Text document

speedai1.diff
Description: Text document