Are the action target kind names clear enough?

Contribute, display and discuss rulesets and modpacks for use in Freeciv here.
Post Reply
sveinung
Elite
Posts: 548
Joined: Wed Feb 20, 2013 4:50 pm

Are the action target kind names clear enough?

Post by sveinung »

The action target kind names will be exposed to the ruleset in 3.1. We have "individual cities", "individual units", "unit stacks", "tiles", "tile extras" and "itself".

Should "individual cities" drop consistency with "individual units" (that must be separated from "unit stacks") and become "cities"? Is it clear that "itself" means that an action targets the actor (unit) or would something like "none" be better?

Asking here since the people that are going to write in the language I change should have a say.
Ignatus
Elite
Posts: 644
Joined: Mon Nov 06, 2017 12:05 pm
Location: St.Petersburg, Russia
Contact:

Re: Are the action target kind names clear enough?

Post by Ignatus »

Not an English speaker (as many of ruleset authors), but it seems strange that you put them in plural. Maybe we should keep singular as in the code. "stack" is even more comprehensive than "units", maybe we should combine it into "units of a stack" to stress that target_reqs are evaluated towards each single one. (And yes I want actions against only some units of a stack to unhardcode "Unreachable", but it's another song.)
nef
Elite
Posts: 324
Joined: Mon Jun 25, 2018 5:01 pm

Re: Are the action target kind names clear enough?

Post by nef »

City, Unit, Stack, Tile, Extra, Self.
sveinung
Elite
Posts: 548
Joined: Wed Feb 20, 2013 4:50 pm

Re: Are the action target kind names clear enough?

Post by sveinung »

Thank you for your feed back. I now know that the names should be improved. Issue created.
Ignatus wrote:Not an English speaker
I myself am not an English native speaker.
Ignatus wrote:it seems strange that you put them in plural.
It felt natural while writing the autohelp. What a non native speaker feels is natural in the auto help shouldn't determine what ends up in the ruleset format.
Ignatus wrote:"stack" is even more comprehensive than "units", maybe we should combine it into "units of a stack" to stress that target_reqs are evaluated towards each single one.
Good point about hinting about each unit. So the question now is "short and easy to write" vs "longer but clearer".
nef wrote:City, Unit, Stack, Tile, Extra, Self.
Would "Self" be better than "None" even if sub actors and sub targets - probably set to None by default - would appear later?
Ignatus wrote:(And yes I want actions against only some units of a stack to unhardcode "Unreachable", but it's another song.)
So do I. Many times I have planned out soft coding Unreachable (and its interaction with the unreachableprotects server setting) to the point of almost doing it before getting distracted and losing my plans. The first was while moving "Attack" to enabler control still was nothing but a plan. (There has been some progress. I once got far enough to add a new requirement type that took boolean server settings.) If I know myself correctly I probably didn't even file an issue in some tracker with some hints about the details of how to split the attack actions, convert the existing rulesets and set up rscompat. But it won't be that hard to plan again since I can "cheat" by remembering old details so I'll probably get around to it some day unless someone else does it first.

Update: I have now filed an issue. Thank you for the reminder!
cazfi
Elite
Posts: 3069
Joined: Tue Jan 29, 2013 6:54 pm

Re: Are the action target kind names clear enough?

Post by cazfi »

As part of the final(?) push to get 3.1 ruleset formats ready, I'm now looking this. As I don't think that there can be a good solution with the marriage of the string used in defining the rules, and in the one that appears in the helptext, I'm about to divorce them - to have two separate sets of strings instead -> https://osdn.net/projects/freeciv/ticket/44402

Then we have to decide what the final rules defining strings are in 3.1. The strings appearing in help can be adjusted also after the ruleset format has been frozen.

As for the rule names nef's list seems concise.
Ignatus has a point in that "Stack" is a bit misleading in that we do not evaluate against the stack as a whole but each unit in it. As far as I see we either make it longish "Units in a stack", or treat this as a documentation issue (and assume that ruleset authors read the documentation).
Ignatus
Elite
Posts: 644
Joined: Mon Nov 06, 2017 12:05 pm
Location: St.Petersburg, Russia
Contact:

Re: Are the action target kind names clear enough?

Post by Ignatus »

cazfi wrote:As for the rule names nef's list seems concise.
Ignatus has a point in that "Stack" is a bit misleading in that we do not evaluate against the stack as a whole but each unit in it. As far as I see we either make it longish "Units in a stack", or treat this as a documentation issue (and assume that ruleset authors read the documentation).
For sure ruleset authors have to RTFM any way.
cazfi
Elite
Posts: 3069
Joined: Tue Jan 29, 2013 6:54 pm

Re: Are the action target kind names clear enough?

Post by cazfi »

I think I'll do one change compared to nef's list. Even the underlying type, not just the old name, for ATK_EXTRAS is plural (it refers to all tile extras, individual extras are its sub targets). For that one plural in name makes sense (especially considering any talk about sub targets: "One extra out of extra" vs "One extra out of extras")
Post Reply