What do YOU want to see improved?

[Implementor37]

What if instead of adding features to the editor, a wmlupdt is created (use similar to wmllint) which would scan an existing WML file for out-of-date segments (i.e. code supported in 1.8, but not 1.12) and update the code to the newer version (if this is not possible, perhaps it could comment out bad code segments to be reviewed by the maintainer). If it is too difficult to create backwards compatibility all the way back to Wesnoth 1.2, maybe starting with 1.10 and moving forward as Wesnoth development continues would help. This would not change anything with regards to add-on creation (avoiding problems with campgen seen in 1.2), but would make it easier for maintainers to update campaigns to the current version of Wesnoth.

Perhaps the best solution to the lack of editor functionality and the difficulty in add-on creation is the WML Workshop Forum

[iceiceice]

• I think wmllint is actually supposed to do that, that's the main purpose of it.
• wmllint has limitations -- it can only do fairly superficial kinds of analysis. It's not going to build a giant database and cross reference things, its just going to scan and look for problems that are "local" and "obvious".
• wmllint is written in python rather than C++. All of our wml tools are like this, and I don't believe that any of them actually parses WML in the same way as the game does. In any case where your wml isn't as easy as possible to parse, they may be error prone. They might not handle some macro constructions the same way. It's really hard to say. IMO this is a design flaw in WML, if we wanted these things to work then either all the tools should be in C++, or WML should have been defined using a formal grammer, and the parsers generated using a parser generator like YACC or ANTLR or something which can make both C++ and python.
• We the developers don't always put backwards compatibility fixes into wml lint. Honestly I don't really know what would go in wmllint besides something like, if we changed the name of a tag or a field in a tag. And that's sort of a bad idea in the first place.
• The most significant compatibility breakage it seems in 1.12 was in mp scenarios, especially campaigns, because all of the code that initializes mp games got changed. Although these bugs were reported, we still haven't fixed alot / most of them. (We fixed / may fix some of the big ones in 1.12.2 soon.) I don't think any of these changes were intended, it just wasn't tested adequately when it all got committed. We would rather to fix the bugs than write wmllint code to encourage people to change their code to work around the bugs I think.

[Anonymissimus]

If someone wants to see what disasters CampGen led to, the old 1.2 add-ons server should still contain some.

I recall a campgen campaign I once played with 1.2...huge maps, all scenarios of type "defeat all enemy leaders", a little dialog or story at the start of each, that's it. Nothing else, no other dialogs, no additional [event]s or even only [message]s (story), no [item]s - fatiguing. I played maybe 1 or 2 scenarios and then skimmed through it with debug mode as I couldn't believe it.
Thanks for agreeing with me E_H - the only one.

@Implementor37
You underestimate the difficulty to keep wesnoth backwards-compatible on one hand and to create lifting wmllint code for everything on the other hand.

There are also tools in perl (e.g. wmlxgettext) or shell script (e.g. wesnoth-optipng). Personnally I have a few in lua, albeit lua is probably not the best choice when run as a standalone language for such a tool. The tools are written in what language the particular dev knows, basically.

We the developers don't always put backwards compatibility fixes into wml lint.

But devs also don't dedicate enough energy to this (as opposed to new C++ features). For instance, it should have been possible to auto-update the "Deprecated AI Targeting Aspects". This is still a pending painful problem in my campaign as there are so many occurrences.

[Kinrany]

What's your experience with other programming languages? If it's "none", then of course you feel confused. Everyone is confused when learning his first programming language... Take your time and don't rush learning it!

I'm a CS student
It's strange that you have to write code and add links manually to create content.

[Nobun]

After thinking so much, I choosed

"Quality of the GUI" (but I have to explain what I mean)
and
"improvements to balance of mainline MP eras"

but I have other two things thing to add, wich are not directly related with wesnoth development itself.... now I will explain my opinion about my choices and other things appeared here.

-----------------------------------------

1) Quality of the GUI

I mean.... GUI should be completely customizable in some way, perhaps introducing a better system to customize it.
This becouse.... when new modifications to GUI will be added, they are not always the best choices in all situations (depending also of the user likes).
Making an easy example: I don't like too much the new "blue" background color of the menus and buttons (I found much confortable for my eyse the old grey color)... but I can't change this thing without modifying wesnoth sourcecode

2) "improvements to balance MP era"

I don't agree with khalifa introduction for a lot of reasons (this is not the place to speak deeply about it) including some doubts about balancement (one for all: why they have a healer +8 lev 1 while in all other faction the MP "rule" was "if you have a healer lev1, the healer must heal +4 at max (and not +8) and NOT cure?)
About all other "standard" factions: some units still appears too powerful or too weak. I think about Drakes cold weaknes -50% that is too much high, seeing that an adept can almost kill a drake lev1 during day with one attack only
Undeads: they are too slow and too disadvantaged for their slowness compared to others: dwarves, in comparison, can have some solutions to solve this problem... but the main problem about undeads slowness is that undeads skeletons and ghouls (the melee units) NEVER can be quick.... this is, imho, a too great disadvantage compared to the "advantage" of not being able to be poisoned.

3) Lua documentation

Coming to the "not directly related to wesnoth", I have to mark that lua documentation is very confused and not clear at all.
It is not clear, for example, when and how you can introduce your lua code in a good way.
A lot of suggestions you found in some pages (for example using wesnoth.fire in order to create a message) are overdated.
Ducumentations leaks about examples of "good lua integration"
etc etc

4) wesnoth sourcecode documentation

Please, please..... add more documentation to wesnoth sourcecode....
Wesnoth sourcecode is very very hard to read, for several reasons:
1) it is not so obvious which file(s) contains the code for a component....
2) there are not enough comment INSIDE code.... reading a code made by someone else is always more difficult than creating from scratch.... so, even if for the creator of sourcecode can appear obvious what a piece of code is doing, this is not always true for anyone else that reads it

So more comments inside sourcecode, please

-----------------------------------------------------------------------

I not voted "AI" becouse developers never ends to develop AI and they always try to improve it.... programming a good AI is very difficult, but I think that we players should appreciate that developers never stop to think about AI and never stop to trying to improve it

About the editor improvements:
I agree with the developer who is feared about the "too much simplicity can make easyer to have too many weird add-ons".
but also the other developer (sorry I don't remember the name) is right: "don't make things needlessy difficult... becouse if you make a thing needlessy difficult it is harder for UMC developer to improve and mantain the add-on)

so... the diffuculty should be "medium", not too hard to add difficulties you could avoid, not too simple in order to avoid that peoples understimate UMC development.

the solution, imho, is not to improve editor introducing new things.... but trying to add more documentations for beginner (like someone did) and add some tools to help the development of the very-difficult-to-develop WML components.
I think about unit creation... that is very complex (also to mantain) so a tool like WesWorkshop (a not official tool) should be considered a nice thing

[Crushmaster]

I voted for quality of the GUI and WML enhancements (not that I profess to know all of what that would entail ).

I made an entire campaign in the past, so I definitely don't think WML is too difficult, however, a little more would be nice. However, I can definitely see the point about it potentially filling the add-on server with poorly produced material. A rating system would help with that some, though, as previously noted.

The AI is quite good - I am (or was ) a reasonably skilled player, and I was never ever able to beat two AI teams with a hundred gold each on a mainline map (I did do it on a random map, though, I believe, which are, of course, unbalanced). But, of course, more improvements are always a good thing...Lots here would be splendid. Though, of course, that generic of an answer isn't too helpful.

[GunChleoc]

I enjoy working on design, so I've started looking into ways the UI could be improved. Not making much progress, but I don't have a lot of experience or time. And I'd like more input on what a redesign could benefit from. What problems y'all have noticed, what features you'd like, etc.

I don't know if that's what the community wants or needs, but I'm partly doing it for my own practice.

You could start by browsing the bug tracker. Personally, I'd like to have this bug fixed, so that the game will become playable again in my language.