Where should Wesnoth go next?

[RedLTeut]

Non-scripted GUI-controlled things are hard to follow and modify.

Okay, simple stuff is easier with GUI, but once there are 20+ triggers even the GUI starts to resort to using names and text to describe them, so in the end the GUI just gets in the way

e.g. If you ever failed to properly use styles in MS Word(ever heard of them? most people gave up), you'd see that sometimes a GUI does not mean to be in control

[wisdomless]

I like where you all seem to be going with this... Keep up the good work!

[k8to]

I'd like to see two things in wesnoth in the future.

- Extra easy difficulty mode. I really mean this. Something that the regulars on this forum think is stupid in its simplicitly would be a great addition to the game.

- Lower CPU usage. I want to be able to scroll the map and choose and move units without noticable lag. Right now I have in some cases multi-second lag just to select units, probably having to do with how the shadowing is done. It's actually bad enough to to interfere cognitively with play.

[im the one you think of]

i think that in wesnoth itself not much could be changed with a positive affect to overall gameplay, however i think some easier (but not tutorial type easy) official campaigns should be included in the game, i think "the legend of wesmere" would be a good candidate as, on easy, i managed to beat it in around 6 hours gameplay, and i am not the best at wesnoth, in fact i find most campaigns difficult and resort to save-loading reasonably often. so basically i think that only basic bug-fixes and more variety is needed

[turin]

I'd like to see two things in wesnoth in the future.

- Extra easy difficulty mode. I really mean this. Something that the regulars on this forum think is stupid in its simplicitly would be a great addition to the game.

Most regulars (like me) don't dislike it because of its simplicity, they dislike it because they thing easy mode is easy enough already.

IMHO, a better solution would be to improve the tutorial... a lot.

- Lower CPU usage. I want to be able to scroll the map and choose and move units without noticable lag. Right now I have in some cases multi-second lag just to select units, probably having to do with how the shadowing is done. It's actually bad enough to to interfere cognitively with play.

Your guess is wrong, the shadowing is not rendered at all.

How slow is your computer in comparison with other games?

[k8to]

First, a suggestion. Don't start off responses with "you are wrong". it's pointlessly contrary and irritating. It's especially annoying when you haven't even really considered what might really be meant by the original post.

I'd like to see two things in wesnoth in the future.

- Extra easy difficulty mode. I really mean this. Something that the regulars on this forum think is stupid in its simplicitly would be a great addition to the game.

Most regulars (like me) don't dislike it because of its simplicity, they dislike it because they thing easy mode is easy enough already.

I meant, of course, simple in terms of strategy due to the ease of victory. And yes, the easy you think is easy enough is not easy enough.

- Lower CPU usage. I want to be able to scroll the map and choose and move units without noticable lag. Right now I have in some cases multi-second lag just to select units, probably having to do with how the shadowing is done. It's actually bad enough to to interfere cognitively with play.

Your guess is wrong, the shadowing is not rendered at all.

How slow is your computer in comparison with other games?

By "shadowing" I am not referring to rendering or any visual effect. If i had meant actual shadows drawn beside units, I would have called it 'drawing of shadows' or something. I mean the region of possible moves which is thrown up onto the hex grid in a colored overlay. I don't know what term you might use for it.

My computer is pretty slow when I play games like Quake 7 and Unreal Tournament 2008, but when playing turn-based, 2d strategy games, it has always been more than fast enough!

Freeciv, for example, runs nicely on this computer. I can scroll the map very responsively in arbitrary distances. In Wesnoth scrolling the map involves obvious redraw and problems with the scrolling continuing for a period after I have pulled back the mouse due to lag.

I suspect Wesnoth simply doesn't handle the "redraws are happening slowly" case very gracefully, although I would think 2d compositing could be done so fast this wouldn't matter. At least, it was achivable back in the early 90s on Amiga and IBM PC systems. I understand wesnoth does use some more expensive effects with the glows which probably involve at least alpha blending if not some kind of realtime generation of the effects, but even with these off, it's still very noticably slow.

[turin]

Most regulars (like me) don't dislike it because of its simplicity, they dislike it because they thing easy mode is easy enough already.

Did you intentionally misunderstand me? I meant, of course, simple in terms of strategy due to the ease of victory. And yes, the easy you think is easy enough is not easy enough.

No, I unintentionally misunderstood you... IMHO you weren't very clear... But, I get what you mean now.

Now, can I ask why it would be preferrable to add easier difficulty levels to current campaigns, instead of adding easier campaigns?

Your guess is wrong, the shadowing is not rendered at all.

How slow is your computer in comparison with other games?

By "shadowing" I am not referring to rendering or any visual effect. I mean the region of possible moves which is thrown up onto the hex grid in a colored overlay. I don't know what term you might use for it.

Telling me my guess is wrong before being sure I meant really rubs me the wrong way.

I means your guess, as in how I understood your guess... obviously if how I understood it was wrong, you can just ignore my post.

My computer is fast enough that I never have problems with speed... however, Wesnoth tries to support most computers, regardless of their speed... since most programmers won't read this thread, if any at all do, you'd probably be best submitting a bug report if you think there is a major problem. But, feel free to ignore my advice.

[k8to]

There is no specific "bug". Wesnoth simply is not streamlined enough to operate pleasantly on my 866mhz c3 cpu. I've submitted performance info when devs have requested it. The numbers speak for themselves, it's obviously bad on this generation of computer, but most people have 1ghz machines or higher and so don't notice.

Given that it's a turn-based app, and so might be interleaved with other activities, and especially given the usage patterns some people put linux machines through, I think it's worth the effort to clean up the CPU load, whether or not it's "fast enough" on most computers. It's quite demonstratably consuming multiple times the CPU cycles of comparable games, and multitasking is useful.

---------
Edit

Now, can I ask why it would be preferrable to add easier difficulty levels to current campaigns, instead of adding easier campaigns?

Hail to the Throne definitely feels like the "wesnoth game", with the alternate scenarios as "bonuses". Given this, I would like to play through the "wesnoth game" without struggling mightily to keep a reasonable number of units alive. Some enjoy trying the same scenario over and over with different strategies, and while I have done so in various smaller games (vandal hearts, ogre battle, etc), somehow in Wesnoth it is just no fun. It is too easy to lose units in a single turn, and too easy to have your 5 attacks on one unit all fail.

I do not feel the problem is really a matter of needing additional scenarios for lame players like me, but a matter of simply needing to tip the capricious odds a bit more in our favor.

But given that I'm certainly not going to invest much in a game that infuriates me experientially, I welcome any sort of lowering of the bar.

[Dave]

There's lots of reasons it could be going slowly and it's unfortunately rather difficult to tell the specific reason.

One thing to note is that writing a cross-platform app that is fast on all platforms it runs on is difficult. It would certainly be alot faster if we had chosen one platform (e.g. X Windows, or OSX, or Microsoft Windows), and optimized for that platform. However, then, we'd have left out all the other platforms, and we didn't consider that worthwhile.

We've had mixed responses from people with slower computers: some say it's surprisingly fast, others complain that it's slow. It can depend on their platform, as well as their specific configuration.

Also, running at a lower resolution should make things faster. You can run at as low as 800x600. (Note that games from the early 90s all ran at resolutions like 320x200, and all at a low colour depth. Wesnoth runs at at least 800x600 at high colour depth. This alone accounts for a reasonably large difference in speed).

You're also right about the shadows: to allow shadows (and other effects) we allow a full 8-bit alpha channel which does slow things down substantially.

David

[k8to]

I'm not really convinced. Freeciv's GTK-based client runs on mac, windows, and UNIX and is snappy on all of them. I think part of this is that they are less ambitious with what they throw up on the screen, but I think the lion's share is that they're using a well established toolkit which knows how to efficiently update the display.

I'm not much for C++, but the poking around I've done suggests wesnoth is rerendering the whole screen a lot of the time, which is definitely going to be slower than GTK which can use a backing store for scrolling a canvas around, etc.

Part of the comparative win is that freeciv is designed around much smaller tiles (for better or worse), but I generaly size the map to fill nearly my entire 1600x1200 screen anyway.

I usually run wesnoth at 800x600. I've tried lower, but the amount of viewable map is so small as to feel cramped at such resolutions. IIRC you can scale the presentation of the graphics, but they look poor at many sizes. There's probably some utilitarian choice that might get by in terms of map and speed for me, but would have seerious aesthetic issues.

As for the 90s comment. Obviously they were compositing into 1/16th the visual area, but they were hundreds of times slower.

[freim]

The comparison to Freeciv isn't very relevant unless you are using the latest dev version since 2.0.x don't use true color RGBA, but indexed mode. I'm working on a tileset for freeciv (amplio) which does use true color RGBA, and with alpha fog and similar stuff it is significantly slower than tilesets which only does simple dithering and use no alpha channel.

[k8to]

The comparison is relevant. Freeciv is able to put up a workable playable game with no lag. Wesnoth does not seem to have this capability.

I am suggesting this would be a good capability to add to wesnoth.

How to achieve this involves issues of interface, technology, features, and optimization. It may not be worth doing! But clearly it is doable because it has been done.

[freim]

Having more capable graphical abilites does of course come with a performance penalty. I tried to explain some reasons why freeciv stable is faster, but it's becoming quite clear to me you're not really bothering to read what people respond to you.

[k8to]

What I don't like is being dismissed. Like you just did. Again.

I brought up freeciv as an example of a high performing cross platform turn based tile based game, like wesnoth. You said that it's irrelevant because there are some technology differences.

What I said was: it's still comparable whether or not there are technology differences. Obviously all applications differ in some way. I acknowledged (and indeed held this position from the outset) that it may not be WORTH changing Wesnoth to have a lower CPU load. And it may not be any of the developer's interest to do so. But to say that the it's irrelevant is dismissive.

Moreover, when i click on a unit in Wesnoth, there is a pause of over a second in length. I do not think this is a fundamental limiation of truecolor. When I launch wesnoth, the entire screen is black for over 5 seconds with no indication of progress. When I start a new game, the screen can lock up for over ten seconds. There are some performance issues.

I do not say that the issues you raise are not real, or do not matter, but I say that the comparison is relevant nonetheless! I am tempted to accuse you of the same as you have me.

[freim]

I've only tried to point out some reasons why performance can be slow. I've never said it's not worthwhile to try to make it faster. I get the feeling you assume I'm against optimization and interpret my responses in that light, even though I've never actually said so.

Being confrontational really isn't going to help your cause much here. I can't imagine any dev is against optimization as long as it's not detrimential to the code otherwise.

As with anything else in a project based on volunters, it will be worked on if anyone wants to.