Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jiri Petru]

A wild idea, sorry for sounding a bit arrogant:

I have the feeling that people here who say "I like the interface" don't actually use it. Which is not my definition of liking something. When people say: "Yeah, it was quite difficult to learn but once I got the hang of it, I like it", they IMHO mean: "Yeah, the interface is difficult to grasp, but once you memorise the commands, you can fortunately ignore it." Not my definition of liking something either.

So a simple question for those of you who say the like the interface: do you really use it? Do you keep the main overwiew menu opened and do you actually read it instead of having the things memorised? When building a workshop or furniture, do you scroll through the unsorted list of items instead of having the shortcuts memorised? When using the stocks screen, do you feel there's something to guide you aside from your own memory ("Okay, so stones are on the third page, about second from the top"). If you answered YES to all these questions then I might believe you do like the interface. But please do not mistake "I memorised it all" for "the interface is good".

[Deathworks]

Hi!

First of all, as was said before, the keyboard controls are also part of the user interface, not only the things you see on the screen. So technically, as long as you play the game, you will be using the user interface.

Secondly, I am probably a rare and amusing case. After understanding the game and figuring out the core controls in 2D, I turned off the overview and the map view so as to get the maximum size for the game view.

However, somewhere during the 3D versions, I changed my mind and switched to using actually the default 1/3 1/3 1/3 setting as it offered me several advantages:
1. I can check on the key commands for the rarer commands. While I know 'k', 't', 'v', and 'q' by heart, the zone controls ('i') are something I use much more rarely, as are the 'R'oom view and the various military things (I prefer peaceful fortresses, so less action there).

2. Actually, the wider game view felt kind of empty. It is hard to describe, but the openness made me feel uncomfortable. Now, with the 1/3 1/3 1/3, I got a comfy "snuggness" to it. Things do not feel so huge and uncontrolled.

3. Probably the initial reason to revert, however, was that I avoid layout changes. That is, when you select certain things, like viewing a tile, if you don't have the overview window there, it will be opened to display the information you are viewing, right? I found that switching back and forth really annoying, and so I rather keep the window opened all the time.

Well, that is all I can think of.

As for the stock lists, I don't really know them by heart; I have a general idea - refuse at the end, so page backwards, bars towards the top on some page, but I don't have memorized them.

Deathworks

[Jiri Petru]

Thanks for your answers, Deathworks.

Now the question is whether you wouldn't benefit from a better organisation of the overwiew menu that would make the rarer commands easier to find. Or from a better organisation of the stocks menu that would order the things using a different logic than pure chaos  Or from a stocks screen that would allow you to have standing orders?

Spoiler (click to show/hide)

(Hope I'm not too aggressive - you seem to be open to argument, which encourages me  )

[Lord Shonus]

My view on the mouse issue.

Right now, DF is very much like DOS. Someone who's familiar with the tree can get anywhere they want in a matter of seconds.

In Windows, at the very least, you add two seconds (the average time for the start menu to pop up) or more to get to anything that lacks a desktop shortcut. If you lack a start menu shortcut, then you use the run command, which can take a while, or you go through the tree in My Computer, which takes comparitively forever.

A mouse with an expert user is inherently slower than a keyboard with an expert user. It's far easier to use in many cases, but you lose a lot in speed.

[Deathworks]

Hi!

Jiri Petru: Well, I wanted to hear what people feel/think about this issue, so I would think it is only natural that I am open to your views (I would not call it argument, as I don't see an aspect of convincing in this).

As for your proposition, please allow me to first clarify that I take your "screenshot" as a schematic, so I will not comment on the font/color scheme etc. (I am one of the classic: black on white for paper, white on black for screen types, so this would get nasty :) :) ). I also would like to point out that it need not be a single screen implementation.

The improvement you suggest for the stocks screen is in my eyes a small improvement (it does only affect the last level and improves a feature that is already there in form of the toggable display of individual items in the stock screen). So, personally, I would welcome that small improvement as the different levels there are indeed an improvement for the user in my eyes.

However, even while I tend to favor your solution on that point, for the sake of fairness, I have to admit that I am currently taking great advantage of the way the lowest level of the stock list is sorted: You see, I always have cats in my fortresses, so thanks to the vermin bug, I have a constant influx of non-rotting vermin remains that I need to dump periodically. For that task the current system is perfectly set up, because at the lowest level, the remains are first divided into free and non-free, and then into individual item types; so, I first get all the toad remains lying around, then all the lizard remains, then all squirrel remains, and so on, until I get to all the forbidden/to be dumped toad remains, forbidden/to be dumped lizard remains, ... This means, for my regular dumping of the trash, I only have to go to the lowest level under remains and can then set everything to dump until I come across something that is already set so or is forbidden. Still, I think that this singular benefit does not outweigh the general benefit I personally see in your approach.

As for the overview, I am more sceptical. The stock screen has the great advantage that the divisions are obvious/pre-defined. There is already furniture as a category, so it makes sense to have it as a category. And the different furniture types followed by the different materials also makes sense especially taking the limitations on beds, bins, and coffins into account. In most cases, the furniture type is more important for the user, I would presume, so having materials as the lowest level makes sense.

The overview is a completely different beast altogether. I think there are possibly different views of how things relate to each other and also different needs of players. For instance, I have deliberately not mentioned the burrows since they are still too new a feature for me. But depending on your playing style, which commands are rare and which are not rare changes as does what exactly you need. So, I think that the overview should probably try to be open to many different approaches, which of course limits the refinement.

And just to avoid misunderstandings - I said I was open to small improvements and twitches here and there, I just don't feel them absolutely necessary and I would not want to change the interface fundamentally. No matter how much sugar-coating you put on it, the moment you suggest pull-down menues, you have definitely left my limits.

Deathworks

[Exponent]

A mouse with an expert user is inherently slower than a keyboard with an expert user. It's far easier to use in many cases, but you lose a lot in speed.

For some actions, this may be true, but since it isn't true for all actions (or necessarily even most of them), "inherent" seems to be the wrong word in my opinion.  Consider Wolfenstein 3D versus any modern FPS.  Keyboard controls are awful for manipulating a camera.  Or selecting a few particular items from a fairly large list is far easier with a scroll wheel and a few mouse clicks than it is with numerous down-arrow presses and a few ctrl+space presses.  Moving around a 2D map can in many contexts be easier with a mouse, whether it be edge-scrolling, click and drag (the distance from the click point determining scroll speed), or especially use of a mini-map.  Browsing the internet with just a keyboard is one of the most painful things I've ever done.

Besides, as I noted before, use of a mouse inherently provides context about which control the user wishes to manipulate simply due to the location of the cursor, whereas the keyboard inherently applies to the whole screen at once, which means that 1) it isn't immediately clear which screen element key presses will be routed to (they might not even be routed to a screen element at all, but some "invisible" element), and B) it becomes far more complicated to allow routing the key presses to different targets depending upon what the user wants.  Every single screen in essence becomes a monolithic control, which I would imagine makes use, design, and modification harder.

[Jiri Petru]

As for the speed issue: I don't really play games to be fast and effective. Have enough of that in my job. I am willing to "suffer" interface that is slower if it is more comfortable to use.

Also, it depends. Assigning labours by clicking in Dwarf Therapist is waaaaay faster and more effective than assigning labours by keyboard in DF.

[Deathworks]

Hi!

First of all, I want to agree with Jitri Petru that the speed issue is not that elemental to me either. Of course, if things run smoothly, they are fun, but that is not a major aspect.

For some actions, this may be true, but since it isn't true for all actions (or necessarily even most of them), "inherent" seems to be the wrong word in my opinion.

I agree with you there. Just as with the wooden training axes, they are superior when used in training but inferior when facing that goblin in iron full body armor: You want the right tool for the right job. (The dwarves happily using the training weapons when going to war does NOT seem to be a user interface issue as far as I have followed the reports/comments on that).

However, the job is defined in part by the user as well, and this is why I consider this entire discussion to be subjective. Some people don't have the patience to wait, and others want more time to think. And accordingly, they focus on different aspects of the user interface, and this may result in them favoring different solutions.

Moving around a 2D map can in many contexts be easier with a mouse, whether it be edge-scrolling, click and drag (the distance from the click point determining scroll speed), or especially use of a mini-map.

While I am fine with your general argument, I have to confess that I find that example extremely unfortunate: Map scrolling with the mouse, while maybe in theory a nice thing, seems to be implemented only in the most awful fashions possible. Whether it is an aggressive edge scrolling that does not allow you to click on the icon bar at the bottom of the screen, or whether the edge scrolling has so much resistance that you push and push until it jumps to the other corner of the map, or whether you get a tennis arm from moving the mouse back and forth to scroll using drag and drop because none of the possible settings for speed or sensitivity allow you to really enjoy this procedure, in my experience, all mouse based map scrolling has been a horrible nightmare. Merely jumping via the mini map or if you scroll one screen at a time by confirming with clicking have worked out for me personally. But the latter approach does not yield any advantage over simply pressing SHIFT + cursor keys as Dwarf Fortress currently does for map scrolling.

Mind you, I agree that there are things the mouse can do better, as there are those the keyboard can do better. I just needed to point out that you might wish to choose an example with a smaller subjective component (^_^;;

Besides, as I noted before, use of a mouse inherently provides context about which control the user wishes to manipulate simply due to the location of the cursor, whereas the keyboard inherently applies to the whole screen at once, which means that 1) it isn't immediately clear which screen element key presses will be routed to (they might not even be routed to a screen element at all, but some "invisible" element), and B) it becomes far more complicated to allow routing the key presses to different targets depending upon what the user wants.  Every single screen in essence becomes a monolithic control, which I would imagine makes use, design, and modification harder.

Actually, that is more of an implementation question. Using marking, you can have very similar effects using the keyboard. When the subwindow or element that is currently being targeted is clearly marked, for example by a red frame (just a spontaneous idea), you can convey that information just as well as you can with the mouse. The web browser you referred to before also uses the "ants" to mark which element is currently subject to your next key press.

And even beyond that, I do not really see how the unity of the screen, so to speak, would inherently be more difficult to use or design. Regardless of what approach you choose, you have to think about how you implement or it will fail horribly.

At least that is the way I see it.

Deathworks

P.S.: Picking up the comment by Jiri Petru, I want to mention that I am not using any outside tools while playing dwarf fortress (the game is already labelled as unstable, so I rather not mess around with it even further), the only outside tool I ever use being the map compressor for uploading the maps at the map archive.

[lordnincompoop]

Since there seems to be quite a few people interested in the interface, let me use this post to invite you to a topic I created months ago in the suggestion forum: Total Interface Overhaul. It was meant as a repository of constructive ideas and graphical proposals for Toady to choose from. It's been going quite slowly (and I myself am mostly ignoring it), but there have fortunately been contributors who kept it alive (most notably Zwei). So if you're interested in interface and like to come up with solutions, we could use you 

NOTE: That topic should if possible remain practical, so let's keep all arguing here. That being said...

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

It's hard for me to believe someone could actually like the current interface. A part of me want to say you'd have to either (a) not understand what an interface is, or (b) never have seen a good interface. No insult meant, It's just hard for me to grasp.

A good example is the main menu:


It's basically just a mess, a confusing wall-of-text. A good interface would in the least visually differentiate between the most important (often used) features and the minor ones. There's nothing like it in DF. No visual distinction, not even any semblance of order. The features are listed alphabetically by the keyboard shortcut, for Christ's sake! I can hardly imagine anything more useless. The menu is unusable - you have to memorise it and then hide it, you can't orientate yourself by it.

A functional DF main menu would at least:
- Group features according to frequency of use and function. Eg: (1) Buildings, Designations, Stockpiles, Zones; (2) View unit, View items in building, Use buildings, (3) Military, Squads, Routes, Burrows, etc.
- Break the wall of text! Differentiate between the functional groups using font size, colour, blank spaces. Distribute them along the whole screen. Anything...
- Reduce the visual clutter! Hide the more obscure functions in submenus. Artifacts and civilisations could go into Stocks menu. (why are they here in the first place, when things like animals or kitchen - which you use more often - are hidden?) Nobles could hide under units, or whatever. Depot acces should be accesible wrom the depot building screen, no need to have it here. Options could easily fit into the (Esc) menu, like in every other game.
- etc. etc.

There are just examples I made up while writing this post! The current interface is completely illogical, there seems to be next to no thought put into its ergonomy or whatever. I really, really can't believe you like it more than for example my half-assed proposal here.

(Also, please note the lack of any mouse support or graphics stuff in this post.)

---

ALSO, before a misunderstanding, because I expect reactions along these lines: you can hide a Tooltip in a submenu and still have a shortcut for it. You can hide the text "(l) artifacts" in the stocks menu and still have the shortcut (l) functional in the main game. Most software works like this, just look for example at your browser. Is there a visual clutter? No. Is there a visual distinction of more and less important controls? Yes.

What? I can read it just fine. Perfectly usable, and even when I was starting out it was fine.

[Exponent]

Moving around a 2D map can in many contexts be easier with a mouse, whether it be edge-scrolling, click and drag (the distance from the click point determining scroll speed), or especially use of a mini-map.

While I am fine with your general argument, I have to confess that I find that example extremely unfortunate: Map scrolling with the mouse, while maybe in theory a nice thing, seems to be implemented only in the most awful fashions possible...

True, I've seen some horrible mouse-based map scrolling interfaces.  I should have just eliminated that example or picked a different one.

Besides, as I noted before, use of a mouse inherently provides context about which control the user wishes to manipulate simply due to the location of the cursor, whereas the keyboard inherently applies to the whole screen at once...

Actually, that is more of an implementation question. Using marking, you can have very similar effects using the keyboard. When the subwindow or element that is currently being targeted is clearly marked, for example by a red frame (just a spontaneous idea), you can convey that information just as well as you can with the mouse. The web browser you referred to before also uses the "ants" to mark which element is currently subject to your next key press.

Having the visual feedback helps, but changing the focus is still difficult.  Sometimes all that is available is a linear tab-key style selection.  Other times, arrows can sometimes be used, though even then it isn't always easy to select the element you want; you press down-arrow, but there are two nearby elements below the current one, and it goes down and slightly left instead of down and slightly right, for example.  (Not to mention you have now lost the ability to use arrow keys to manipulate the element itself.)  Either way, you will often end up repeatedly pressing a keyboard key to move to the element you want, whereas with a mouse you'd just quickly move the mouse right to it and click (or simply hover, if all you need is info in a tooltip).

And even beyond that, I do not really see how the unity of the screen, so to speak, would inherently be more difficult to use or design. Regardless of what approach you choose, you have to think about how you implement or it will fail horribly.

At least that is the way I see it.

When the screen is a unified whole, any change to any part of it needs to take into account everything else on the screen.  If a feature or option is added to perform a specific task, then one has to make sure that the controls used to initiate and control that task don't interfere with already existing features.  Adding search to various screens has been a good example of how this is a problem.  With the addition of search, just about every key on the keyboard becomes valid for standard typing, and can no longer be used for commands.  Thus we get funny keys to cancel out or confirm, raising complaints of inconsistency.

Additionally, it becomes harder to reuse components when every screen is built as a whole.  It might be useful to display the same type of information in multiple screens.  But pulling a part of one screen out and adding it to another is not always simple.  Maybe some component uses a certain set of keys on Screen A without problem, but on Screen B, some of those keys are already used for something else.  The options at that point are to 1) rework Screen A and Screen B so that the shared component can use the same set of keys on both screens without conflict (possibly a lot of work, and will throw people off because the interface changed), 2) make the component use a different set of keys on Screen B (leading to that dreaded inconsistency), or 3) deciding that the shared component would only be nice but not necessary on Screen B and therefore leave it off because it's too much hassle.  I suspect option 3 gets taken quite often, to the point that many opportunities for providing useful information in various parts of the game aren't even considered due to a subconscious recognition that doing so would simply take too much effort.

Heh, unfortunately, I haven't played the game at all in the last year and a half, so all the specific experiences that caused me to form these general thoughts have left my memory, leaving just the general ideas.  Which I realize makes my argument less persuasive.  But I don't think I'm saying anything revolutionary; it's a common design principle in engineering to design stuff such that it can be broken down to a reasonable level, and each individual component can be largely worked on in isolation from all the other components.

[Goron]

Since there seems to be quite a few people interested in the interface, let me use this post to invite you to a topic I created months ago in the suggestion forum: Total Interface Overhaul. It was meant as a repository of constructive ideas and graphical proposals for Toady to choose from. It's been going quite slowly (and I myself am mostly ignoring it), but there have fortunately been contributors who kept it alive (most notably Zwei). So if you're interested in interface and like to come up with solutions, we could use you 

NOTE: That topic should if possible remain practical, so let's keep all arguing here. That being said...

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

It's hard for me to believe someone could actually like the current interface. A part of me want to say you'd have to either (a) not understand what an interface is, or (b) never have seen a good interface. No insult meant, It's just hard for me to grasp.

A good example is the main menu:


It's basically just a mess, a confusing wall-of-text. A good interface would in the least visually differentiate between the most important (often used) features and the minor ones. There's nothing like it in DF. No visual distinction, not even any semblance of order. The features are listed alphabetically by the keyboard shortcut, for Christ's sake! I can hardly imagine anything more useless. The menu is unusable - you have to memorise it and then hide it, you can't orientate yourself by it.

A functional DF main menu would at least:
- Group features according to frequency of use and function. Eg: (1) Buildings, Designations, Stockpiles, Zones; (2) View unit, View items in building, Use buildings, (3) Military, Squads, Routes, Burrows, etc.
- Break the wall of text! Differentiate between the functional groups using font size, colour, blank spaces. Distribute them along the whole screen. Anything...
- Reduce the visual clutter! Hide the more obscure functions in submenus. Artifacts and civilisations could go into Stocks menu. (why are they here in the first place, when things like animals or kitchen - which you use more often - are hidden?) Nobles could hide under units, or whatever. Depot acces should be accesible wrom the depot building screen, no need to have it here. Options could easily fit into the (Esc) menu, like in every other game.
- etc. etc.

There are just examples I made up while writing this post! The current interface is completely illogical, there seems to be next to no thought put into its ergonomy or whatever. I really, really can't believe you like it more than for example my half-assed proposal here.

(Also, please note the lack of any mouse support or graphics stuff in this post.)

---

ALSO, before a misunderstanding, because I expect reactions along these lines: you can hide a Tooltip in a submenu and still have a shortcut for it. You can hide the text "(l) artifacts" in the stocks menu and still have the shortcut (l) functional in the main game. Most software works like this, just look for example at your browser. Is there a visual clutter? No. Is there a visual distinction of more and less important controls? Yes.

What? I can read it just fine. Perfectly usable, and even when I was starting out it was fine.

You can read it, yea, but not as well as if it was better organized.
Suppose you had to designate every single tile one at a time that you wanted to be mined out, would you like that? Would you say "whats wrong with it? I can do it, its perfectly usable, and even when I was starting out it was fine!" Sure, thats all true but that still doesn't make it reasonable.

I find many people here seem to think that criticizing an interface element is akin to criticizing a gameplay element. They are not the same. It is OK to admit something could be done better. Thats why so many people use Dwarf Therapist, for example. It presents a much better method of completing in-game actions. Do you not agree that therapist is a nice utility to use? Well news flash! Therapist is the result of a poor DF interface! It makes doing things easier and better. Same goes for any of the designation macro apps.
Did the introduction of a zoomable interface ruin DF? Newsflash: thats a usability enhancement!
Did the introduction of an on-the-fly re-sizable screen ruin the game? Newsflash! interface enhancement!

[Deathworks]

Hi!

Goron: I do agree that gameplay elements and user interface seem to get mixed up in the discussions, together with the bugs. And that mix-up does not really help as comments from both camps then get confusing.

However, the user interface does in any case influence the game experience, so changing it does have great consequences for the perception of the player.

And just for an aside before someone tries to misuse this: Yes, it is of course theoretically possible to make the gameplay elements dependent on what the chosen user interface can support, but I don't think we have to worry about Toady One ever going in that direction. So, that is also not a relevant issue, I think.

As for the 40d# elements, I have to admit that I have not tested any of the 40d# versions because their update cycle was too fast for me at the time, and I was a bit wary given how useless and broken partial print had been for me. So, I can't really comment on any details there, especially whether these are really enhancements for me (I just installed Heroes of Might and Magic V on my other computer, and the 3D graphics during the adventure/map mode are driving me crazy. I have to see if I can somehow fix that - I hope there is some option or something).

:) :) :)

But you could have also pointed out that the tileset had corrupted the '.' next to One-Step :) :) :) :)

Having the visual feedback helps, but changing the focus is still difficult.  Sometimes all that is available is a linear tab-key style selection.  Other times, arrows can sometimes be used, though even then it isn't always easy to select the element you want; you press down-arrow, but there are two nearby elements below the current one, and it goes down and slightly left instead of down and slightly right, for example.  (Not to mention you have now lost the ability to use arrow keys to manipulate the element itself.)  Either way, you will often end up repeatedly pressing a keyboard key to move to the element you want, whereas with a mouse you'd just quickly move the mouse right to it and click (or simply hover, if all you need is info in a tooltip).

First of all, let me re-iterate that I am not trying to convert anyone. I am just describing the things that come to my mind when reading these things.

Dwarf Fortress being a tile-based game, we are at least spared the two units too close to select separately syndrome with the mouse. However, Dwarf Fortress does pose its own implementation problem by allowing units to be stacked quite liberally. Personally, I find the paging through the units using 'v' quite easy as I have an entire third of the screen giving me information on the currently selected unit. And I didn't need to move the mouse cursor to some additional control element or anything.

As for the "loss" of keys - that is really a very weak argument. I mean, you got 26 letters, 10 numbers, some more things like '.' and ',', each of which can be combined with shift and control, plus a number pad plus the cursor keys and a few extra function keys. Dividing keys up in a consistent and elegant manner should not be impossible under such circumstances.

Heh, unfortunately, I haven't played the game at all in the last year and a half, so all the specific experiences that caused me to form these general thoughts have left my memory, leaving just the general ideas.  Which I realize makes my argument less persuasive.  But I don't think I'm saying anything revolutionary; it's a common design principle in engineering to design stuff such that it can be broken down to a reasonable level, and each individual component can be largely worked on in isolation from all the other components.

Well, breaking down a problem into smaller bits is basically the definition of programming I learned, so that is not something I can argue with.

However, I am not fully convinced that the impact of the unity of the screen is so great. Modern high level languages like C++ offer quite a lot of possibilities to intelligently copy units and modify them, so that with some thought, it might be that much of a problem.

Deathworks

[The_Fool76]

Reading this whole discussion (phew) has made me realize there is one thing that would help the interface more than anything else.

Pulling it out of the binary and making it so we can edit it. 
Say something like:

<screen name="build">
<menu name="furniture" key="f">
<menu_item name="chair" key="c" action="build_chair" icon="chair_tile" />
</menu>
</screen>

[Aklyon]

Reading this whole discussion (phew) has made me realize there is one thing that would help the interface more than anything else.

Pulling it out of the binary and making it so we can edit it. 
Say something like:

<screen name="build">
<menu name="furniture" key="f">
<menu_item name="chair" key="c" action="build_chair" icon="chair_tile" />
</menu>
</screen>

Thats in the Eternal Suggestions. its #16, 145 votes.

[ManaUser]

So a simple question for those of you who say the like the interface: do you really use it? Do you keep the main overwiew menu opened and do you actually read it instead of having the things memorised? When building a workshop or furniture, do you scroll through the unsorted list of items instead of having the shortcuts memorised? When using the stocks screen, do you feel there's something to guide you aside from your own memory ("Okay, so stones are on the third page, about second from the top"). If you answered YES to all these questions then I might believe you do like the interface. But please do not mistake "I memorised it all" for "the interface is good".

To my mind, keyboard commands are part of the "interface". And yes clearly the game was designed with the expectation that you will eventually memorize commonly used commands, and would be rather horrible if you couldn't. Some parts of the interface have more two-way communication like the stocks screen, but alot of the menus really are more like "cheat sheets" of commands, to be ignored once you know them.

But I'm not saying that as any kind of value-judgment. That doesn't mean it's not a "real interface", just a different style. And again, personally I don't have any problem with it. It's actually the more interactive parts like the custom stockpile screen and of course the new military stuff that I tend to find confusing.