about using winboard

[gbtami]

"You seem to have grave objections against text as as a handle for discoverability. Why is that?"

Maybe because G=Graphical in GUI?

[Ras]

when you start WinBoard through the Windows start menu item specifying what you want to do

When I click on "winboard.exe", that contraption comes up. I've yet to see another GUI where such nonsense is even possible. See Arena how to do that one right. The items in the Winboard folder of the Windows start menu are overloaded with tons of random stuff anyway. Even the start menu entries are badly designed.

That is "play against an engine", right? It doesn't say "Play against two engines"...

Yeah, so why is the option "play against A chess engine", which is SINGULAR, and then there are two engines? And if singular suggests "A" chess engine, why can't I actually do that but have to choose TWO engines? Do they both play? But chess is a game for TWO players. So with two engines, that suggests one for White and one for Black - leaves me with playing the Ninja pieces or what?

Why the heck would you do that???

In order to get the stupid dialogue to do what it announces, to play against ONE chess engine.

But I guess it would be very easy to make it more idiot-proof here, by simply using the first engine also as second engine, when the user doesn't specify one.

No, the right way would be not to mingle two exclusive options into one radio button selection. Play against A chess engine requires ONE chess engine, matching TWO engines requires TWO engines, and that doesn't make sense.

Never mind that the actually right thing would be to get rid of that dialogue at all. See Arena how to do it - I can match engines, play tournaments or against an engine without dealing with such a nonsense dialogue before startup.

This item is just for deleting or re-ordering the list. Otherwise you should never need it.

Then it doesn't make sense to include it on that menu level. See Arena how to do it right, where you also can reorder the list of installed engines.

Standard text-editing functions are actually a very efficient way to do this.

Only if one has no idea how to implement functionality in a GRAPHIC user interface. Resorting to text editing is unconditional capitulation.

In my experience, people would have no clue what it meant if you provided a menu item "register engine".

Then call it somehow else, but accessing the same functionality (putting in a new engine) via two different menu items is nonsense. On top of that, it makes me wonder why installing a new engine 1 is different from installing a new engine 2 or not.

And please don't try to defend this brokenness. It doesn't get better.

that engines they used before appears in a list

Next broken feature that shows complete lack of understanding even basic GUI principles. This kind of effect should not have to be detected by the user after repeated use. It should have an obvious feedback directly after the action. See Arena how to do that right.

It seems you have been so much exposed to the crummy way it tends to work in other Chess GUIs

LOL. Winboard is the most broken one, and I'm far from the only who tells you this over the years.

that you consider it normal that engines have to be 'registered' before you can use them, and get confused when this silliness is abolished.

It's at least better than clumping two different functionalities into the same dialogue.

And of course they would never paste a FEN.

The G in GUI, what do you think it stands for, hm? For pasting the position via text data? That understanding is so broken that it's no wonder why Winboard is such a bad GUI.

Oh, and "moving around pieces". Did it even occur to you that I may want to "edit" a position by adding new piece, possibly even one of a type that isn't on the board right now? The option to "clear" a board doesn't make any sense if there is no OBVIOUS way how to add any kind of piece.

That is why an editor or something like that would be necessary so that I can drag & drop items. Or maybe select them. Any kind of "tool box".

Funny that you expect anything to open.

Giving user feedback is essential. The "feedback" is a nonsensical text string in a place where it doesn't make sense (the clocks).

That is an interesting point of view. For me, as a user, it is not so much important what I can do with a GUI in the first minute, but rather what I can do in the 20 years after that.

If a GUI comes up with total nonsense like clicking the clock for emptying the board, which has NOTHING to do with each other, than I can expect the rest of the GUI to be the same kind of randomly duct taped hotch-potch. Means, it is disqualified because I'm not willing to spend time to learn a bad GUI.

You can reasonably expect domain knowledge, in this case that the user knows how to play chess. Anything else has to be self-explaining, that is why people like GUIs. Uhm, good GUIs, I mean.

heaven forbid, would require me to look in the man**l

If you have to look up the manual, that's because the GUI is bad. The manual is not the appropriate place to patch up lack of usability.

You write a nice comedy act, but in fact you did discover that you had to click the clock, eve though it was the absolutely crazy thing to do.

That's because I was using Linux as only system over 9 years, and I know that Winboard is an Xboard port, so I could reasonably expect it to be riddled with usability bugs. And I was right, of course.

It seems you are trying to make a mountain out of a mole hill here.

Don't try to defend this garbage. It IS garbage. OK, you can argue that you don't care because the garbage works for you and you don't mind. Which also is why you will keep the garbage as is since anyone who objects could repair it himself, given that Winboard is GPL'ed.

Meanwhile in the real world, people simply instead use GUIs where you can see that they were designed by people who have at least a beginner's idea on how the graphic paradigm is supposed to work.

As I said, it's your right as author to keep Winboard in its bad shape because it works for you. But then don't wonder why people use Arena despite its occasional bugs which are never fixed. It's because Arena is a bit buggy at times, which Winboard isn't - but Winboard is bad by design, which Arena isn't.

You seem to have grave objections against text as as a handle for discoverability. Why is that? Are we supposed to design for dyslectics?

Google what the G in GUI means.

Every time XBoard starts, it prints in large letters over the board "Right-clicking menu item or dialog text pops up help on it".

Also nonsense. GUIs are supposed to lead the user, and by design. This isn't design, this is patching up the lack of proper GUI design.

You know, developing a GUI is hard in two ways. One, the GUI programming itself. That's the programming task. Second, it's the design task.

Xboard is the product of capable programmers without even the faintest competence in GUI design, probably not even knowing what it is and mixing it up with pretty or fancy display effects. Nothing short of a rewrite could have made a Windows version that is actually usable. Not with that bad original version.

[Evert]

Xboard is the product of capable programmers without even the faintest competence in GUI design, probably not even knowing what it is and mixing it up with pretty or fancy display effects.

XBoard was never designed to be a chess GUI, of course. It was made to provide a graphical board for gnuchess, nothing more or less. That's also why there are a bunch of weird idiosyncrasies in CECP. Any functionality additional to that original goal was added later, by a diverse group of programmers.
That it works as well as it does (and it actually works pretty well) after all these years is a miracle.

[Ras]

Any functionality additional to that original goal was added later, by a diverse group of programmers.

Which is also a red flag in itself, see the link I gave earlier in the thread. The resulting mess is exactly what you can expect from that kind of development process. Process becomes product, as an industry saying goes.

And the mess doesn't stop at the GUI; in practice, it has extended into the whole CECP ecosystem. No surprise, either.

[hgm]

"You seem to have grave objections against text as as a handle for discoverability. Why is that?"

Maybe because G=Graphical in GUI?

That seems a fallacious argument. Text is a subset of graphics, and GUIs typically have text-based menus, and print their error messages as text. I don't see anything wrong in that; we are not designing for the MacDonalds here.

Perhaps I should first make a general remark, to prevent that this discussion goes off in a completely wrong and unproductive direction. Let me first say that I very much appreciate any feedback.

But although I would not say that discoverability carries exactly zero weight, it only makes a minute impression on what I consider the quality of a GUI. To 'discover' a GUI takes me perhaps a day. The rest of my life I am just using it, fully aware of how it works, and the dicoverability plays no role at all.

Unfortunately a power/complexity and discoverability are somewhat conflicting requirements. A car obviously has a far lower discoverability than a bicycle, which is why we need driving lessons, and yet many people prefer a car over a bicycle.

So IMO focusing on discoverability and letting it restrict what you can design is 'putting the cart in front of the horse'. In the first place I want a design that works fast and easy for those that know how to use it. Because after the first day everyone will know how to use it, and 99.9% of the use will take place in that situation. Within this boundary condition we can then start to work on the discoverability.

So I am susceptible to arguments like "the explanatory text in the message field between board and clocks is not conspicuous enough, people will not notice it, and consequently get stuck trying to use the feature". A possible solution would be to put that same message on a red background, or make it flash on and off. Or write it diagonally over the board. There certainly exist ways to show it in such a way that they cannot possibly miss it.

But, provided they cannot miss it, I see no real drawback in using this method to provide/enhance discoverability. If the method it would make me discover is convenient and easy, then that is the only thing that I, as a user, would care about.

[hgm]

When I click on "winboard.exe", that contraption comes up.

Going to the installation folder and clicking .exe files there would already be considered 'hacking' by the average Windows user. This is not what they do to start FireFox, MS Word, MS Paint... It seems an awful lot like you use your expert knowledge to seek out situations where you can wreck havoc once you magically turn into an idiot user after the situation is reached.

That being said, I actually use the Startup Dialog very often for non-trivial purposes, (i.e. startig with an engine other than the default, without having to go through the Load Engine menu dialog), so as a user I am very happy that it is there. It allows me to quickly and easily select an engine, and that is a real time saver.

I've yet to see another GUI where such nonsense is even possible.

Well, so one more reason for me to prefer WinBoard over other GUIs. The latter apparently are missing a feature that I often use, and allow me to do the same thing only in a cumbersome way.

See Arena how to do that one right. The items in the Winboard folder of the Windows start menu are overloaded with tons of random stuff anyway. Even the start menu entries are badly designed.

"Tons of random stuff". Not really criticism that I can work with. I did not throw any dice to create menu items; there are items to connect to various ICS, to start in game-viewer mode (normally done by clicking on a PGN file, so not needed very often), and several items to start with a specific engine (in a sub-group). And, yes, one to explicitly summon the Startup Dialog, for those who like it. So far not very random, but carefully chosen as tasks that might be the sole reason why people want to use WinBoard in the first place. Then there is an 'Uninstall' item, which I believe is also pretty common in any software package.

Is your gripe that there also is an entry for a README file? It should be pretty easy to delete menu entries you don't like/need. So I am curous what exactly prompts your dislike here.

[Fulvio]

Let me first say that I very much appreciate any feedback.

It doesn't seem true if you keep defending absurd design choice like clicking on the clock to clear the board.
The problem is well known: programmers have a tendency to think how a feature can be implemented, which contrasts with the GUI's purpose to hide technicalities.

That been said, I think this thread is a little bit ungrateful. It's free software and I'm sure that a lot of time was spent in its development. With limited time and resources it is reasonable to prioritize functionality over easy of use, and indeed the result is that many free software have usability problems. However many people can cope with that, they are free to decide if it's worth spending time to learn how to use it and if the occasional quirks are a deal breaker.

[MikeGL]

Let me first say that I very much appreciate any feedback.

It doesn't seem true if you keep defending absurd design choice like clicking on the clock to clear the board.
The problem is well known: programmers have a tendency to think how a feature can be implemented, which contrasts with the GUI's purpose to hide technicalities.

That been said, I think this thread is a little bit ungrateful. It's free software and I'm sure that a lot of time was spent in its development. With limited time and resources it is reasonable to prioritize functionality over easy of use, and indeed the result is that many free software have usability problems. However many people can cope with that, they are free to decide if it's worth spending time to learn how to use it and if the occasional quirks are a deal breaker.

Also, the source is online. Anyone can just grab the source and
include or remove a feature until it fits their needs.

[hgm]

It doesn't seem true if you keep defending absurd design choice like clicking on the clock to clear the board.
The problem is well known: programmers have a tendency to think how a feature can be implemented, which contrasts with the GUI's purpose to hide technicalities.

But that is a totally wrong diagnosis. This design choice was not made at all because I wouldn't know another way to implement it. I know many other ways to implement it. In fact it was implemented in another way (which can still be activated by changing a configure option), namely by having to right-click the board to summon a context menu, and then select white/black to move there. But this, and most more conventional alternative ways were quite annoying to me as a user.

So it seems this whole discussion already starts on the wrong footing because it considers me a programmer / developer. While I am in fact primarily a WinBoard user, who happens to like easy-to-operate software, and is not afraid to adapt what exists to give him the optimal user experience. I would never have touched WinBoard if I had not needed it for my own purposes.

The problem is that 'absurd design choice' is a rather unhelpful qualification, if it doesn't tell me why this is considered so absurd. 'Innovative', perhaps. But it doesn't sound particularly difficult or cumbersome to have to click the clock. Nor especially taxing that you have to remember that you have to click the clock. And it doesn't suggest any competitive (= single-click) alternative method to do this.

But since people seem to make such a big deal out of this, perhaps the following would be considered an improvement: the clocks are not really needed in Edit Position mode. So instead of displaying the time in the 'clock fields' above the board, they could be made to display the texts 'Clear Board' and 'Set White/Black to move', so that they visibly turn from clocks into buttons.

[Ras]

So instead of displaying the time in the 'clock fields' above the board, they could be made to display the texts 'Clear Board' and 'Set White/Black to move', so that they visibly turn from clocks into buttons.

That would be a starter, yes. However, that doesn't solve the problem that editing a position still is not possible because you can only move pieces, but you cannot add new ones. That is why the Shredder GUI opens an "edit board" along with piece icons at the side. And side to move. And castling rights.

Discoverability, btw., is not in conflict with power usage, that is a cheap pretext. But it's rampant especially in open source where the willingness to suffer from bad user interfaces often is coined into a sign of intelligence. Quite funny, actually.

Let's take keyboard shortcuts, which usually are what users prefer once they are proficient with a GUI. The right way to implement that is to make them discoverable via the GUI. Means, each (important!) menu item should display the shortcut how it could be accessed alternatively. The user will easily remember the commands he uses often while still not be stuck on those he uses rarely.

In the Shredder GUI, I can do several things in analysis mode, e.g. doing multi variant analysis. How to increase/decrease the number of PVs? By hitting F6/F5. How do I know that? Manual? Pffft. Right-click on the PV outputs because right-click usually opens a context menu. That's as always in Windows because even the file explorer has it that way. Among the available options, there is increase/decrease, and they tell also F6/F5.