Crafty UCI version

[hgm]

Well, what you can do is pester ChessBase for the defects in their 'industry standard' product, rather than engine authors. I am sure they will be much more forthcoming to paying customers than authors of free engines would be to let them to work around defects in commercial GUIs.

But perhaps you should elaborate on your animosity towards adapters. Because adapters is just a technology, and objecting against them seems irrationality at the level where one insists on a Crafty rewrite in C++, because as a matter of principle you don't want to run programs that are written in plain C. What a user cannot notice, is not his business to care about.

Suppose you could download a zip file that contained a Crafty.exe, a Crafty_UCI.exe, a wb2uci.ini, crafty.rc and some other files (opening book?), and you could run Crafty in your crummy ChessBase GUI by 'installing' (actually: registering) Crafty_UCI.exe like any other UCI engine. Would you be unhappy? Well, tough luck for you: Crafty_UCI.exe, although UCI, happens to be not an engine, but an adapter, that (without you knowing it) will start Crafty.exe to ply the actual Chess.

Why would that be unacceptable to you?

[chessico]

Well, what you can do is pester ChessBase for the defects in their 'industry standard' product, rather than engine authors. I am sure they will be much more forthcoming to paying customers than authors of free engines would be to let them to work around defects in commercial GUIs.

But perhaps you should elaborate on your animosity towards adapters. Because adapters is just a technology, and objecting against them seems irrationality at the level where one insists on a Crafty rewrite in C++, because as a matter of principle you don't want to run programs that are written in plain C. What a user cannot notice, is not his business to care about.

Suppose you could download a zip file that contained a Crafty.exe, a Crafty_UCI.exe, a wb2uci.ini, crafty.rc and some other files (opening book?), and you could run Crafty in your crummy ChessBase GUI by 'installing' (actually: registering) Crafty_UCI.exe like any other UCI engine. Would you be unhappy? Well, tough luck for you: Crafty_UCI.exe, although UCI, happens to be not an engine, but an adapter, that (without you knowing it) will start Crafty.exe to ply the actual Chess.

Why would that be unacceptable to you?

I just tried to bring a real world voice into the discussion. I don't have any animosities against adapters. Where do you take that from? The industry standard may have as many defects as you can find, people in the real world won't care and won't start using winboard. They also use windows, no matter how much I would like them to change to linux.
I have used all the available adapters, geek solutions, config files or whatever in the past; I am even a bad programmer and experienced user of all kinds of software, so I could figure out and make viable whatever I want.
But like I said, my geeky days are over. I knew all the little parameters of crafty by heart; then they were changed, and I did not bother to learn the new set. I am also aware of your successful efforts to improve on winboard, thx for it. But most chessplayers are not as knowledgable and it's up to the programmers to make the decision: Do they want their product to be used or ignored? Well, of course I know the answer, but I wanted to make the point, nevertheless.

[hgm]

I don't have any animosities against adapters. Where do you take that from?

Well, in your previous post you wrote: "I certainly won't use adapters"...

The industry standard may have as many defects as you can find, people in the real world won't care and won't start using winboard.

No, and I was not asking them to. But at least they should aim their grievances against the makers of the defective GUIs, rather than against engine programmers.

I have used all the available adapters, geek solutions, config files or whatever in the past; I am even a bad programmer and experienced user of all kinds of software, so I could figure out and make viable whatever I want.
But like I said, my geeky days are over. I knew all the little parameters of crafty by heart; then they were changed, and I did not bother to learn the new set. I am also aware of your successful efforts to improve on winboard, thx for it. But most chessplayers are not as knowledgable and it's up to the programmers to make the decision: Do they want their product to be used or ignored? Well, of course I know the answer, but I wanted to make the point, nevertheless.

The problem is that the discussion about technical issues here is polluted by a discussion of who should do what. These are really separate issues. You make the assumption here that using an adapter unavoidably means you have to install and configure it yourself.

But that is a fallacy. To run my engine Joker under the ChessBase GUI it would be necessary to invoke WB2UCI, with some .eng (?) settings file. Geeky users would be able to figure out how to do this themselves. But if I wanted to make life easy for the digibetic masses, I could bundle (a renamed) WB2UCI with the required .eng file and the .exe of the (WB) engine. They would use the disguised WB2UCI as if it was a UCI version of Joker, and would not have the technical expertise to ever figure out that they are really running a WB engine through an adapter.

And if I, as engine author, are not interested to bundle Joker that way, someone else could do it, and host the bundled Joker_UCI package on his website. This would be a hell of a lot easier for them than to convert Joker to UCI (even if they could get the source code for it).

So it seems that the dominant part of the reasoning here is: "If I run WB engine XXX through WB2UCI, this is cumbersome, but still not so difficult that I couldn't conceivable do it myself if I tried hard. So it is better to require XXX is converted to a UCI engine, because that is such a big and complex job that everyone understands I cannot do it myself, so all the work has to be done by someone else." Well, that the user does not want to do any extra work just because his GUI supplier also did not want to do it (i.e. bundle the required adapters with the GUI, as I do for WinBoard), and wants to dump the extra work on a random volunteer might be natural. But that still doesn't justify making the job far more elaborate than needed.

In WinBoard we have of course had the opposite situation, where it required cumbersome and geeky editing of ini files to run UCI engines in it. This was remedied by improving both the Polyglot adapter, and to a lesser extent WinBoard, so that UCI engines now run just as easily as WB engines in WinBoard, without the user ever knowing he was running through an adapter. So it is not exactly science fiction. Transparent usage of WB engines in UCI-only GUIs should be just as easy. And to achieve it would be easier than converting a single WB engine to UCI.

Of course if the engine doesn't really do what you want even in a WB-capable GUI, (like not supporting multi-PV in analysis mode), that is an entirely different issue still. It would require extensive modification of the core of the engine. You might as well require that it was 300 Elo stronger. Wouldn't it be great if someone would make a version of Crafty for us that was 3400 Elo? But I think it is highly unrealistic to expect that anyone but the author would be able and willing to fix such things.

[Roger Brown]

I just tried to bring a real world voice into the discussion. I don't have any animosities against adapters. Where do you take that from?

Perhaps when you said this...

In these days, after my geeky days are over, I certainly won't use an adapter, however well written it may be, in order to use an engine. Most chessplayers I know would not do such a thing.

The industry standard may have as many defects as you can find, people in the real world won't care and won't start using winboard.

This is fascinating. This industry standard that you point to and support sounds incredibly arrogant to me. You are a paying customer, among many, non-geeky chessplayers who would like to use Winboard engines in your Chessbase guis. Yet that company refuses to support your likes even though you pay for their product.

You will support them, defects and all, rather than a "geeky" product which is responsive to user needs.

Free guis such as ChessGUI, Arena and yes, Winboard, support both protocols effortlessly. Their developers receive not a cent from their users. Chessbase seems to be ignoring what their own customers want.

The customers of Chessbase then turn around and expect engine authors who contribute their efforts for free to write their engines to suit the gui. A many to one solution sounds logically more painful than a one to many.

They also use windows, no matter how much I would like them to change to linux.

Not so sure about this comment and I would not want to open up this can...

I have used all the available adapters, geek solutions, config files or whatever in the past; I am even a bad programmer and experienced user of all kinds of software, so I could figure out and make viable whatever I want.

What if you had no idea you were interacting with an adapter? Fabien is rightly recognised for Fruit but in my opinion his true genius is in coding polyglot. In Winboard, you have no idea that you are interacting with it as you simply tick a box. I have installed engines in several guis, including your industry standard and believe me, it is not as geeky as it once was. Arena same. ChessGUI same.

But like I said, my geeky days are over. I knew all the little parameters of crafty by heart; then they were changed, and I did not bother to learn the new set. I am also aware of your successful efforts to improve on winboard, thx for it. But most chessplayers are not as knowledgable and it's up to the programmers to make the decision: Do they want their product to be used or ignored? Well, of course I know the answer, but I wanted to make the point, nevertheless.

I appreciate your perspective. I just think that asking many to suit one is far more painful for everyone concerned rather than demanding that your commercial gui support over a hundred chess engines. One change and you could get over a hundred more engines for your enjoyment.

I think that is worth some agitation.



Later.

[chessico]

I think that is worth some agitation.

Agitation is certainly what I am least interested in. I was speaking from a practical point of view, of the average user, agitated fighting for ideologies is what I wholeheartedly leave to the discussion that has been going on here for several pages already. Have fun with it!

[chessico]

I don't have any animosities against adapters. Where do you take that from?

Well, in your previous post you wrote: "I certainly won't use adapters"...

The industry standard may have as many defects as you can find, people in the real world won't care and won't start using winboard.

No, and I was not asking them to. But at least they should aim their grievances against the makers of the defective GUIs, rather than against engine programmers.

I have used all the available adapters, geek solutions, config files or whatever in the past; I am even a bad programmer and experienced user of all kinds of software, so I could figure out and make viable whatever I want.
But like I said, my geeky days are over. I knew all the little parameters of crafty by heart; then they were changed, and I did not bother to learn the new set. I am also aware of your successful efforts to improve on winboard, thx for it. But most chessplayers are not as knowledgable and it's up to the programmers to make the decision: Do they want their product to be used or ignored? Well, of course I know the answer, but I wanted to make the point, nevertheless.

The problem is that the discussion about technical issues here is polluted by a discussion of who should do what. These are really separate issues. You make the assumption here that using an adapter unavoidably means you have to install and configure it yourself.

But that is a fallacy. To run my engine Joker under the ChessBase GUI it would be necessary to invoke WB2UCI, with some .eng (?) settings file. Geeky users would be able to figure out how to do this themselves. But if I wanted to make life easy for the digibetic masses, I could bundle (a renamed) WB2UCI with the required .eng file and the .exe of the (WB) engine. They would use the disguised WB2UCI as if it was a UCI version of Joker, and would not have the technical expertise to ever figure out that they are really running a WB engine through an adapter.

And if I, as engine author, are not interested to bundle Joker that way, someone else could do it, and host the bundled Joker_UCI package on his website. This would be a hell of a lot easier for them than to convert Joker to UCI (even if they could get the source code for it).

So it seems that the dominant part of the reasoning here is: "If I run WB engine XXX through WB2UCI, this is cumbersome, but still not so difficult that I couldn't conceivable do it myself if I tried hard. So it is better to require XXX is converted to a UCI engine, because that is such a big and complex job that everyone understands I cannot do it myself, so all the work has to be done by someone else." Well, that the user does not want to do any extra work just because his GUI supplier also did not want to do it (i.e. bundle the required adapters with the GUI, as I do for WinBoard), and wants to dump the extra work on a random volunteer might be natural. But that still doesn't justify making the job far more elaborate than needed.

In WinBoard we have of course had the opposite situation, where it required cumbersome and geeky editing of ini files to run UCI engines in it. This was remedied by improving both the Polyglot adapter, and to a lesser extent WinBoard, so that UCI engines now run just as easily as WB engines in WinBoard, without the user ever knowing he was running through an adapter. So it is not exactly science fiction. Transparent usage of WB engines in UCI-only GUIs should be just as easy. And to achieve it would be easier than converting a single WB engine to UCI.

Of course if the engine doesn't really do what you want even in a WB-capable GUI, (like not supporting multi-PV in analysis mode), that is an entirely different issue still. It would require extensive modification of the core of the engine. You might as well require that it was 300 Elo stronger. Wouldn't it be great if someone would make a version of Crafty for us that was 3400 Elo? But I think it is highly unrealistic to expect that anyone but the author would be able and willing to fix such things.

Yes, I agree with most of what you say, as a matter of fact.
From the point of view of the tech noob it would have to be a transparent solution, and I agree, an adapter used without the inexperienced user knowing it is quite acceptable. One level of complexity, where something can go wrong, more than necessary, but in most cases it should make no difference as regards usability.

[hgm]

Free guis such as ChessGUI, Arena and yes, Winboard, support both protocols effortlessly. Their developers receive not a cent from their users. Chessbase seems to be ignoring what their own customers want.

This is not as strange as it seems. ChessBase has no interest at all in people using their GUI with engines other than their own. It seems it has been a hard fight even making them support UCI, to which they gave in only very reluctantly. Their attitude is: "if you want to use an engine, buy Fritz!"...

I appreciate your perspective. I just think that asking many to suit one is far more painful for everyone concerned rather than demanding that your commercial gui support over a hundred chess engines. One change and you could get over a hundred more engines for your enjoyment.

Indeed, this is really the point. The solution to provide a UCI 'version' of every engine by bundling a pre-configured adapter with it, (assuming a sufficiently high-quality adapter exists), although far easier then converting engines to other protocols, is still far more cumbersome globally than bundling an adapter with the GUI. I really see this as a task of the GUI supplier.

Note that there never are any requests for converting UCI engines to WB protocol, by users of WB-only GUIs.

[chessico]

is still far more cumbersome globally than bundling an adapter with the GUI. I really see this as a task of the GUI supplier.

Very true, of course. But neither will CB support winboard, (they might have to exclude the feature from support), nor will Bob support uci. And god is not a woman. Although she could be, if only she wanted.

[Roger Brown]

Indeed, this is really the point. The solution to provide a UCI 'version' of every engine by bundling a pre-configured adapter with it, (assuming a sufficiently high-quality adapter exists), although far easier then converting engines to other protocols, is still far more cumbersome globally than bundling an adapter with the GUI. I really see this as a task of the GUI supplier.

Note that there never are any requests for converting UCI engines to WB protocol, by users of WB-only GUIs.

Interesting, that point in bold...

Could it be that the poster is correct and that as geeks only use WB protocol, that they solve their problems in a manner that involves the least pain for everyone?

Fabien has been hailed for Fruit when polyglot was his invention I use the most.

Perhaps one day he might be persuaded to something like that for wb2UCI...

Later.

[Roger Brown]

I think that is worth some agitation.

Agitation is certainly what I am least interested in. I was speaking from a practical point of view, of the average user, agitated fighting for ideologies is what I wholeheartedly leave to the discussion that has been going on here for several pages already. Have fun with it!

Hmmmmm, you really missed the point of what I wrote it seems.

Indeed, if you perceive an ideological battle, it is not coming from me or from most of the posters in the thread.

And yes, I am having fun.

Later.