WinBoard, exotic version

[Daniel Shawul]

i haven't uploaded the new one. I will do so now.
Edit: Uploaded now

c1x1kxli/pp1ppppp/1w6/8/1P1P1l2/8/P1P2PPP/ILXKWXLC w - - 1 4
It works but square a3 goes off on a hover on a red square. Probably a problem with merging.

[hgm]

Indeed, there was a bug left, that still cleared the first empty square of any numeric stretch. I uploaded a WinBoard-Alien.zip now that fixes it.

I am still not very happy with the way victims are highlighted on Long Leaper moves. E.g.

1lxwkxli/p2ppppp/7P/2pcW3/1P6/6P1/1PPP1P2/ILXK1XLC b - - 5 6
[d]1nbqkbnr/p2ppppp/7P/2prQ3/1P6/6P1/1PPP1P2/RNBK1BNR b - - 5 6

When I hover Lb8 over f4, Pb4 and Pg3 remain green, while they are not capture victims of that move.

[Daniel Shawul]

Yes that is not good. But for that I need to make intermediate moves, recalculate highlight and send it to winboard Or use the method I discuss below. In that case we will not need superposition of highlights. For checkers this can introduce highlighting problems because if you recalculate the highlight while being on the second leg, you could add additional branches that don't exist on the first leg move..

I will try implementing a scheme which generates the moves once, and then narrows down the list when a leg is being made. More like what I proposed winboard should do when we started this discussion. This is the right way of handling the highlights since now no branches will be added and unnecessary highlights (like in the example you just gave) will be dropped when more and more moves are made. Currently I am doing regenration of moves which is the problem.

I guess when I am done with this new scheme we could either drop the current superposition method or offer it as an alternative (it comes useful in many games). What do you think ?

[hgm]

OK, I proably don't understand your highlight method well enough to judge this. But what is so different about the Long Leaper compared to other pieces? For other pieces you also don't highlight all possible target victims in green on lifting. You wait until the hover over a capture square, and only then highlight victims when that would be to-square. I also don't understand how legs come in. A Long Leaper move is always a single leg, even if it jumps several pieces. In that respect the situation is very different from Checkers, where you would have to make an intermediate stop between the pieces.

Btw, I am trying to debug the ignored-clicks problem, with many printfs to the debug file. It seems that the cause is 'bouncing' of the mouse button. When you Ctrl+click on g8 (after i6), the first leg is performed, and causes sending "put g8," "lift g8" to the engine. But if a second Ctrl+click on g8 come in from the mouse button, before the engine could respond to put with click, this deselects the piece that was left selected (on g8) after the first leg. So the "click e6," of the engine now hits an empty to-square while no piece is selected, and thus is ignored (as well as further clickcommands).

I am not sure what to do about this. It is a hardware or user error, and it seems hard to make the software immune to this. Perhaps I should use a rule that click commands hitting empty squares or opponent pieces must be intended as to-squares, and if no piece is selected, should undo the last deselect. Or I should start timing the mouseclicks, and ignore a deselect that occured too quickly after a to-square entry of a partial move.

For now I just want to leave this, as the chances of it happening do not seem to be very large, and teh consequences not very severe. So you miss a one-click move now and then, and have to reselect the piece,and make the remaining legs by hand. No big deal.

[Daniel Shawul]

OK, I proably don't understand your highlight method well enough to judge this. But what is so different about the Long Leaper compared to other pieces? For other pieces you also don't highlight all possible target victims in green on lifting. You wait until the hover over a capture square, and only then highlight victims when that would be to-square. I also don't understand how legs come in. A Long Leaper move is always a single leg, even if it jumps several pieces. In that respect the situation is very different from Checkers, where you would have to make an intermediate stop between the pieces.

Ah..I see what you mean just now:) The long-leaper captures were completely missing from the code which highlights on _hover_. I had code for all other pieces but not the long-leapers somehow I forgot to copy paste.

The thing I was talking about concerns making moves leg by leg for checkers and similar variants. I liked the superposition for checkers because it alteast do not add branching highlights since it had the original
highlight. I am half way through re-implementing but it might take me some time to finish it.

Btw, I am trying to debug the ignored-clicks problem, with many printfs to the debug file. It seems that the cause is 'bouncing' of the mouse button. When you Ctrl+click on g8 (after i6), the first leg is performed, and causes sending "put g8," "lift g8" to the engine. But if a second Ctrl+click on g8 come in from the mouse button, before the engine could respond to put with click, this deselects the piece that was left selected (on g8) after the first leg. So the "click e6," of the engine now hits an empty to-square while no piece is selected, and thus is ignored (as well as further clickcommands).

I am not sure what to do about this. It is a hardware or user error, and it seems hard to make the software immune to this. Perhaps I should use a rule that click commands hitting empty squares or opponent pieces must be intended as to-squares, and if no piece is selected, should undo the last deselect. Or I should start timing the mouseclicks, and ignore a deselect that occured too quickly after a to-square entry of a partial move.

For now I just want to leave this, as the chances of it happening do not seem to be very large, and teh consequences not very severe. So you miss a one-click move now and then, and have to reselect the piece,and make the remaining legs by hand. No big deal.

Agreed.

[Roger Brown]

Indeed, there was a bug left, that still cleared the first empty square of any numeric stretch. I uploaded a WinBoard-Alien.zip now that fixes it.

Hello H.G.,

I have four versions of Winboard to be looked at:

Winboard Nebiyu installer
Winboard TM
Winboard 4.5
Winboard Alien

For ordinary chess I suppose that the TM version is the latest state of development, correct?

Precisely how do I use the other files?

Sorry to be a bother...

Later.

[Daniel Shawul]

Sorry to be a bother...

Indeed you are a bother atleast to me. You want the download links I am sure if you search this thread you will find what you want. You constantly keep on posting in a thread that you contribute nothing to viz a viz winboard _alien_ development, and yet you insist to be updated like you are some kind of authority.
You want the TM version? the standard one? the 4.5 version?.. all can be found if you just make a little effort and search..

[Roger Brown]

Sorry to be a bother...

Indeed you are a bother atleast to me. You want the download links I am sure if you search this thread you will find what you want. You constantly keep on posting in a thread that you contribute nothing to viz a viz winboard _alien_ development, and yet you insist to be updated like you are some kind of authority.
You want the TM version? the standard one? the 4.5 version?.. all can be found if you just make a little effort and search..

Hello Daniel,

I certainly am no authority nor do I insist on being updated and if my requests for information were taken as such by the contributors here then I apologise. I am an enthusiastic Winboard user and quite frankly I am excited by H.G.'s unceasing efforts to widen its use. I simply wanted to know from the developer what version did what at this time. There seemed to be a melding of the various Winboards now there seems to be a splitting off.

This topic is 30+ pages long. I hardly think I contributed more than half a page in total to it so I struggle to understand your use of constantly and keep on BUT out of respect for your sheer rudeness, general lack of class, ignorance and arrogance I will stay out of what you seem to regard as your private business.

Later.

[Daniel Shawul]

This topic is 30+ pages long. I hardly think I contributed more than half a page in total to it so I struggle to understand your use of constantly and keep on BUT out of respect for your sheer rudeness, general lack of class, ignorance and arrogance I will stay out of what you seem to regard as your private business.

Scream all you want. I just approved to what you said " that is to being a bother". Now go ahead serve your master.