chessdb-users — User mailing list - used by developers too.

Re: [Chessdb-users] Tablebase Directory

[Dr. D. K. <dav...@on...>]

Linda VanArsdale wrote:

> 
> Hi Dave,
> 
> Under the options menu it would not allow me select the folder path.  I was
> able to type it in and it did select, recognize and load the table bases.
> My operating system is Win2000 Pro if that help.
> 
> Thanks for the Quick response.
> 
> Phil VanArsdale AKA Pawnd

Can anyone else reproduce this problem? I have just tried it under 
Windows XP Pro with SP2 and had no problems at all. I just clicked on 
the 3 dots '...' then navigated to the directory

C:\Program Files\ChessDB\tablebases

picked a tablebase at random and it all worked. Here is a screen shot 
showing the startup screen once I had did this.

http://chessdb.sourceforge.net/tutorial/img/xp-with-tablebases.jpg

As you can see, I only have the 3-piece tablebases loaded here, but i 
have every reason to believe it would work with the 4 and 5 piece ones - 
I can't at the minute test it, since there is no drive letter mapped to 
the directory with them in. (They are stored on another computer).


I doubt this is a bug in ChessDB, since that part of the code has not 
changed since Scid and it works both on XP Pro and Solaris (UNIX). But 
I'm willing to be corrected!

Perhaps someone else can check 3.6.12-beta-1 with the tablebases.

[Chessdb-users] Help using pgnchessdb

[Richard & H. D. <ri...@di...>]

Hi,
When I run this program (pgnchessdb), the Windows DOS box opens for a =
millisecond & closes!
Can anyone tell me how I can run this program normally to create chessdb =
files from pgn files?
I am most grateful for any help & look forward to a reply,many thanks,

Richard 

Re: [Chessdb-users] Tablebase Directory

[Linda V. <van...@ve...>]

----- Original Message -----
From: "Dr. David Kirkby" <dav...@on...>
To: "Linda VanArsdale" <van...@ve...>
Cc: <che...@li...>
Sent: Wednesday, January 17, 2007 3:32 AM
Subject: Re: Tablebase Directory

>
> http://scid.sourceforge.net/tutorial/t_tool_egtb.html
>
> and it particular the sentence:
>
> "It may be confusing to select a file instead of a folder, but the
> reason is so you can be sure you are selecting a folder which actually
> contains tablebase files."
>
>
> Just select a file and it should work.
>
>
> Dave

Hi Dave,

Under the options menu it would not allow me select the folder path.  I was
able to type it in and it did select, recognize and load the table bases.
My operating system is Win2000 Pro if that help.

Thanks for the Quick response.

Phil VanArsdale AKA Pawnd

[Chessdb-users] Using pgnchessdb

[Richard & H. D. <ri...@di...>]

Hi,
When I run this program, the Windows DOS box opens for a millisecond & =
closes!
Can anyone tell me how I can run this program normally to create chessdb =
files from pgn files?
I am most grateful for any help & look forward to a reply,many thanks,

Richard 

[Chessdb-users] Right click in pgn window sometimes crashes

[Bryan T. <bt...@bw...>]

I've noticed this under both scid and chessdb: when I have a game, 
especially one with heavy annotation and I right click in the pgn 
window, the application crashes with no error -- it just "vanishes".

It's supposed to render a board with the move I clicked on. Has anybody 
else seen this?

Re: [Chessdb-users] Move annotation comments

[Bryan T. <bt...@bw...>]

Dr. David Kirkby wrote:

> I'm not sure why some messages here are being marked as spam. They look 
> fine to me.

It's actually my fault -- but it's already been addressed.

Sourceforge evidently sells a footer to www.techsay.com. My email 
provider's spam protection gets a hit on that, because evidently a lot 
of real spam comes from or spoofs that site and the footer's "earn cash" 
etc... looks like spam. The result is that everything in my inbox from 
this list is labelled as spam. The long and short of it is that I have 
already had them fix the false positive when it originates from 
sourceforge.com, but I was careless and failed to remove the tag before 
I replied.

Re: [Chessdb-users] Move annotation comments

[Dr. D. K. <dav...@on...>]

Bryan Taylor wrote:
> Good plan. Make sure everything (checkboxes & thresholds) stores its 
> state in the config.

Yes, that would be a good idea.

> Another idea is that when it annotates one of these symbols, you could 
> have it add a variation with the better line or just the better move. So 
> maybe instead of checkboxes, you could have a four-way drop down 
> (nothing, ?!/?/??, move, line).
> 
> Dr. David Kirkby wrote:


I'll look at that possibly. But like Jeremy I am keen to keep things 
relativly simple.

I'm not sure why some messages here are being marked as spam. They look 
fine to me.

Dave

[Chessdb-users] New release of ChessDB database

[Dr. D. K. <dav...@on...>]

I've put a new release on the Sourceforge web site. There is both source
code and a Windows executable.

I've put them in a package called 'unstable-for-testing' since I have
not fully debugged this release. It is a *lot* better than the alpha one
I released a week or so ago,.

Get it from:

http://sourceforge.net/project/showfiles.php?group_id=184864

A list of the changes is at:
http://chessdb.sourceforge.net/changes/#3.6.12-beta-1

The main feature which are improved are connected with downloads from TWIC
http://chessdb.sourceforge.net/tutorial/twic-retriveal.php
(which now work under Windows or UNIX) and with checking the revision of
the software.

http://chessdb.sourceforge.net/tutorial/check-software-revision.php

You should now be able to quickly tell if you have the latest release,
whether it is a stable release, and whether there are any 'beta' releases.

A lot more of the user messages are now in the language files, so
translations should be better (if someone will do the translations of
course). Any help with that appreciated.

CVS is updated and tagged to this release, although it implies the 116th
Jan, which I think is a bit odd, but I don't know how to change the tag.
Its probably not worth worrying about anyway, as the version is there
and the CVS will tag the date correctly.

Let me know of any problems. I hope to make another beta release then a
stable one in the near future. But for now I am keeping it as beta,
until I am sure it is almost bug free.

PLEASE LET ME KNOW OF ANY BUGS AND COMMENTS YOU HAVE

Dave

[Chessdb-users] New release (3.6.12-beta-1) of ChessDB

[Dr. D. K. <dav...@on...>]

I've put a new release on the Sourceforge web site. There is both source 
code and a Windows executable.

I've put them in a package called 'unstable-for-testing' since I have 
not fully debugged this release. It is a *lot* better than the alpha one 
I released a week or so ago,.

Get it from:

http://sourceforge.net/project/showfiles.php?group_id=184864

A list of the changes is at:
http://chessdb.sourceforge.net/changes/#3.6.12-beta-1

The main feature which are improved are connected with downloads from TWIC
http://chessdb.sourceforge.net/tutorial/twic-retriveal.php
(which now work under Windows or UNIX) and with checking the revision of 
the software.

http://chessdb.sourceforge.net/tutorial/check-software-revision.php

You should now be able to quickly tell if you have the latest release, 
whether it is a stable release, and whether there are any 'beta' releases.

A lot more of the user messages are now in the language files, so 
translations should be better (if someone will do the translations of 
course). Any help with that appreciated.

CVS is updated and tagged to this release, although it implies the 116th 
Jan, which I think is a bit odd, but I don't know how to change the tag. 
Its probably not worth worrying about anyway, as the version is there 
and the CVS will tag the date correctly.

Let me know of any problems. I hope to make another beta release then a 
stable one in the near future. But for now I am keeping it as beta, 
until I am sure it is almost bug free.

PLEASE LET ME KNOW OF ANY BUGS AND COMMENTS YOU HAVE

Dave

Re: [Chessdb-users] Move annotation comments

[Dr. D. K. <dav...@on...>]

Jeremy White wrote:
>>How would you feel if I implemented this, but with a checkbox to do so
>>and have a radio button which defaulted to not doing it for '?!' and
>>'?', but only '??'. (I know that sentence would end with a ?, but I
>>think it would confuse matters more).
> 
> 
> I'm not going to object when someone else does the work <grin>.
> 
> Cheers,
> 
> Jer
> 


perhaps I'll do that at some point. It will not be in the next beta 
release (which I hope to release within the next 3 hours or so), but 
hopefully some time in the near future.

Re: [Chessdb-users] Move annotation comments

[Jeremy W. <jw...@co...>]

> How would you feel if I implemented this, but with a checkbox to do so
> and have a radio button which defaulted to not doing it for '?!' and
> '?', but only '??'. (I know that sentence would end with a ?, but I
> think it would confuse matters more).

I'm not going to object when someone else does the work <grin>.

Cheers,

Jer

Re: [Chessdb-users] [SPAM?] Re: Move annotation comments

[Bryan T. <bt...@bw...>]

Good plan. Make sure everything (checkboxes & thresholds) stores its 
state in the config.

Another idea is that when it annotates one of these symbols, you could 
have it add a variation with the better line or just the better move. So 
maybe instead of checkboxes, you could have a four-way drop down 
(nothing, ?!/?/??, move, line).

Dr. David Kirkby wrote:
> Dr. David Kirkby wrote:
> 
>> How would you feel if I implemented this, but with a checkbox to do so 
>> and have a radio button which defaulted to not doing it for '?!' and 
>> '?', but only '??'. (I know that sentence would end with a ?, but I 
>> think it would confuse matters more).
> 
> 
> I mean 3 checkboxes - one for ??, one for ? and another for ?!. No radio 
> buttons at all. Make the deffault just ??
> 
> 
> 
> 
> 
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Chessdb-users mailing list
> Che...@li...
> https://lists.sourceforge.net/lists/listinfo/chessdb-users

Re: [Chessdb-users] Move annotation comments

[Dr. D. K. <dav...@on...>]

Dr. David Kirkby wrote:

> How would you feel if I implemented this, but with a checkbox to do so 
> and have a radio button which defaulted to not doing it for '?!' and 
> '?', but only '??'. (I know that sentence would end with a ?, but I 
> think it would confuse matters more).


I mean 3 checkboxes - one for ??, one for ? and another for ?!. No radio 
buttons at all. Make the deffault just ??

Re: [Chessdb-users] Move annotation comments

[Dr. D. K. <dav...@on...>]

Jeremy White wrote:

> Sorry for the delayed response; I was travelling all last week.
> 
> Yes, this would be easy.  I have to confess that I have a range
> of reservations about doing that automatically.  First, I find
> that crafty's score varies a fair amount; a swing of 0.2, for example,
> does not seem uncommon, even if you follow the 'best move'.


True. I have never looked in detail how much it swings, but making the 
default 0.5 would probably prevent them too often.

>  Second,
> I'm a big fan of keeping things simple (I still question whether
> adding '?' was the right decision, to be honest).  Third, my sense was that
> the 'add a variation when chosen move is not the best move' was
> the better tool for the serious chess player.  (Although I've
> never really used it, and so I don't know how well it works
> out in practice).

> However, I'm also uninterested for purely selfish reasons:  for what
> I'm doing (analyzing kids games), I'll be thrilled if I can
> reduce the number of ?? moves, let alone ? moves; ?! is a subtlety
> beyond my (admittedly young) kids.  So if I added a '?!' category,
> I'd have to also add a way of ignoring it, because I don't want
> it in the annotations I give the kids :-(.

How would you feel if I implemented this, but with a checkbox to do so 
and have a radio button which defaulted to not doing it for '?!' and 
'?', but only '??'. (I know that sentence would end with a ?, but I 
think it would confuse matters more).

That would seem to offer (to me anyway) a reasonable compromise - and 
software is always going to be a compromise.

Re: [Chessdb-users] Move annotation comments

[Jeremy W. <jw...@co...>]

Dr. David Kirkby wrote:
> Bryan Taylor wrote:
> 
>>When I annotate my games with a chess engine I tend to use these 
>>guidelines.
>>
>>?! - a dubious move: evaluates to 0.20 to 0.50 below the best move
>>? - a mistake: evaluates 0.50 to 1.00 below the best move
>>?? a blunder: evaluates to 1.00 or more below the best move
>>
>>It would be cool if we support three such thresholds for these auto 
>>annotations. 
> 
> 
> Doing that would be quite easy I suspect - what does Jeremy think, since 
> he wrote it?

Sorry for the delayed response; I was travelling all last week.

Yes, this would be easy.  I have to confess that I have a range
of reservations about doing that automatically.  First, I find
that crafty's score varies a fair amount; a swing of 0.2, for example,
does not seem uncommon, even if you follow the 'best move'.  Second,
I'm a big fan of keeping things simple (I still question whether
adding '?' was the right decision, to be honest).  Third, my sense was that
the 'add a variation when chosen move is not the best move' was
the better tool for the serious chess player.  (Although I've
never really used it, and so I don't know how well it works
out in practice).

However, I'm also uninterested for purely selfish reasons:  for what
I'm doing (analyzing kids games), I'll be thrilled if I can
reduce the number of ?? moves, let alone ? moves; ?! is a subtlety
beyond my (admittedly young) kids.  So if I added a '?!' category,
I'd have to also add a way of ignoring it, because I don't want
it in the annotations I give the kids :-(.


>>combination/tactical moves and label them accordingly. I'd love to let 
>>the app chew on a collection of games in a batch and present me with 
>>tactical problems to solve taken from the games.

Yeah, I've been chewing on a way to do this in a batched way.
Sadly, I think it requires an alternate program that talks directly
to the chess engine, and I'm not quite up to that yet.  I'll see
how I feel if/when I succeed in getting the kids to send me
batches of their games (right now I just have dribbles, which
is quite easy to handle manually).

Cheers,

Jeremy

[Chessdb-users] Feature Request: Filter Opening Tree

[Bryan T. <bt...@bw...>]

It would be really cool if the opening tree could be configured to apply 
a filter before displaying it's results. For example, I sometimes want 
to see only games between 2600+ opponents or games from the last 2 years.

I'd like to enter this into a filter and see the opening tree breakdown 
only for games matching my filter, and I want the filter to remain in 
place for subsequent moves, including in "lock" mode.

I often end up creating multiple databases with overlapping content just 
because I don't have this feature.

[Chessdb-users] Interesting review

[Dr. D. K. <dav...@on...>]

I thought you might like to take a look at this web site

http://freechess.50webs.com/index.html

it says on the front page "ChessDB knocks off Playchess from Pick of the 
Best"

then has quite a nice review

http://freechess.50webs.com/chessdb.html


dave

Re: [Chessdb-users] positive feedback: Chess-DB

[Dr. D. K. <dav...@on...>]

Regine F. wrote:
> Dear Dr. Kirby,
> 
>  I have just downloaded (for testing) chess-db. I was
> reluctant to do so, as I am using SCID for many years
> and especially all the LaTeX related stuff to create
> opening reports or game documentation...

> I was afraid that the LaTeX stuff might not work or
> was even removed from the new Chess-DB. But, every
> feature I have used so far works.

Good.

You have raised an interesting point there. You may not be the only one 
who feared a missing feature. Nothing at all has been removed.

I've put a page up on Sourceforge showing the differences between Scid 
and ChessDB

http://chessdb.sourceforge.net/Scid/

I decided to the fact nothing has been removed. It would be a shame if 
someone failed to upgrade for fear of loosing something.

> And best of all: it's 
> -reliable and faster than before on SCID.

I'm surprised you find it faster, since there have not been any changes 
in ChessDB which were designed to improve the speed. However, there may 
have been in the gzip library it links to, the compiler, or new 
instructions in your CPU.

Perhaps the C++ compiler you have has done a better job of optimisation 
than the last time you build it. Compilers tend to improve a bit over time.

It is also possible that you have a newer CPU than when you compiled it, 
so the compiler has made use of new instructions in that CPU. To take an 
extream example, if you built Scid on a standard 386 machine, even the 
floating point would have to have been emulated in software since the 
386 had no floating point processor. So Scid would not have use the 
floating point processor in your newer machine - it would have still 
used emulation, which would be slower.

It's difficult from me to say why it is faster, but I don't think there 
is anything changed in ChessDB that should improve the speed.

You *may* be able to improve the speed by removing unnecessary language 
files. Just edit the Makefile and take out any you don't need from this 
line.

LANGUAGES = tcl/lang/deutsch.tcl tcl/lang/francais.tcl 
tcl/lang/italian.tcl tcl/lang/nederlan.tcl tcl/lang/spanish.tcl 
tcl/lang/portbr.tcl tcl/lang/swedish.tcl tcl/lang/norsk.tcl 
tcl/lang/polish.tcl tcl/lang/czech.tcl tcl/lang/hungary.tcl 
tcl/lang/serbian.tcl

I've not tested that I must say.


> Thank you and all co-working Chess-DB developpers for
> this positive development.

You are welcome.

BTW, you posted your message to che...@li... 
without first registering I believe. That results in the post being held 
until I log in and approve it. That obviously takes time, so for best 
response, you should visit

https://lists.sourceforge.net/lists/listinfo/chessdb-users

and register. If you want to look at past messages to the mailing list, 
you *should* be able to find them at

http://sourceforge.net/mailarchive/forum.php?forum=chessdb-users

That said, Sourceforge have had a few problems with this and it often 
happens these are delayed. You message for example was not visible when 
I just checked.

> Looking forward for the future release,

There will be another release soon, which will improve the code for 
downloading from the Week in Chess (TWIC) site.

That is quite a difficult task. Mainly knowing what to do for the user 
and what to let them decide. If I gave them all the control I could, the 
interface would be very confusing.

> regine
> 

[Chessdb-users] positive feedback: Chess-DB

[Regine F. <reg...@ya...>]

Dear Dr. Kirby,

 I have just downloaded (for testing) chess-db. I was
reluctant to do so, as I am using SCID for many years
and especially all the LaTeX related stuff to create
opening reports or game documentation...

I was afraid that the LaTeX stuff might not work or
was even removed from the new Chess-DB. But, every
feature I have used so far works.

And best of all: it's 
-reliable and faster than before on SCID.

Thank you and all co-working Chess-DB developpers for
this positive development.

Looking forward for the future release,

regine


	
	
		
___________________________________________________________ 
New Yahoo! Mail is the ultimate force in competitive emailing. Find out more at the Yahoo! Mail Championships. Plus: play games and win prizes. 
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk 

[Chessdb-users] An FAQ (list of Frequently Asked Questions)

[Dr. D. K. <dav...@on...>]

I'm thinking of creating an FAQ for chessDB - lost of Frequently Asked 
Questions.

If you know of a question which has been asked a few times, or can think 
of a non-obvious question, and YOU KNOW THE ANSWER, then can you seed 
the question *and* the answer as an email to either this mailing list 
(best) or me alone if you would rather do that.

There are currently only two quesions in the FAQ (which are both from 
me), so I have not bothered yet to provide a link to the FAQ from 
CheesDB's homepage. Once there are a a dozen or so questions, I'll make 
a link from the homepage.

Re: [Chessdb-users] Move annotation comments

[Dr. D. K. <dav...@on...>]

Bryan Taylor wrote:
> When I annotate my games with a chess engine I tend to use these 
> guidelines.
> 
> ?! - a dubious move: evaluates to 0.20 to 0.50 below the best move
> ? - a mistake: evaluates 0.50 to 1.00 below the best move
> ?? a blunder: evaluates to 1.00 or more below the best move
> 
> It would be cool if we support three such thresholds for these auto 
> annotations. 

Doing that would be quite easy I suspect - what does Jeremy think, since 
he wrote it?

> It would be very cool if other symbols could also be auto 
> annotated. This could include !?, !, !! 


That would be a lot harder and need a different approach. The chess 
engine does not give sufficient information to make it obvous how this 
would be done.


> or also the position evaluations 
> like +=. 


That again would be easy, although I guess you would only want it when 
the score changes in one direction. You don't want every move annotated 
as a win for white if the mate is not for another 20 moves or so.

> It would be awesome if there was a way to find key 
> combination/tactical moves and label them accordingly. I'd love to let 
> the app chew on a collection of games in a batch and present me with 
> tactical problems to solve taken from the games.

It's not obvious how one is going to be able to do this *easily* from 
the feedback given by a chess engine - i.e. a score and a move sequence.

> These are the symbols for Informant/ECO and by implication FIDE, which I 
> propose is the best reference standard. An advantage of this is that 
> they have official translations into about a dozen languages.
> 
> The position evaluations += etc... are easy to define as scores, but 
> it's harder to decide when to add them (perhaps when they change more 
> than some amount?). Tagging good  moves is a challenge:
> !? - a move deserving attention
> !  - a very good move
> !! - an excellent move

> It's hard to figure out a way to make this precise. 

yes, I feel that too.

> Some ideas:
> - ! or !! should be a move that increases the score relative to the 
> previous evaluation or be "the only move" which keeps the score up, 
> without be forced (ie recapturing after an exchange is not !)

But so often there is only one move which keeps the score up and that 
move is very obvious, so hardly deserves ! or !!

> - ! need not be unique in this regard, but !! should be
> - !! should be "hard to find" which perhaps means it's a sacrifice

Perhaps one way is to let the engine analyse the problem to a depth of 
N, then redo it to N+2. If the score is significantly different, then 
there is something buried N+1 or N+2 moves away that leads to an 
advantage and is do finding it would probably deserve a ! or !!

It would not be too hard to do this sort of thing. Give crafty a pgn 
file, limit the search depth and record the scores. Then do it again at 
a deeper search. Find where the scores are significantly different, 
which indicates there is a good move available that is not too obvious 
to see, as the chess engine missed it when it was not searching 
sufficiently deep.

But that is quite a different problem. Interesting yes, but a very 
different problem. Perhaps not too hard to do, but it is not really an 
extension of what Jeremy wrote, but something quite different. It would 
need a new tool to do it.

> - For endgames only Nunn (and Muller/Lamprecht) use ! to mean the only 
> move which doesn't change the evaluation (win/loss/draw) and ? to mean a 
> move which allows the other side to change the evaluation.

Although once again, sometimes (often) the only sensible move can be 
very obvious in places. So there is not a lot of point in annotating a 
move with !! if it is very obvious is there? But how you get a computer 
to tell you when it would be obvious to a human is not too clear.

> As for !? I sometimes use this for moves that score within 0.1 of the 
> best move and somehow is "different" than the main line. I don't know 
> how to make that precise.

I think it one is not careful, one could end up with a annotated file 
that is a bit messy.

[Chessdb-users] Move annotation comments

[Bryan T. <bt...@bw...>]

When I annotate my games with a chess engine I tend to use these 
guidelines.

?! - a dubious move: evaluates to 0.20 to 0.50 below the best move
? - a mistake: evaluates 0.50 to 1.00 below the best move
?? a blunder: evaluates to 1.00 or more below the best move

It would be cool if we support three such thresholds for these auto 
annotations. It would be very cool if other symbols could also be auto 
annotated. This could include !?, !, !! or also the position evaluations 
like +=. It would be awesome if there was a way to find key 
combination/tactical moves and label them accordingly. I'd love to let 
the app chew on a collection of games in a batch and present me with 
tactical problems to solve taken from the games.

These are the symbols for Informant/ECO and by implication FIDE, which I 
propose is the best reference standard. An advantage of this is that 
they have official translations into about a dozen languages.

The position evaluations += etc... are easy to define as scores, but 
it's harder to decide when to add them (perhaps when they change more 
than some amount?). Tagging good  moves is a challenge:
!? - a move deserving attention
!  - a very good move
!! - an excellent move

It's hard to figure out a way to make this precise. Some ideas:
- ! or !! should be a move that increases the score relative to the 
previous evaluation or be "the only move" which keeps the score up, 
without be forced (ie recapturing after an exchange is not !)
- ! need not be unique in this regard, but !! should be
- !! should be "hard to find" which perhaps means it's a sacrifice
- For endgames only Nunn (and Muller/Lamprecht) use ! to mean the only 
move which doesn't change the evaluation (win/loss/draw) and ? to mean a 
move which allows the other side to change the evaluation.

As for !? I sometimes use this for moves that score within 0.1 of the 
best move and somehow is "different" than the main line. I don't know 
how to make that precise.

Re: [Chessdb-users] [Patch] Store analysis engine name, use it in blunder comments

[Jeremy W. <jw...@co...>]

> 
> 1) You use the terms
> 
> ?? - blunder
> ? - mistake.

I confess to relying entirely on Google; and the first entry it
pointed to was one in Wikipedia:

  http://en.wikipedia.org/wiki/Punctuation_%28chess%29

I also felt that 'mistake' was more accurate than 'weak move';
the loss of a pawn (my default threshold) is pretty significant
in my opinion.

However, I am quite far from being a chess scholar, so I am quite
cheerful to have that say anything (in theory, that's just a change
to the language file).

> So there seems to be a whole range of meanings attacked to '?' and '?!'.
> So its hard to say that ChessDB should use X rather than Y, but I think
> its reasonable to say it should use the same term consistently.

That's persuasive to me.  Someone can later change all of them in
one patch if they feel strongly it should be something else.
I tweaked the text to reflect that.

> 
> 2) Also, the tick box you have added to enable this useful feature says
> 
> "When game is an obvious blunder"
> 
> but it will annotate moves which are less week/poor/dubious (pick you
> term) than an "obvious blunder".
> 
> So why not use
> 
> "When game move is a poor move or blunder"
> 
> or similar. (again, the poor/week/bad should be consistent throughout).

<grin>

Because when I wrote it, I only had the blunders, and didn't
have the second threshold.  I tweaked the language.

> 
> 
> 3) The defaults for both '??' and '?' are 1.0
> 
> Would it not be better to have the default for '?' to be less than the
> default for '??' - I've no idea how much less, but it would make it more
> obvious one was considered less weak than the other.

I'm pretty sure the default is 1 and 3, although it will save
whatever value you enter there in your personal
~/.chessdb options area.

> 
> 4) It is possible to set the score for '??' to be less than for '?',
> which is rather pointless.  Would it not be preferable to add a test
> which throws a warning message if the score for '?' was greater than the
> score for '??'.

I debated that.  The code should, I believe, more or less function if you
do that, and do the reasonable thing (which is ignore the blunder score).
To some extent, I thought that perhaps that was okay - the user has asked
for blunders to be less important than mistakes, so we ignore them.
But the honest truth was that I don't know tcl/tk well enough to do
a validation, so I didn't!

A patch attached this time (sorry about that), with the few tweaks
in language discussed above, as well as the original proposed
change to add the name to the score in the comments.

Cheers,

Jeremy

Re: [Chessdb-users] [Patch] Store analysis engine name, use it in blunder comments

[Dr. D. K. <dav...@on...>]

Jeremy White wrote:
> This tracks the name of the chess engine used for a given
> analysis mode, and then uses that name in the comments
> for blunder analysis.
> 
> Apply with:
>   patch -p0 < analysis.name.diff
> 
> Cheers,
> 
> Jeremy

Thanks Jeremy,

I was hoping to get around that to pretty soon, but have been busy 
working on the facility to download new games from The Week in Chess 
(TWIC) every week. Hence adding the name of the anlaysis engine was not 
done.

A few comment on your new code on which I admit I am being 
picky/pedantic, but critical comment often useful to aid to improvements 
in my experience.

1) You use the terms

?? - blunder
? - mistake.

in this part of the software.

That is different to other parts of the old Scid (which ChessDB has not 
changed) where '?' is 'poor move' , rather than 'a mistake' (try 
hovering the mouse over '?' in the comment editor).

It seems to me these should be consistent with each other. A poor move 
in both places, or a mistake in both places, but not a mix of the two.

I've tried a few books / resources to see if there is much agreement on 
the symbols used - the answer is no, although none of the resources use 
the term 'mistake'.

a) Oxford Companion to chess

?! - move of doubtful value.
? - weak move
?? - blunder

b) Nunn's chess openings
?! - dubious move
? - bad move
?? blunder

c) Batsford Modern Chess Openings

?! - Speculative attempt to complicate.
? - Bad or weak move
?? - blunder

d) Scid 3.6.1 (last official release)
?! - dubious move
? - poor move
?? - blunder

e) Annotation comments in the new code in ChessDB

? - mistake
?? - blunder


So there seems to be a whole range of meanings attacked to '?' and '?!'. 
So its hard to say that ChessDB should use X rather than Y, but I think 
its reasonable to say it should use the same term consistently.

2) Also, the tick box you have added to enable this useful feature says

"When game is an obvious blunder"

but it will annotate moves which are less week/poor/dubious (pick you 
term) than an "obvious blunder".

So why not use

"When game move is a poor move or blunder"

or similar. (again, the poor/week/bad should be consistent throughout).


3) The defaults for both '??' and '?' are 1.0

Would it not be better to have the default for '?' to be less than the 
default for '??' - I've no idea how much less, but it would make it more 
obvious one was considered less weak than the other.

4) It is possible to set the score for '??' to be less than for '?', 
which is rather pointless.  Would it not be preferable to add a test 
which throws a warning message if the score for '?' was greater than the 
score for '??'.

[Chessdb-users] [Patch] Store analysis engine name, use it in blunder comments

[Jeremy W. <jw...@co...>]

This tracks the name of the chess engine used for a given
analysis mode, and then uses that name in the comments
for blunder analysis.

Apply with:
  patch -p0 < analysis.name.diff

Cheers,

Jeremy