HGM's Polyglot reference material

[AndrewGrant]

HGM, I hope you'll read this thread.

I've just been looking at http://hgm.nubati.net/book_format.html, in order to confirm the work of a colleague. Under the heading "Sample code", there is pg_key.c and pg_show.c, but neither file exists still on your website, nor are they archived by the wayback machine.

Do you still have those files on hand?

[Rebel]

pg.zip

[AndrewGrant]

Thanks, Ed. This will be very helpful.

[hgm]

I think I never had those. The HTML page with the format description was copied from Michel VandenBergh's website, and I added the part about variants to it. I never noticed that it contained relative links, which therefore started referring to my own website.

[smatovic]

Old website of Michel Van den Bergh is archived on archive.org:

https://web.archive.org/web/20180117040 ... ormat.html
https://web.archive.org/web/20180122140 ... a/pg_key.c
https://web.archive.org/web/20180115133 ... /pg_show.c

--
Srdja

[Dave Gomboc]

Is there a definitive upstream repository for Polyglot somewhere? While I do see a repository or two on Github, but nothing published by F. Letouzy (original author) or H.G. Muller (author of updates in "2.x" version).

[hgm]

Unfortunately the GitWeb interface for my on-line source repository is down, and I haven't had time to repair it. But isn't Polyglot available as a Debian (source) package? I haven't done anything on Polyglot for many years, so one supposes that the latest changes I made should have trickeld through to Debian.

I consider Polyglot obsolete, and have focused on UCI2WB as adapter for XBoard.

[Dave Gomboc]

Unfortunately the GitWeb interface for my on-line source repository is down, and I haven't had time to repair it. But isn't Polyglot available as a Debian (source) package? I haven't done anything on Polyglot for many years, so one supposes that the latest changes I made should have trickeld through to Debian.

I consider Polyglot obsolete, and have focused on UCI2WB as adapter for XBoard.

AFAIK, the tip of GNU's XBoard repository is currently using polyglot to communicate with UCI engines. Is switching to use UCI2WB one of the improvements that you have made to Winboard that isn't yet upstreamed? Let's try to get all your updates integrated into the GNU XBoarc repository. I'm happy to not care about Polyglot once it can be eliminated from the next point release of Winboard/XBoard!

[hgm]

Neither Polyglot nor UCI2WB is a GNU project. They cannot be in the XBoard repository at Savannah without transferring the copyrights to GNU. And Polyglot has many contributors.

Originally UCI2WB started as a rule-agnostic adapter for UCI dialects, so that it could handle arbitrary variants. Downside compared to Polyglot was that it was a pure protocol adapter. It did not store engine settings, and did not convert the PV to SAN. (Which is not possible without rule knowledge.)

But when CECP was extended to support engine-define options too, outsourcing these additional tasks to an adapter became undesirable, and I added the possibilty to save such settings and produce SAN PVs to the GUI itself.

[Dave Gomboc]

Neither Polyglot nor UCI2WB is a GNU project. They cannot be in the XBoard repository at Savannah without transferring the copyrights to GNU. And Polyglot has many contributors.

Well, the copyright transfer is definitely preferred, but I do believe that one acceptable alternative to GNU is that the author(s) explicitly contribute their work to the public domain. Even that would still require specific sign-off from every contributor, though, so it sounds like that's definitely not a practical option for Polyglot.

I did a web search to learn more about UCI2WB. https://www.chessprogramming.org/UCI2WB indicates that while there was some original program named the same back a couple of decades ago, that the one you're referring to is actually something that you created beginning with USI2WB in 2009 and released a version 4.0 of back in 2018 (viewtopic.php?t=69203). So, I wonder: are you the only author of the current code in the current edition UCI2WB, and are you (and any other authors that might exist) amenable to placing that code upstream in Savannah?

I was hoping to get all of your outstanding Winboard changes merged into Savannah's XBoard, then publish an official point release that contains it. However, if that's not a realistic option and WinBoard is just never going to be fully integrated back into Savannah, then it's better that I know that sooner rather than later so that I don't keep barking up the wrong tree.

Dave

P.S. Would calling it UCI2CECP instead make things less or more confusing if it's integrated into GNU XBoard?