Latest UCI spec?

[MartinBryant]

I have a copy of the UCI spec dated April 2006 from the ShredderChess website.
Has it been updated/maintained since then? If so, does anyone have a link?

The reason I ask is that we seem to have certain additions like 'SyzygyPath' nowadays but have these been officially adopted or are they just de-facto standards by now?

[phhnguyen]

I guess that is the latest version due to that site is from the authors and UCI has only one update about 2004.

'SyzygyPath' is just an option and there is no standard for that name and engines' developers are free to use any string they want. In fact, most engines use 'SyzygyPath' since it is simple, clear but some others use other strings such as 'SyzygyTbPath'. For other options, it is worse, for example, people may use 'SyzygyProbeDepth', 'Syzygy Probe Depth', 'Syzygy Depth'...

I have worked with those issues for BanksiaGUI. I don't want my users to have to set Syzygy paths manually for all their engines, say, over 100 ones. Thus I need to detect the option for automatically setting the Syzygy path. I have used the simple trick: match the option name with two words, "syzygy" and "path". If anyone uses something different than that, bad luck for him

[MartinBryant]

Thanks for the reply Nguyen. Ironically it was my first attempt at using Banksia this morning that prompted my question!

Back in 2006-8 I wrote my own GUI because Arena was missing some functionality that I wanted and I remember the issues I had then coping with the numerous interpretations of the rather loose UCI spec, so I feel your pain!

It seems bizarre that nobody has updated it in 15 years?!?!
There have been numerous postings in this forum over the years discussing interpretation of the spec and surely it would make sense that the conclusions/best-practice/advice from those discussions were captured in the spec? It would certainly make the job easier for GUI writers!
I'm sure it would be a thankless job but perhaps a small team of dedicated volunteers could bring it up to date and iron out some of the creases?

[Ras]

It seems bizarre that nobody has updated it in 15 years?!?!

That's also an upside of UCI - it's stable and doesn't get "improved", becoming ever more convoluted until you have basically no support in the wild for most of the features so that from a user POV, every engine becomes a snowflake - and that's not what users want.

[CMCanavessi]

It seems bizarre that nobody has updated it in 15 years?!?!

That's also an upside of UCI - it's stable and doesn't get "improved", becoming ever more convoluted until you have basically no support in the wild for most of the features so that from a user POV, every engine becomes a snowflake - and that's not what users want.

It also has some issues that could be vastly improved though... maybe it's time for chess programmers to get together and come up with a new up to date standard. Things have changed A LOT recently.

[AndrewGrant]

It seems bizarre that nobody has updated it in 15 years?!?!

That's also an upside of UCI - it's stable and doesn't get "improved", becoming ever more convoluted until you have basically no support in the wild for most of the features so that from a user POV, every engine becomes a snowflake - and that's not what users want.

It also has some issues that could be vastly improved though... maybe it's time for chess programmers to get together and come up with a new up to date standard. Things have changed A LOT recently.

One update per 15 years is a nice compromise.

I'll never touch xboard due to its ever-changing nature.

[Ferdy]

Thanks for the reply Nguyen. Ironically it was my first attempt at using Banksia this morning that prompted my question!

Back in 2006-8 I wrote my own GUI because Arena was missing some functionality that I wanted and I remember the issues I had then coping with the numerous interpretations of the rather loose UCI spec, so I feel your pain!

It seems bizarre that nobody has updated it in 15 years?!?!
There have been numerous postings in this forum over the years discussing interpretation of the spec and surely it would make sense that the conclusions/best-practice/advice from those discussions were captured in the spec? It would certainly make the job easier for GUI writers!
I'm sure it would be a thankless job but perhaps a small team of dedicated volunteers could bring it up to date and iron out some of the creases?

I have a github repo on uci, you may raise an issue or open a topic for discussion.