Frivolous Engine Feature Ideas

[hgm]

My only problem with WB engines was the lack of some searchmoves/exclude best move/MultiPV feature which made them being used for analysis impractical because there was no way to quickly make it show an eval for its second favorite move.

I now even wonder how people used WB engines for analysis at all.

My WB engines usually do support multi-PV. Fairy-Max even supports move exclusion.

[Uri Blass]

My only problem with WB engines was the lack of some searchmoves/exclude best move/MultiPV feature which made them being used for analysis impractical because there was no way to quickly make it show an eval for its second favorite move.

I now even wonder how people used WB engines for analysis at all.

My WB engines usually do support multi-PV. Fairy-Max even supports move exclusion.

many engines do not support multi-PV but it is not a question of WB or UCI.
I am not sure but I guess that in the past the interface for winboard engines did not support multi-pv

I started with a winboard engine many years ago.
I did not work on developing my engine in the last years but I do not remember seeing multi-pv option when I implemented the WB commands.

This is a link that explain the protocol and I do not see something about multi-pv in the analyde mode section(edit:note that it is from 2009 and I guess since then there are new commands but the history is one of the reasons why many old WB engines do not support multi-pv).

Many authors do not support every possible option and it is an additional work to implement more options but
the fact that I did not read commands to change the number of lines in the winboard protocol at the time I developed my engine did not encourage me to implement multi-pv.

https://www.gnu.org/software/xboard/engine-intf.html#12

[hgm]

It is up to the WB engine how many PVs it will show. To allow the user to control this interactively it can define an option multi-PV. Before the protocol supported engine-defined options engines often had to be controlled by their config files or command-line arguments, but even then there would be methods to control the number of Thinking Output lines per depth in a more interactive way. Such as the askuser command. E.g. the engine could have asked for the number of lines to display every time you switched it to analysis mode. And asking for 0 lines could have been used as a request to ask this on a per-move basis.

[Ovyron]

An interesting question is why UCI won the protocol wars against Winboard, most top engines are UCI and even Chessbase dropped its proprietary format and moved its engines to UCI.

[hgm]

Marketing?

The lowest-quality system usually conquers the market. Happened also with video recorders. And operating systems. Internal-combustion engines are also theoretically inferior to hot-air (Stirling) engines.