Back around 2000-2001 Adrien Regimbald wrote the very nice little
chess program Faile, stopping at release 1.4. It was said to play at
the 2000-2200 ELO level.

Now, recently we've seen the small program BlindC, designed to play a
reasonable game in "blind chess" mode. Alas, that was only a DOS
executable; I wanted something for Linux, and BlindC in DOS emulation
was slow.

So I picked up Faile and got enthused enough to do some significant
work. I couldn't reach Adrien for specific permission, but his
license (the MIT license) does allow re-release by others.

In any case, I've added features such as save and restore, reading
commands from a file, output of PGN/move lists, and a few other
things, including an overhaul of how book moves are chosen. It's very
much suitable for "blind chess" mode, but will work with WinBoard and
XBoard, and can print its own ASCII boards if you wish.

My new distribution, version 1.5, includes a Windows and a Linux
executable, as well as an 11-page PDF manual, and the original opening
book which is "good enough" for a program of this type.

So, you say, who cares about an old program that is a good 800 or more
ELO weaker than other strong programs, including free ones, and has
only rather basic features by today's standards?

This is true minimalism; the compressed sizes of the Windows and Linux
binaries are about 28k and 48k respectively. The opening book is
about 240k. Minimalism is something which intrigues me; so much used
to be done in small footprints, and it still can. DVD sized
distributions don't have to be the only thing in existence.

99.9% of the work is of course Adrien's; I take credit for new bugs
and problems and little more.

If this sort of thing interests you, it's he

http://www.bobnewell.net/filez/faile15.zip

Source code isn't in the distribution but it's available free on
request.

Short follow up: I've gone to version 2.0.0 and renamed the fork
"Little" as it is unethical, I think, to retain the name "Faile"
without being the authorized maintainer, and moreover, to associate
Adrien with the bugs I've surely introduced. Still, all credit to
Adrien for fabulous base code and remarkable playing strength for the
size of the program. All I did was add things and probably break a
few things along the way

This is my last set of feature additions. I revised the opening book
building so now very large books are possible (anomalous with a
minimalist program!) with no hash conflicts. The manual is up to 18
pages or so. XBoard mode is cleanup up a little and so is ICS
mode. I added to the time controls and made them work better.
Etc., etc.

I tested a while with a guest account on FICS and no human did better
than draw over a series of many games. I think the estimate of
2000-2200 ELO playing strength at blitz levels is probably right.

It's been fun. The windows and linux executables, compressed, remain
in the 30k size range. Not 30 megabytes, 30 KILObytes.

Distributions he

http://www.bobnewell.net/filez/little/little200.zip (windows &
linux executables, small opening book, PDF manual)

http://www.bobnewell.net/filez/little/littled.zip (djgpp DOS
executable only)

Read the manual for info on getting several additional opening books
and the source code all of which is freely available.