Highest number of programs??

[Tony Thomas]

I'm not very familiar with Viper but maybe you should add it to Tord's list!

Now Tord also got four engines.

[Tony Thomas]

Who is the chess programmer that you know of who has the highest number of chess programs? I know that Tord got three, Gothmog, Glaurung 1X and his new Glaurung 2X(a complete rewrite). I know that Tom Kerrigan got 2 programs (TSCP, and Stobor), and Evgeniy Korniloff got 4 programs(Qchess, Chess, Woodpecker, Obender), Fabien Letuzey (Fruit, Chess 64), Daniel Shawul (Danchess, Scorpio)Who else that you know got more programs? A complete rewrite count as another program even if it doesnt have a different name.

I don't agree with your "complete rewrite" idea. But if you go by that, we had 6 complete rewrites of blitz over the years. from blitz I to blitz 6. Cray Blitz was a modification of blitz 6 to use vectors (major change but not a rewrite). Crafty has had two rewrites, it started as a chess 4.x type program, then was re-done to use rotated bitboards which changed everything again...

So, you have 8 total rewrites, so that would currently put you at the top. When I said a complete rewrite, I was talking about major changes in code. For example, adding history pruning and gaining 100 points from it does not make it a new program.

Also Daniel Mehrmann( Homer 1X, and Homer 2x)

[bob]

Who is the chess programmer that you know of who has the highest number of chess programs? I know that Tord got three, Gothmog, Glaurung 1X and his new Glaurung 2X(a complete rewrite). I know that Tom Kerrigan got 2 programs (TSCP, and Stobor), and Evgeniy Korniloff got 4 programs(Qchess, Chess, Woodpecker, Obender), Fabien Letuzey (Fruit, Chess 64), Daniel Shawul (Danchess, Scorpio)Who else that you know got more programs? A complete rewrite count as another program even if it doesnt have a different name.

I don't agree with your "complete rewrite" idea. But if you go by that, we had 6 complete rewrites of blitz over the years. from blitz I to blitz 6. Cray Blitz was a modification of blitz 6 to use vectors (major change but not a rewrite). Crafty has had two rewrites, it started as a chess 4.x type program, then was re-done to use rotated bitboards which changed everything again...

So, you have 8 total rewrites, so that would currently put you at the top. When I said a complete rewrite, I was talking about major changes in code. For example, adding history pruning and gaining 100 points from it does not make it a new program.

Also Daniel Mehrmann( Homer 1X, and Homer 2x)

That is correct for me. blitz 1 was completely rewritten (at least the engine part, some support stuff did not change, such as opening book format, etc) to create blitz 2. All the way to blitz 6 which became Cray Blitz. The two crafty rewrites were re-dos where we threw out the chess 4.x approach (both attacks_from and attacks_to incrementally updated in Make/Unmake) to the current far simpler/faster way of doing attacks (I do not count rotated to magic as a major change at all). That period of time covers about 40 years, from 1968 to today... In that context 8 total versions doesn't seem quite so big...

[swami]

Pedro Castro,Danasah and Caligula.

You better not take rewrite with same engine name into account, It's kinda confusing because many authors will claim they rewrote it from the scratch etc although if it is true.Better go ahead with multiple engine releases.

[Tony Thomas]

Pedro Castro,Danasah and Caligula.

You better not take rewrite with same engine name into account, It's kinda confusing because many authors will claim they rewrote it from the scratch etc although if it is true.Better go ahead with multiple engine releases.

I do want to take rewrites with same names to account, if program A shares no code with Program A 2.0, then program A 2.0 is indeed a different program. Some people release engines with only personality changes as different engines, Lordking comes to mind.

[Edsel Apostol]

In my case I wrote 4 Winboard programs before, namely Alpha, Gamma, Lambda, and Sigma. I started with Alpha, my first chess program. I was learning the C programming that time. It is using 0x88.

They are full of bugs and crashing so often. I decided to scrap them all.

I came across the source of the pre-alpha version of Olithink 5. I studied its implementation and I adapted bitboard for my engine. I released the first version of Twisted Logic when it was about the strength of TSCP. I developed it until the 0.099x version where I decided that the data structure is so poor it needs a complete rewrite.

The latest versions of Twisted Logic now is completely different from the 0.0xxx versions.

A total of 6 engines for me.

Edsel Apostol

[Matthias Gemuh]

3 engines in my case.


BigLion 1.x, BigLion 2.x, ArcBishop 1.x


Taktix is based on BigLion. Chancellor is based on ArcBishop.

Matthias.

[Tony Thomas]

3 engines in my case.


BigLion 1.x, BigLion 2.x, ArcBishop 1.x


Taktix is based on BigLion. Chancellor is based on ArcBishop.

Matthias.

You were on my list.
http://www.talkchess.com/forum/viewtopi ... 85&t=17474

[Tord Romstad]

I'm not very familiar with Viper but maybe you should add it to Tord's list!

Actually, I don't think he should. Viper is nothing more than a heavily lobotomized version of Glaurung 1, and shouldn't be counted as an independent program.

Tord

[hgm]

One that I missed who got more than two is H.G (Joker, Micromax, Fairy Max).

You missed more. But first, I do not count Fairy-Max an independent engine. Its code is 95% micro-Max 4.8, with just a few lines added to implement the weirder fairy pieces, and to read its move tables from a file in stead of having them hard-coded.

But I do have a third WB compatible engine in active competition, N.E.G., which is a resurrection of an old blitz program I wrote around 1978 as a joke. It is a program that does not search at all, but just picks a move from the move list based on the number of times the From and To square are attacked, what is in them, and what is the lowest attacker. (Not even a full SEE.) It does not know what Checkmate is, but still managed to checkmate its opponent twice in the 11 rounds of the ChessWar Promo division. Currently it is also participating in the top-500 of Système du Suisse, where it has 1/2 out of 6 upto now.

[Event "ChessWar XI Promotion 40m/20'"]
[Site "DELL-E6600"]
[Date "2007.07.31"]
[Round "4.68"]
[White "NEG 0.3d"]
[Black "Eden2"]
[Result "1-0"]
[ECO "A00"]
[WhiteElo "1099"]
[BlackElo "1000"]
[Annotator "2. -0.01"]
[PlyCount "11"]
[EventDate "2007.??.??"]

1. Nc3 f6 2. e3 2... a6 3. Qh5+ 3... g6 4. Qe2 4... h6 5. d4 5... g5 6. Qh5# {White mates} 1-0

A more difficult game for N.E.G. was:

[Event "ChessWar XI Promotion 40m/20'"]
[Site "DELL-E6600"]
[Date "2007.08.06"]
[Round "7.70"]
[White "POS 1.14"]
[Black "NEG 0.3d"]
[Result "0-1"]
[ECO "A00"]
[WhiteElo "1000"]
[BlackElo "1099"]
[Annotator "2... +0.13"]
[PlyCount "84"]
[EventDate "2007.??.??"]

1. a4 Nc6 2. d3 e6 3. Qd2 Bb4 4. Kd1 Bxd2 5. Nxd2 g5 6. Ra3 d5 7. Rc3 Bd7 8. Ke1 b6 9. Ngf3 Nf6 10. Ng1 Qe7 11. Ndf3 g4 12. Kd2 gxf3 13. Ke3 fxe2 14. Kxe2 Nd4+ 15. Kd1 c6 16. Kd2 Rd8 17. Ke1 Qb4 18. Bf4 Qxa4 19. Be3 Nxc2+ 20. Ke2 Nd4+ 21. Kd2 Qb4 22. Kc1 Nb3+ 23. Kb1 Rb8 24. Rc2 Qe1+ 25. Ka2 Qxf1 26. Kxb3 Qxd3+ 27. Kb4 Qxc2 28. Ka3 Qd3+ 29. Ka4 b5+ 30. Ka5 Qf5 31. Kb4 a5+ 32. Kc3 Ne4+ 33. Kc2 Nxf2+ 34. Kb3 Nxh1 35. Ba7 Rc8 36. Ka3 Qd3+ 37. Ka2 Qc4+ 38. Ka3 b4+ 39. Ka4 Qc2+ 40. b3 Qxg2 41. Bd4 Rf8 42. h4 Qa2# {Black mates} 0-1

Around 1980 I wrote a Chess program Usurpator I & II, first for 6800, then for 6502. (Both written in assembler, necessitating a rewrite from scratch, but as the algorithm at the time I switched CPU was the same, I count it as a single continuous development.) I have scanned the source from the book it was published in, and I was able to run it in a 6502 emulator. I plan to fuse the emulator (which I have in C source) to the WB driver of my other engines, to play Usurpator in the next Chess-War Promo.

Before I started Joker just over a year ago, I worked for one year on a C code that incorporated the ideas of Usurpator II, combining them with modern techniques like hash tables. I called it Usupator V, and it played in Leiden twice (October 2005 and May 2006). It was even WB compatible, on the second occasion, but still was a dismal failure. As it contained for more than 75% new ideas, I do count that one as an independent program. It turned out the old ideas could not be integrated well with new techniques (e.g. everything was path-dependent in Usurpator, leading to big trouble with hashing). So I decided to abandon these ideas completely, and start from scratch on a more conventional engine (which became Joker). Unfortunately Usurpator V was lost with the crash of my laptop HD.