"Derivatives" - The "D" Word......

[Christopher Conkie]

We would like to bounce off you some thoughts with regard to derivatives, how they should be presented, what might be acceptable to authors and how that relates to computer chess..

To facilitate this we will use Plague as an example. Plague is a derivative of Smaug which in turn is a derivative of Glaurung by Tord Romstad which is published under the GNU Public Licence.

1. The derivative needs to conform to the license of the originating engine.

2. Naming the derivative can be done but only provided that the name is presented in the following format, both in the executable and in the source that is supplied along with it. The reason for this is to show the lay user who is the original author. It should be visible. Since Plague is UCI, I show UCI code.

Code: Select all

id name Glaurung - Plague Derivative
id author Tord Romstad - Updated by Christopher Conkie

3. A derivative can be presented to people as diff files provided the naming convention used in 2. is used in them.

4. The entry of a derivative into a tournament is at the behest of the tournament organiser. Since tournament organisers do not want two of the same thing in their tournaments, permission must be sought from the author of the original engine and the tournament director has to want that to be allowed in their tournament.

These are just a few thoughts. We would now like to ask authors and tournament directors what they think of these points and to supply some of their own. It is not that we don't want others to contribute. It is just that those are the people best placed to decide such things since it concerns them.

The floor is now open......



Christopher

[sje]

From the event organizer's viewpoint, the derivative history is quite important.

The organizer should be able to award victory titles, rankings, and prestige according to author and program history.

If I were running a CC event, there would be awards like:

1) Best score by a non-derivative program by a single author, and

2) Best score by a non-derivative program by multiple authors, and maybe

3) Best score by an open source, multi-platform program.

[Stan Arts]

-Name of program has to stay the same, only the original author can change the name. Add something to the versionnumber to indicate the difference instead. Example: I'm cloning Fruit, but added the mighty Doom's BFG9000, Now my engine is Fruit 2.1BFG ! (I'm surprised no one thought of this, it would make it much stronger)

-Name of original author stays the same, mentioned first.

-No derivative in tournaments, unless the original author clearly initiates this himself.

-Don't. Others have worked on their honest engines for years, and you'll (should) be outcast like you have the plague. Instead, experiment in private, have something that works? Tell it to the original author.

-Write an original engine. You'll be very surprised how fulfilling it is. Have no time? Getting something to play it's first chess-like game will take hours, not weeks. In the meantime you learn loads and loads more about computerchess and programming then you would of cloning some extremely sophisticated engine, plus you'll be extremely and rightfully proud of your effort.

Stan

[Matthias Gemuh]

Code: Select all

id name Glaurung - Plague Derivative

Christopher

There already exists a stand naming convention, as in

"TogaII based on Fruit 2.1".

Matthias.

[bhlangonijr]

2. Naming the derivative can be done but only provided that the name is presented in the following format, both in the executable and in the source that is supplied along with it. The reason for this is to show the lay user who is the original author. It should be visible. Since Plague is UCI, I show UCI code.

Code: Select all

id name Glaurung - Plague Derivative
id author Tord Romstad - Updated by Christopher Conkie

You just started way wrong. Plague is also a derivative of Smaug. What about the credits of Joona?

[Christopher Conkie]

2. Naming the derivative can be done but only provided that the name is presented in the following format, both in the executable and in the source that is supplied along with it. The reason for this is to show the lay user who is the original author. It should be visible. Since Plague is UCI, I show UCI code.

Code: Select all

id name Glaurung - Plague Derivative
id author Tord Romstad - Updated by Christopher Conkie

You just started way wrong. Plague is also a derivative of Smaug. What about the credits of Joona?

No need, only the original author needs credit, and......Plague contains nothing that Joona did.......but it was a nice try and I do appreciate a nice try.



Christopher

[Christopher Conkie]

Code: Select all

id name Glaurung - Plague Derivative

Christopher

There already exists a stand naming convention, as in

"TogaII based on Fruit 2.1".

Matthias.

I'm afraid I will have to differ. In fact what you state is the start of a slippery slope.

eg

Plague based on Smaug and no one sees Glaurung which should be first anyway.

How about....

id name Glaurung 2.2 - Plague 1.0 Mod

To be fair I much prefer Stan's suggestion although.

Christopher

[Christopher Conkie]

-Name of program has to stay the same, only the original author can change the name. Add something to the versionnumber to indicate the difference instead. Example: I'm cloning Fruit, but added the mighty Doom's BFG9000, Now my engine is Fruit 2.1BFG ! (I'm surprised no one thought of this, it would make it much stronger)

-Name of original author stays the same, mentioned first.

-No derivative in tournaments, unless the original author clearly initiates this himself.

-Don't. Others have worked on their honest engines for years, and you'll (should) be outcast like you have the plague. Instead, experiment in private, have something that works? Tell it to the original author.

-Write an original engine. You'll be very surprised how fulfilling it is. Have no time? Getting something to play it's first chess-like game will take hours, not weeks. In the meantime you learn loads and loads more about computerchess and programming then you would of cloning some extremely sophisticated engine, plus you'll be extremely and rightfully proud of your effort.

Stan

I pretty much agree with everything you have said.



Christopher

[Christopher Conkie]

From the event organizer's viewpoint, the derivative history is quite important.

The organizer should be able to award victory titles, rankings, and prestige according to author and program history.

If I were running a CC event, there would be awards like:

1) Best score by a non-derivative program by a single author, and

2) Best score by a non-derivative program by multiple authors, and maybe

3) Best score by an open source, multi-platform program.

These are good suggestions but can you see that happening in say the WCCC where there is limited time and possibly resources?

I do agree that derivatives need to be kept completely separate if they are allowing in some tournament however.

I also think that the main thrust of any tournament is the original authors competition and there is no place for derivatives in that.

Christopher

[Ryan Benitez]

I am not sure about a standard but I do think proper credit to the original programmer and other contributors (such as beta testers) is very important.