Page 1 of 3
What was the verdict on Protector 1.2.7?
Posted: Sat Oct 03, 2009 11:46 pm
by Graham Banks
Original engine or Toga clone?
At present I'm treating it as original, unless the consensus says otherwise.
Cheers,
Graham.
Re: What was the verdict on Protector 1.2.7?
Posted: Sat Oct 03, 2009 11:58 pm
by Graham Banks
Leo Dijksman
it seems nothing else then another TogaII, this time "bitboard based" with "many small" changed code!?
I asked some people but they say "its nearly ALL Toga", which sounds like the readme.txt part:
"In short it's kinda Bitboard-based TogaII with many small improvements."
Patrick Buchmann
Olivier give me the advice in my forum: another Fruit derivate.
Protector author
my name is Raimund Heid and I am the author of Protector. Protector is new to you and I'd like to clarify its relation to Toga/Fruit in order to give you a solid base of information for obvious questions like the one issued by K. Inuen.
When I claimed 'Protector is kinda bitboard-based Toga' I wanted to point out that Protector is based on many ideas implemented in Toga/Fruit. However, you have to keep in mind that Protector was written from scratch (about six years ago) as a bitboard-based program of its own. As a consequence it has a data structure that differs fundamentally from Toga's data structure. And of course Protector's algorithms refer to this data structure. For this reason it doesn't make sense to copy any *code* from Protector to Toga or vice versa.
This could seem to be a problem, but I think it isn't. Experience tells that chess programming is not a matter of coding but a matter of having clever ideas (or copying them ). For this reason in most cases it should't be a big effort to carry successful ideas from one program to the other.
(But, agreed, this still doesn't answer the question why I've written Protector instead of simply enhancing the Toga code. I don't think that this is a topic for the current discussion thread and for this reason I want to give only a short answer here. In fact I saw (and see) only one benefit in the option to start with the Toga code: saving a lot of time. On the other hand I saw a lot of benefits in a new program written from scratch, the most prominent being: 1) When I started with Protector back in 2003 there was no Toga or Fruit around 2) Bitboards. I like Bitboards and I think a modern chess program should be bitboard-based in order to benefit from 64 bit architectures as much as possible. 3) Learning curve. I felt that it's only possible to write a good chess program if you fully understand what's going on. One easy way to get this experience is writing a program from scratch
Coming back to the original question from K. Inuen: I guees it would take about half a day for a decent programmer with profound knowledge of the Toga data structures to attach the tablebase code. I assume the reason why this didn't happen until now is simply the copyright from Eugene Nalimov and Andrew Kadatch, which would't allow the Toga sources to be released any further under GNU licence.
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 12:18 am
by Graham Banks
For the record, Protector 1.2.7 seems to be around the same strength as Frenzee Feb08, Delfi 5.4 and Twisted Logic 20080620.
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 12:46 am
by CRoberson
The author claims it makes no sense to copy any Toga code given the difference in data structures.
This is false.
Due to good programming with levels of abstraction, one can change the base level data structures and the base level code while leaving the
higher level code untouched.
Thus, you could change the data structures for Toga and cut and past the search. Also, you could cut a paste most of the eval.
If you do this, then its a clone, because you have fully cloned the parts that make up the personality of the program.
The only remaining question is whether or not Toga/Fruit is written well enough to pull this off. The answer is YES.
To that end, I suggest that Sloppy is practically a clone because he cloned the eval. Sloppy's author clearly states that he cloned the
eval in a comment at the top of the eval code.
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 12:08 pm
by Guenther
Graham Banks wrote:For the record, Protector 1.2.7 seems to be around the same strength as Frenzee Feb08, Delfi 5.4 and Twisted Logic 20080620.
Here it is around the same strength as Toga-FruitII 1.41SE, which is of
course much stronger than your mentioned ballpark above.
(No surprise knowing its origin)
Guenther
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 8:40 pm
by Graham Banks
Guenther wrote:Graham Banks wrote:For the record, Protector 1.2.7 seems to be around the same strength as Frenzee Feb08, Delfi 5.4 and Twisted Logic 20080620.
Here it is around the same strength as Toga-FruitII 1.41SE, which is of
course much stronger than your mentioned ballpark above.
(No surprise knowing its origin)
Guenther
You could be right Guenther. I based my estimate on a small number of games.
Does Protector use ideas from Toga as opposed to copied code?
Cheers,
Graham.
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 10:42 pm
by Dr.Wael Deeb
Speaking of which,a new version of Protector is out....
Protector 1.2.8
http://sourceforge.net/projects/protector/
Dr.D
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 10:46 pm
by Graham Banks
Looks like linux only at this stage or am I missing the link to a windows exe?
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 10:49 pm
by Dr.Wael Deeb
Graham Banks wrote:
Looks like linux only at this stage or am I missing the link to a windows exe?
No,the Windows compiles are there Graham....
Dr.D
Re: What was the verdict on Protector 1.2.7?
Posted: Sun Oct 04, 2009 11:14 pm
by Eelco de Groot
Graham Banks wrote:Leo Dijksman
it seems nothing else then another TogaII, this time "bitboard based" with "many small" changed code!?
I asked some people but they say "its nearly ALL Toga", which sounds like the readme.txt part:
"In short it's kinda Bitboard-based TogaII with many small improvements."
I have not seen anybody point out specific code that is based on Fruit/Toga code rather than on ideas that could have come from Toga, or claim it should be published under the GPL from Fruit so maybe "some people" should be more specific then. If this wasn't discussed yet in previous threads about Protector. There is no reason to declare Protector a clone when it is using Toga's ideas rather than specific code.
Eelco