which certainly don't look convincingSalut,
Deux raisons:
1- thinkingalot
2- Grapefruit
Patrick


P.S. Gull 0.8 is out: https://sourceforge.net/projects/gullch ... p/download. Estimated strength: 2900 - 3000 CCRL.
which certainly don't look convincingSalut,
Deux raisons:
1- thinkingalot
2- Grapefruit
Patrick
Go and click on "mp" under his post.ThinkingALot wrote: Does anybody know how to contact him in order to clarify this issue? I failed to find the e-mail.
.
Thanks!ThinkingALot wrote:Gull 0.9 is out: https://sourceforge.net/projects/gullch ... p/download. This is a bugfix release. 10 - 30 ELO increase is expected.
At least it would be nice to know what license you are using -- I couldn't find that.Martin Thoresen wrote:Thanks!ThinkingALot wrote:Gull 0.9 is out: https://sourceforge.net/projects/gullch ... p/download. This is a bugfix release. 10 - 30 ELO increase is expected.
I would very much appreciate some sort of readme file in the archive, with version changes and general info...
I plan to add a readme file to the first "release" version: 1.0.Martin Thoresen wrote: Thanks!
I would very much appreciate some sort of readme file in the archive, with version changes and general info...
Gull is registered at sourceforge.net as a public domain project.UncombedCoconut wrote:At least it would be nice to know what license you are using -- I couldn't find that.
At least in the nearest future.UncombedCoconut wrote:Another question, do you intend this to be a Windows-only engine forever?
Simple assembler functions can probably be borrowed from anywhere. I don't think it would be a GPL violation to use these functions and keep the code public domain.UncombedCoconut wrote:I've checked and it's doable... if your license lets you borrow the ASM from a project like SF, I'm willing to try it.
Porting prefetch isn't absolutely necessary. This function just gives some speedup. It doesn't change the search logic.UncombedCoconut wrote:The parts that aren't totally trivial are BSF/BSR, TT prefetch, GetTickCount(), and non-blocking read from stdin.
I am sorry but I don't agree here.ThinkingALot wrote:Simple assembler functions can probably be borrowed from anywhere. I don't think it would be a GPL violation to use these functions and keep the code public domain.
Sure. However you forget about two important issues.mcostalba wrote:I am sorry but I don't agree here.
In no part of GPL licence it is written that if you copy just a small part of a GPL sorce you are allowed to skip GPL enforcements.
I would rather suggest you to careful read this:ThinkingALot wrote:Sure. However you forget about two important issues.mcostalba wrote:I am sorry but I don't agree here.
In no part of GPL licence it is written that if you copy just a small part of a GPL sorce you are allowed to skip GPL enforcements.
1) If some code is published under GPL it only means that all its parts comply with the GPL or some "weaker" license. This "weaker" may be "public domain". It's pretty plausible that all the assembler intrinsics from SF were implemented long before its appearance.
2) I'm not sure, but "obvoius" algorithms/ideas probably can't be copyrighted. Correct me If I am mistaken.
P.S. Could you please provide a link to the CCC discussion on this topic?
The things he's talking about are already public domain. You can't take public domain things put them in GPLed program and claim GPL. Things simply don't work that way.mcostalba wrote:I am sorry but I don't agree here.ThinkingALot wrote:Simple assembler functions can probably be borrowed from anywhere. I don't think it would be a GPL violation to use these functions and keep the code public domain.
In no part of GPL licence it is written that if you copy just a small part of a GPL sorce you are allowed to skip GPL enforcements. Actually it does not change anything if you copy a big or a small part, a difficult or a trivial one, in any case you are forced to accept GPL terms, i.e. to release also your sources under GPL license.
If you want to avoid this simply do not take any code chunk from GPL sources. There are no other legal ways to overcome this. The "quantity" of code copied is not a valid argument at all.