In defense of Pedro

[Ovyron]

I personally think copying code is wrong, and copying a net is wrong.

When something is wrong there's always a victim. Who's the victim on this case, exactly?

[algerbrex]

I think there are three dimensions to this issue: Legal and ethical, and personal preference.

I can't particularly speak to the legal dimension, but I assume that at least some of the various licenses can still be used to cover the usage of NNs from other engines. So that's what I would let dictate my legal opinion.

Ethically, as HG said, different tournaments will have different rules. And as long as these rules are respected, I don't think there's any issue. Also, I think common courtesy and the wishes of different engine authors should be respected when reasonable. So if engine author X shares that they don't mind others using the game data they generated and their evaluation, but would prefer if others didn't directly copy their NNs and publish them officially as part of their engines, I don't see a huge issue either.

As far as personal preference goes, I would prefer some hybrid between 2 and 3. Personally, for Blunder, I'd like to use my own HCE to generate my own games and to train my own NN. I'd like the experience and challenge of learning how to do all of that on my own. But I also recognize for many people, including myself, that accomplishing such a feat could potentially be very computationally expensive, and make it difficult for many to experiment with NN. So I can't fully layout what my approach will be right now (unless I'm willing to splurge a bit on some better hardware), and I think others should be free to choose the option easiest for them.

Of course, I'm not a strong engine author, so take my views with a grain of salt, but that's my two cents.

[Frank Quisinsky]

To use the knowledge of generations in chess programming but to condemn others in own contributions is reprehensible.
Disrespectfully with a lot of double moral.

I await a bit more humility for the presents we got from a big group of experts in programmings from the past.

Furthermore, I await more understandings.
At first ... the other could be right!

[Steve Maughan]

I personally think copying code is wrong, and copying a net is wrong.

When something is wrong there's always a victim. Who's the victim on this case, exactly?

I agree you need a victim. To expand on what I was saying: by saying it's "wrong" I mean that claiming your engine is your own works while copying code or copy a net is wrong. So the end user that thinks they're using a new original engine is the victim. Even if the engine is free you're wasting their time. If you say you've copied the code or the net then that's OK, but don't claim it a completely original engine. Put another way I don't think an engine with either copied code or copied nets should be able to enter the WCCC. For other tournaments it's up to the organizer but I wouldn't want to see a load of clones or partial clones in the WCCC. However, using Stockfish in any way in the development and testing of the engine is OK (IMHO).

Best regards,

Steve

[Ovyron]

So the end user that thinks they're using a new original engine is the victim.

Okay, so we're on the realm of plagiarism. The problem is not that people are doing such and such and such, but that they're claiming they're doing something else, and we don't have a way of knowing what they did.

[dkappe]

So the end user that thinks they're using a new original engine is the victim.

Okay, so we're on the realm of plagiarism. The problem is not that people are doing such and such and such, but that they're claiming they're doing something else, and we don't have a way of knowing what they did.

Well, there are those that claim that their license (open database, GPL, etc.) says something it doesn’t say. That is also wrong.

[Frank Quisinsky]

GPL:
Stallman's "code of conduct".
More his problem with Lawyers.

Basically, the ideas are good.
In praxis is the meaningfull questionable!

If 66% of open-source used GPL3 is that not a free ticket.
Unfortunately, too many legal violations and very welcome for each lawsuit.

More or less ...
If 30 person jump from the bridge ... number 31 is thinking ist must be good!

[Rebel]

To use the knowledge of generations in chess programming but to condemn others in own contributions is reprehensible.
Disrespectfully with a lot of double moral.

I await a bit more humility for the presents we got from a big group of experts in programmings from the past.

Furthermore, I await more understandings.
At first ... the other could be right!

You are told a fact, that FireNN uses the SF12 network, what you do with that information is your business. Nothing to with your above character qualifications of those who teach you something.

[Frank Quisinsky]

Be right or not ...
On third position!

Understand reasons of others ...
On second position!

And on first position:
Is the situation important for me and my own contribution.
What it does with me?

Wrong is:
Should I start an official campagne to a witch burning?

More wrong is:
If so ... I need others for more attention.
With the power of others "witch burning" is more welcome!

Pure dogmatism never changed any situation if third position in foreground!
Often it's better if you are right not to get right.

Best
Frank

PS: ... without Hello Ed ... you are older!
Not for you directly!

[Modern Times]

You are told a fact, that FireNN uses the SF12 network, what you do with that information is your business. Nothing to with your above character qualifications of those who teach you something.

I had thought that the network in question had not been used at all by Stockfish. Still, SF12 is four versions old now - not sure how much the NNUE has changed in that time - so this doesn't change much for me. I think Norm has demonstrated very low sim test results Fire vs SF, demonstrating that just because a similar NNUE is used does not mean the engines become similar even though it was a good Elo gain for him.