Programmer code of honor (update)

[Rebel]

http://www.top-5000.nl/programmer_code.htm

So far 18 programmers subscribed to the code. Thanks all those who support the initiative.

Still a lot of names are missing and I want to collect the objections in order to investigate if the code needs an update.

So please contact me by email, PM or put your objections here.

Thanks.

Ed

I suggest you add the year when the form was sign ... One can swears for now and the past but an original work can become a derivative later ...

+1

I wonder if a derivative can becomes an inspired or an original work

Ferret started from GNUchess and became original.

[hgm]

1. You never used an IBM PC compatible computer? Isn't your current PC not a derivative? By logic like this drop your PC out of the window right now?

Utter nonsense. PCs are not a copyright or patent violation. Adopting a compatibility standard is not copying. As it happens, I have studied the schematic of the original IBM PC in the 80s, because I was developing hardware for the extension bus. I can assure you there isn't anything in there that is patentable, it is all standard stuff taken directly from the Intel application notes of the involved chips. Of course the exact layout of the circuit board is subject to copyrights, but all competitors used their own design for that.

2. Shall the Americans give back the land to the Indians? It's stolen land.

Good idea. But not my problem.

3. Never buy something from Microsoft they stole Stacker.

What the heck is stacker? You can rest assured that I have not consciously bought that, and if they bundle it with Windows, I probably never use it. Actually, I am well aware that MicroSoft is a bunch of crooks, and if I can screw them in any way, I don't hesitate to do so. The only money they have ever seen from me is what they collect from the OEM installs of Windows that you cannot avoid buying together with most hardware. But buying stuff from a known crook that he did obtain legally doesn't make you a fence.

etc.

Well, better elaborate on the etc. to come up with something better. Because so far it sucks.

[tpetzke]

This is a bit dangerous because if you publish the source of your non competing product the idea might get extracted and used by others back in the original competing area. They might not even realize that they are relying on stuff that was extracted without the permission of the original author as they only look at legally obtained source code.

Even if you don't publish the source the idea might be much more obvious in you product than it was in the originals ones.

If want to know something ask the author for permission to fiddle around with its stuff, most likely you don't have to RE then because the author will tell you what you want to know. Just ask nicely. And if he denies it after all, respect his wishes.

Thomas...

[lkaufman]

It is obvious that you and Don mean two different things by "RE". I often try to discover things about other programs by observing their behavior, which is what Don refers to when he says I do RE. I am utterly incapable of doing disassembly myself, which is what you define as RE. I do feel free to make use of any published comments by anyone who himself has done RE. So at the very minimum you should change RE to disassembly in the document.
But I would also say that I see nothing wrong with using disassembly on any program that itself is based on a program which did use this technique. Since this applies to so many engines now the whole idea is pretty much moot. I also think it makes no sense to prohibit something that is completely undetectable and unprovable. You cannot tell if someone desassebled another program, you can only tell if they copied the details to a point that proved that they had to have decompiled the program.
So I would rather see something like a prohibition on copying exact details (parameter values for example) from another program without permission than a prohibition on decompiling. The former is verifiable (if only by decompiling !), the later is not.

[SuneF]

It is obvious that you and Don mean two different things by "RE". I often try to discover things about other programs by observing their behavior, which is what Don refers to when he says I do RE. I am utterly incapable of doing disassembly myself, which is what you define as RE. I do feel free to make use of any published comments by anyone who himself has done RE. So at the very minimum you should change RE to disassembly in the document.
But I would also say that I see nothing wrong with using disassembly on any program that itself is based on a program which did use this technique. Since this applies to so many engines now the whole idea is pretty much moot. I also think it makes no sense to prohibit something that is completely undetectable and unprovable.

To me RE means disassebling the code.
I do belive there are "ethical" isssues with doing it.

It takes a long time and hard work to come up with real novelties in computer chess. When it does happen you want to keep your findings to yourself for as as long as possible. Otherwise you are just raising the bar for everyone instead of leaping forward yourself. To most serious contenders this would simply mean you have wasted your time.

I mean, imagine if Gelfand could have see all Anand's opening preparations during the WC. Anand would simply have wasted his time for nothing.

I don't see how RE can be defended. It is simply looking at someone elses ideas without their permission. They didn't mean to tell you about their tricks, if they did they could have made the engine open source or written about it on the forums.

Illegal perhaps not, unethical absolutely!

You cannot tell if someone desassebled another program, you can only tell if they copied the details to a point that proved that they had to have decompiled the program.
So I would rather see something like a prohibition on copying exact details (parameter values for example) from another program without permission than a prohibition on decompiling. The former is verifiable (if only by decompiling !), the later is not.

I think the odds that the exact same values will fit into your own engine is tiny, a re-tuning would be needed which makes detection much harder.

[lkaufman]

I would be happy if decompiling were impossible. I expect that once Komodo is number one others will try to decompile it to learn its secrets. But there is nothing to be done about this. To me the analogy is if Anand and Gelfand were forced to do their preparation with their seconds with open mikes that the other side could hear. They would just have to adapt to the situation.
So while I agree with you in principle, I don't like unenforceable rules that only harm those who chose to abide by them. I agree with Richard Vida, it's not the act of decompiling that is wrong, it's a question of what you do with the information. If you copy original ideas in detail, that is provable (with enough work and decompiling) and wrong. If you just get some general idea, let's say it inspires you to work more on time management for example, that is unprovable and in my view not objectionable. Of course there is no absolute rule that says what is ok and what is not, it's a matter of judgement.

[syzygy]

I also think it makes no sense to prohibit something that is completely undetectable and unprovable. You cannot tell if someone desassebled another program, you can only tell if they copied the details to a point that proved that they had to have decompiled the program.

In general such acts are neither undetectable nor unprovable. A person committing the act might e.g. openly admit it, or the results of the act or traces of it can be detected on his computer system, or in documents or e-mails.

[Houdini]

I also think it makes no sense to prohibit something that is completely undetectable and unprovable. You cannot tell if someone desassebled another program, you can only tell if they copied the details to a point that proved that they had to have decompiled the program.

In general such acts are neither undetectable nor unprovable. A person committing the act might e.g. openly admit it, or the results of the act or traces of it can be detected on his computer system, or in documents or e-mails.

A couple of real examples:

1) Regarding Larry.
There was a public forum discussion in which Larry encouraged other persons to disassemble Houdini and discuss the findings. He even reported back on testing certain changes to Komodo as a result of these discussions.

2) Regarding Richard.
There was a public forum discussion in which Richard discussed his RE of Houdini 1.5.
He has always claimed nothing of his findings ended up in Critter, but it would be very easy for me to demonstrate that Critter post version 1.0 contains a non-trivial number (more than 5) of ideas/code taken directly from Houdini 1.5a. The OpenCritter Pascal source codes are quite revealing.

The bottom-line is that this "code of honor" remains empty talk when the #2 and #3 engines freely use RE of the #1 engine, apparently to the universal approval of the forum members.

Robert

[diep]

I also think it makes no sense to prohibit something that is completely undetectable and unprovable. You cannot tell if someone desassebled another program, you can only tell if they copied the details to a point that proved that they had to have decompiled the program.

In general such acts are neither undetectable nor unprovable. A person committing the act might e.g. openly admit it, or the results of the act or traces of it can be detected on his computer system, or in documents or e-mails.

A couple of real examples:

1) Regarding Larry.
There was a public forum discussion in which Larry encouraged other persons to disassemble Houdini and discuss the findings. He even reported back on testing certain changes to Komodo as a result of these discussions.

Disassembling houdini isn't needed. it's total trivial that it's using the same evaluation like all the clones do.

2) Regarding Richard.
There was a public forum discussion in which Richard discussed his RE of Houdini 1.5.
He has always claimed nothing of his findings ended up in Critter, but it would be very easy for me to demonstrate that Critter post version 1.0 contains a non-trivial number (more than 5) of ideas/code taken directly from Houdini 1.5a. The OpenCritter Pascal source codes are quite revealing.

The bottom-line is that this "code of honor" remains empty talk when the #2 and #3 engines freely use RE of the #1 engine, apparently to the universal approval of the forum members.

Robert

With Richard i assume you mean Richard Vida. Is that correct?
Can you point out which ideas Richard Vida took over?

[Houdini]

Disassembling houdini isn't needed. it's total trivial that it's using the same evalution like all the clones do.

You haven't got a clue.

With Richard i assume you mean Richard Vida. Is that correct?
Can you point out which ideas Richard Vida took over?

What exactly do I gain by pointing out in detail what was copied?

Robert