[Fsfe-ie] Re: ethical interpretations of FS

[Fergal Daly]

On Tue, Feb 03, 2004 at 03:46:24PM +0000, Ian Clarke wrote:
> It is the definition of "using" that we differ on.  If your definition 
> of "using" is that you run it on your computer, then I agree, but if 
> your definition of using is that you have some interaction with a remote 
> server on which it is running, then I don't.

It's not my definition that I'm talking about. It's the FSF's and what I can
find on their website implies that remote use is included and will be
specifically catered for in the next edition of GPL.

> I do not accept that it can be the considered opinion of the FSF that no 
> software can ever be kept secret, because I don't think it is possible 
> to make that argument for software without making it for all types of 
> information.  If you disagree with this, you will have to explain how 
> you decide what information can be kept secret, and what information 
> can't, and justify the distinction.

They don't quite say no software can ever be kept secret. They say it should
not be kept secret from it's users. So they do not want your source code
unless they are actually using your software. If you keep it to yourself or
you manage to keep the number of users limited then they would say you have
no responsibility to anyone outside that group of users.

I don't follow how this spreads to all types of information but I won't try
to argue one way or the other because I am not the FSF, I am just pointing
out what they have said. If you believe that these ideas are fundamentally
flawed and logically leads to all information being public then either I've
misrepresented the FSF's ideas or you disagree with them. That's all I'm
trying to point out.

> To prove they are not isomorphic you must explain why your argument 
> applies to software but doesn't apply to all other types of information. 
>   How do you draw a distinction, and how do you justify that distinction?

No, to show 2 things are not isomorphic all I have to do is find a single
thing that's true for one of them and false for the other. Here's the
simplest difference I can think of "all passwords available to everyone"
implies no password protected bank accounts. Whereas "all source code
available to users" does not imply no password protected bank accounts (or
at least I cannot see why it would).

This shows that facts proved about one don't necessarily carry over to the
other (in either direction).

You are appealing to analogy/isomorphism to transfer an argument from one
domain to another. This is a very powerful tool for arguments but it is only
valid if the isomorphism really does hold. This power comes at a price, if
someone can poke a single hole in your isomorphism then you lose it
completely. I have given a simple example where your isomorphism doesn't
hold. You either need a new one or you need to start arguing directly in the
domain at hand, without making analogies.

> >I did not say that public interest
> >criteria _should_ be applied to every aspect of life I showed that applying
> >public interest criteria to these 2 situations gives quite different
> >results. Therefore they are not isomorphic.
> 
> You are arguing that public interest criteria should be applied to the 
> secrecy of all software.  I am asking why, if you believe that, the same 
> criteria shouldn't be applied to all other types of information.  If 
> they can't, then I am asking where, how, and why you draw the line 
> between the types of information that can't be secret, and the types 
> that can.

Let me correct that paragraph.

--
I did not say that public interest criteria _should_ be applied to __any__
aspect of life I showed that applying public interest criteria to these 2
situations gives quite different results. Therefore they are not isomorphic.
--

I was not arguing about _should_ or _shouldn't_, I was only showing that
your isomorphism does not hold and so you cannot automatically transfer
conclusions from one idea to the other. In particular, showing that "no
secret passwords" is bad does not prove or disprove that "all source code
available to users" is bad.

I should have used a simpler counterexample, one that didn't try to make
another point at the same time.

> >I agree that "no personal secrets" is a bad thing but it is _not_ 
> >isomorphic
> >to "all source code must be free for all users". If you want to argue about
> >the second idea, then you will have to argue about it directly and avoid
> >using analogies that are not applicable,
> 
> If you want to argue that software can't be kept secret, but other types 
> of information can, then I can ask you to justify the distinction, ask 
> you where you draw the line between what can and can't be kept secret, 
> and why.

This is what the FSF are doing but (rahter unwisely) I'll give it a go too.

Program: instructions which encode an algorithm which determines the actions
of a computer.

User of program A: a human who interacts with a computer who's output is
materially determined by program A. I say materially determined to exclude
stuff like the effect your private number crunching program may have on the
execution of program A.

Now, let's say I make a law that says any user of program X is entitled to
the source code for program X. Why would that imply that I am entitled to
your password or some steamy celebrity's videos? Even if you argue that your
password or an mpg is in some way a program, I am not a user of them and so
I am not entitled to them or their source code,

F