[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Barriers to an ASP loophole closure

Jeremy Hankins <nowan@nowan.org> writes:

> bts@alum.mit.edu (Brian T. Sniffen) writes:
>> Jeremy Hankins <nowan@nowan.org> writes:
>>> Is it users of programs or owners of copies of programs that should
>>> have freedom?  As far as I can see the answer is clearly users.
>>> Currently those two groups are roughly the same, and the second
>>> group is *much* easier to draw a line around.  So we use ownership
>>> of a copy to pin freedoms to.
>> The two groups are vastly different.
> Only when you're playing the game of trying to push the definition of
> "user" as far as you can push it.  And that's a perfectly legitimate
> and good thing to do when you're discussing a license text, but in
> doing so you shouldn't forget that there's also an ordinary, every-day
> definition that doesn't get pushed so far.

When I say you're a user of router software, I'm not pushing the
definition of user any further than you are when you say I'm a user of
PHP-nuke or Apache.

> The ordinary, every-day use of the term is fuzzy, acknowledges
> inconsistencies, and works anyway.  Ask the next guy you meet that
> uses computers but doesn't care about licensing whether they're using
> apache when they look at apache.org.  Do you honestly think he'd say
> yes?  Or the NYT's typesetting software when they read the paper?

I think he'd say he doesn't use either.  But then when I asked him
what he used to read his mail, he'd either say "Oh, I use Netscape" or
"Oh, I use Yahoo! mail".  I've observed a number of the nontechnical
users I support failing to distinguish between their OS, browser, home
page, and other web sites.  I'm not sure we should be guided by such
opinions, though: it would be nice to meet the expectations of the
uninformed masses, but given the choice of their expectations, which
are often internally inconsistent, and the expectations of those who
will actually be modifying or distributing code, I'd rather satisfy
the latter.

In particular, it seems to me that drawing a technical distinction
among users (i.e., to say that users at a console or over an IP
network are users who get source, but users via sneakernet or over a
voice telephone link are not) is unwise.

>> Maybe.  I don't believe it's possible to delineate users in a way
>> that doesn't discriminate strongly against fields of endeavor or
>> against particular technologies.  I think your three tests summarize
>> what's needed pretty well... I just don't think they're mutually
>> compatible.
> Maybe not, I'm not sure.  It's hard to prove a negative.
>>> So we have three meta-options:
>>> * Decide that freedoms *should* attach to copies rather than users
>> There's something to be said for that option.  Here's a further
>> extrapolation of the ASP nightmare for you: everybody who writes code
>> makes it available only on their machine, in some .NET horror by which
>> code is patched at runtime via RPC... so I want to set up a
>> webserver.  It uses Apache.NET, Zope.NET, ZWiki.NET, Perl.NET, and
>> each of these (and each Perl module and Apache module) lives on its
>> own separate server.
>> To what are the users of my site entitled?  My glue code?  The kernels
>> of each of these servers?
> Yeeks, dunno.  Probably your glue code, but once I install that myself
> and use the other sites I'm using them as well (i.e., if they're under
> ASP-loophole-closing licenses I get access to that code too).

Oooh.  Neat point.  So if I access a site -- even just far enough to
try to authenticate myself and be refused, then use that refusal in
some computation -- I can demand access to the source.  That's no
good.  This very quickly hits the case where "If you make it available
to anyone, you must make it available to everyone," which I think we
all agree is not acceptable.

Why not the code of the routers in between?  Why not the code
supplying power to the computers -- or does that count as a major
component of the operating system?

What happens twenty years from now, when Transmeta-style reconfiguring
processors are everywhere, and I'm not so much running emacs as I am
rebuilding my computer into a fixed machine which implements emacs?


Brian T. Sniffen                                        bts@alum.mit.edu

Reply to: