Re: new webpage: rikai.com

["Stephen J. Turnbull" <turnbull@example.com>]

Easy ones first.

>>>>> "Todd" == Todd Rudick <trudick@example.com> writes:

    Todd>   The analogy to bison is that a company can't distribute a
    Todd> program built with bison under a non-GPL license (although I
    Todd> think bison has changed now, no?).

That was true.  I'd have to check carefully, but I believe that what
happened was that the FSF changed the license terms for the files
bison.simple and bison.hairy, which contain the yyparse() skeleton.

Since Bison 1.24 it has been legal to use bison parsers in proprietary
programs.  From `info bison conditions':

       As of Bison version 1.24, we have changed the distribution terms for
    `yyparse' to permit using Bison's output in non-free programs.
    Formerly, Bison parsers could be used only in programs that were free
    software.

There is further rationale in the info file, but that's the bare fact.

Now for the hard one.

    Todd> << The bison analogy is probably false.  [...] >>

    Todd>   You misunderstood. What I mean is, if I GPLed rikai, and

I don't think I misunderstood; I just don't see this as a problem for
would-be competitors.  They're not going to try to compete by
restricting their customers' rights to reproduce the pages they have
glossed.

    Todd> if you run a page through the rikai engine, BECAUSE it now
    Todd> contains code, and that code is part of rikai, it (the
    Todd> result document) can't be released under an incompatible
    Todd> license. Someone wanting to do so would need to deal with me
    Todd> directly.

Right.  So what?  Maybe you have a different scenario in mind, but the
one I have is:  I go to www.M$-rikai.com, feed in
"http://turnbull.sk.tsukuba.ac.jp/" as the URL.  What comes back is a
rikai-ized HTML version of _my own home page_ with a notice at the top
that this document is GPL.  Does M$-rikai care?  I think not.  In most
cases I would expect rikai to be used for a single copy.  I suppose I
might want to redistribute the pages with the gloss included (eg, for
a class).  But I doubt they would lose all that much from the
relatively rare case where they could get multiple hits.

True, _you_ can in theory take advantage of the "bison <1.24" problem.
Under copyright law, without a license I wouldn't be allowed to
redistribute copies of my own rikai-ized home page, I suspect.  But
getting it from a GPL site would actually give me more rights, making
it more valuable to me.  To compete with those GPL sites, you really
would have to provide a special license on documents processed through
"rikai".  And you'd have to write that license _very_ carefully, or it
might become possible for M$-rikai to redistribute the code snippets.
(In the very hypothetical instance that anybody would actually pay
attention to your copyright on portions of "their" home page!!)

You could turn this to your advantage by allowing your customers to
restrict third parties, I guess.  But this is getting very
hypothetical!

On the other hand, in terms of creating an improved proprietary rikai,
if M$-rikai really wanted to use a different license, all they would
have to do is figure a way to replace (using sed in a separate
process) the code snippets that actually get inserted in the result
document.  This replacement code would be copyright them, and they can
multiply license it as they please!  Under that scenario, they could
make arbitrary changes to their rikai server, and not have to publish
them.  Dunno if this scenario would really fly, but it's very
plausible.  You don't want to fight that in court.

I don't think any of this bothers me, although it clearly weakens the
GPL as a tool for encouraging open source.  The point is that the GPL
probably hardly protects you at all from competition, even though
rikai code is being distributed in result pages in the current
implementation.

But I really don't think even the current situation where you don't
redistribute code at all protects you from competition.  GPLing is
hardly worse.

- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."