Subject: Re: [fsb]Re: Franklin Street Statement and Free Network Services
From: "Stephen J. Turnbull" <>
Date: Tue, 02 Sep 2008 13:18:45 +0900

Russ Nelson writes:

 > We don't have cost-free publication of binaries either, so obviously
 > any software published in binary form is not gratis software.  But we
 > don't give a whit about gratis software.  It's libre software we care
 > about.  And what could be MORE freeing than being able to get the
 > source of code even if you haven't gotten the binary?

 > Remember: it's not about freedom for the developer.

C'mon, Russ.  Be civil.  MJ is not one of those who needs reminding.

And you're wrong, anyway.  None of the definitions of software freedom
(whatever it's called) permit excluding people just because they're
software developers.  The question is, as always, when the exercise of
freedom by upstream users inhibits the freedom of downstream users,
who gets precedence?

The answer of the "permissive" school is "upstream: the only right we
take from those who choose non-libre distribution modes is the right
to use words like 'libre', 'open source', or 'free' to describe their
distribution; downstream can and should demand libre terms depending
on their needs."  The answer of the-school-I-like-a-lot-less is
"downstream: those users need to be forced to be free because they are
generally unwilling to pay the cost of negotiating a free distribution
mode with upstream."[1]

Licenses implementing *both* extremes impose externalities.  Strong
copyleft licenses take away my freedom to negotiate away my right to
source in return for a lower price, or even the very existence, of a
derivative work.  Permissive licenses, on the other hand, allow
intermediaries to refuse access to source of derivatives, and thus
restrict the freedom of downstream users who do care to get source.

I don't see an a priori liberal answer to this problem of the second
best, unless you're willing to go down Stallman's narcissistic path of
asserting that there is no freedom but software freedom that needs to
be considered here.

 > Bittorrent allows you to control the cost of that copying, however.
 > Buy hosting that allows X megabytes of download per day (which, for a
 > small value of X, is extremely cheap), then configure bittorrent to
 > stay under X.  That doesn't limit the number of copies (which I think
 > you would agree is improper), but it limits your cost.

You will probably argue that users (ie, those who want to download via
bittorrent) are free to find better ISPs, but in my experience ISPs
often interfere with bittorrent.  Bittorrent also sucks for long-term
support, as it can take forever to download if all you've got out
there is the original seed.  This is a PR liability that must be
borne.  There's also the permanent legal/admin cost of keeping track
of all these specific rules.  I don't think it's fair to just pooh-
pooh the costs proposed to impose on others, when a careful study of
what those costs might be has not been done.

[1]  Note that I do not object to copyleft licensing, or Affero
licensing, if that's what the owner of the work wants.  I just don't
consider those licenses "better" from the viewpoint of enhancing
freedom than the permissive ones.