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

Re: Is Debian a common carrier? Was: package rejection

On Tue, Dec 07, 2004 at 02:36:35PM -0600, Manoj Srivastava wrote:
> On Tue, 07 Dec 2004 11:41:42 -0800, Bruce Perens <bruce@perens.com> said: 
> > I don't think we have the slightest chance of proving to any court
> > that Debian is a common carrier, given the several inches of policy
> > manual that specify the nature of the content, etc.
> 	Say what? Where is this policy that specifies on the nature of
>  content? I see a technical policy that specifies on how something is
>  packaged, but nothing at all that states what the content may be. The
>  closes I can see is stating what licese something is released under.
> 	Oh, and if we do not specify what the nature of what we
>  package, would it be easier to prove we merely carry packages?  That
>  would really be nice.
> 	manoj

I work for one of the prime examples of a common carrier (and in fact,
one of very few proven by case law): a former Regional Bell Operating
Company (RBOC) / current Incumbent Local Exchange Carrier (ILEC) with a
Long-Distance affiliate. FCC stamped, approved, and intended. Or, in short,
(a) the Telephone Company.

The rules around what constitutes a CC are *very* thorny, and it's very
easy to screw up (and very painful to do so, as well, due to the liability
issues). But the primary set of conditions are as follows:

* You must carry traffic between points for third parties (with narrow
  exceptions for control traffic, any traffic you origionate or receive, you
  are liable for as a normal sender/receiver).

* You may not choose to carry, or not carry, traffic on any basis except
  capacity to do so (note that your contract with a customer can limit
  capacity available to them). The key point here is that you not only
  can't make judgement calls on whether to carry or not carry something
  based on signal content, in general you can't even LOOK at the signal
  content. About the only exception to this is routing information
  related to the traffic (for the obvious reason that you have to know
  where to send it, and said content is thus implicity directed to you as the

* In general, you are heavily restricted on what criteria you may apply
  when deciding who will be a customer in the first place; "common" is
  the important word here, with it's implication that you provide a
  service to the commons, and that denial of said service requires a
  reasonable justification (such as failure to pay, or an expectation of
  an inability to pay, for the service).

None of this is legal advice in the slightest; if you want that, in regards
to common carriers, you're going to be paying very specialized lawyers a
very large amount of money.

The fact that many ISPs have networks that can, in theory, operate as
"common carriers" (if you do routing purely on traffic routing data, and
utterly ignore both content, and consequences of delivering the content),
does not mean they qualify (as has been pointed out, so far no case law has
established them as such, and it appears to be running the other direction,
for the most part). Most ISPs run enough filtering simply as a requirement
to be able to move traffic at all that they are worlds away from CC status.

In my estimation (based primarily on having worked for national, regional,
and local ISPs, and a telco, including all of the big lengthy legal
warnings and requirements about what one can do or not do when dealing with
customers or their data), Debian isn't even playing the same *sport*, much
less is it in the ballpark - and it doesn't want to be, either.

You could argue that an archive which accepted uploads of DEB packages from
anyone at all, with no review and only automated enforcement of certain
base sanity checks such as "Is it a well-formed DEB package" - *not* "does
it meet the following policies" - could qualify as a CC, but frankly, the
term was never meant to be applied to such a beast, and I strongly suspect
the US Court system would take note of that when deciding. And it would be
a damned useless archive anyway, so.
Joel Aelwyn <fenton@debian.org>                                       ,''`.
                                                                     : :' :
                                                                     `. `'

Attachment: signature.asc
Description: Digital signature

Reply to: