[rfc-i] issue: canonical formats

John Levine johnl at taugh.com
Fri Jun 1 10:12:51 PDT 2012

>"While the canonical source format MUST be easily converted in to a variety
>of other formats, a single canonical display format must exist to satisfy
>the requirements of legal and content disputes"

I see that, but based on my experience in court, I think it's wrong.

The IETF needs to be able to hand a document to a court and say this
is RFC 9999, but the IETF is also allowed to use a small amount of
intelligence and send an appropriate version.  In the current model,
is the canonical version the file of ASCII characters and formatting
codes?  Is a printout of that file canonical?  If an attorney makes a
PDF, by using enscript, by printing one of the web page versions of
RFCs, or by scanning a printout, and then uploads the PDF to the PACER
court document system, which I can assure you is what would happen to
any RFC used in federal court, is that canonical?

The answer is that so long as they all say the same thing, it doesn't
matter.  Ditto if one version happens to be fixed pitch ASCII and
another one a PDF made from proportionally spaced HTML.

>I'm more worried about contracts than IPR.  The use case is that a contract
>specifies that software X complies with RFC Y.  How does an expert witness

She looks at the document, and if there is doubt about its accuracy,
at the XML or whatever authoritative version.  Having done rather a
lot of expert witnessing, I can report that this sort of thing simply
is not a problem.  Courts are not stupid, and sorting out variant
versions of documents has been a well understood problem for
centuries.  It used to be a lot harder, when copies were made by hand.


More information about the rfc-interest mailing list