[rfc-i] URIs in RFC references, was: feedback on draft-iab-styleguide-01

Dave Crocker dhc at dcrocker.net
Tue Mar 25 09:51:05 PDT 2014

On 3/25/2014 9:00 AM, Ted Lemon wrote:
> Why wouldn't the link lead to a document in the same format as the document from which the link came?   I'd expect PDF to link to PDF, HTML to link to HTML, etc.   The only one that would really work nicely would be HTML, of course, unless you had a special PDF reader that fetched the URL rather than launching the browser.

+1, although there is a competing concern about having multiple 
"primary" links for a document, which could be confusing.

Still, the premise of your point is to consider what the user will 
expect.  I wonder whether there are common conventions for handling the 
issue of documents in multiple forms, and what template for links to 
them to use?  If there is already common practice, we should use it.

URLs for RFCs:

      1.  A link for a document should produce that document.  It should 
not produce a page requiring another keyclick, before getting that 
document.  Contrary to many of the folk in this discussion, regular 
users do not like being given lots of choices all the time; they see it 
as burdensome.

      2.  We now have 3 different 'base' URLs for RFCs.  rfc-editor.org, 
tools.ietf.org, and datatracker.ietf.org.  This smacks of incoherent 
organization by our community, adding ambiguity and wasteful operations. 
  We need to standardize on a single, common base reference.  IMO, the 
datatracker interface is by far the best of the 3:

         http://datatracker.ietf.org/doc/rfc2821/ leads directly to a 
page with a rich set of meta-data as well as at least the initial part 
of the document, though I'd wish the "?include_text=1" weren't needed in 
the URL to get the whole document.

        http://www.rfc-editor.org/info/rfc2321 replicates access to 
meta-data.  Not sure whether it is as complete, but I do see that it 
requires another click to get at the actual document.  Bad.

        http://tools.ietf.org/html/rfc2821 produces the whole document, 
but appears to have a subset of the meta-data.  (No, I didn't do a 
careful audit.)

The IETF spends scarce resources and expensive money to maintain all of 
this.  It's wasteful and confusing.


Dave Crocker
Brandenburg InternetWorking

More information about the rfc-interest mailing list