ietf
[Top] [All Lists]

Re: Getting the latest version of an RFC specification

2017-03-29 07:20:06
Hi, Dave

This is definitely doable, although I’d prefer that /latest redirect to the 
latest rather than display it.

I question how useful it is.  The obsoletes/obsoletedBy relationship is 
semantically overloaded. Consider the header of RFC 4306:

   Obsoleted by: 5996
   Updated by: 5282
   Request for Comments: 4306
   Obsoletes: 2407, 2408, 2409

RFC 5996 defines IKEv2, same as RFC 4306, but the three RFCs that 4306 
obsoletes define the protocol IKEv1, a different protocol.

So what should https://tools.ietf.org/html/rfc2407/latest 
<https://tools.ietf.org/html/rfc2407/latest> return?  2407, because that is the 
latest RFC defining IKEv1, or 5996 (actually, 7296) because it’s the latest IKE?

The 822 —> 2822 —> 5322 path is one we should follow. I’m not sure the same 
applies to different versions of a protocol.

Yoav

On 29 Mar 2017, at 6:51, Dave Crocker <dhc(_at_)dcrocker(_dot_)net> wrote:

G'day.

The RFC labeling model is to assign a unique serial number to a static 
document.  A new version of a spec gets a new serial number. This basic model 
has the benefit of both simplicity and predictability.

To this we've added an overlay model, using Obsoletes/ObsoletedBy. This makes 
it dramatically easier to see that something has been obsoleted and to find 
its replacement.

However the seeing and the finding are an essentially manual process. One 
must go to the online older document, then notice the Obsoleted By tag and 
then click to follow it.

Sometimes it would be helpful for the requester to be able to say 'give me 
the latest' more easily.

So I'm wondering whether the IETF should consider adding a citation feature 
for this.

Something like:

    https://tools.ietf.org/html/rfc822/latest

would display the contents of:

    https://tools.ietf.org/html/rfc5322

by having the fetching system automatically traversing the Obsoleted By links 
in RFC 822 and then RFC 2822.

Some sort of display banner would flag this, to help the user see that they 
are getting a different version than they cited.


Thoughts?

d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net


Attachment: signature.asc
Description: Message signed with OpenPGP