Making a 'publish it as an RFC or discard it' claim may be nice
rhetorically, but it's not a very useful answer to the problem
being raised.
So, lots of places have facilities for publishing technical reports.
If an I-D author believes that something should be kept around but
for whatever reason (e.g., lack of concensus) it is not going to be
published as an RFC let the author stamp a TR number on it. Of
course, these would not be as easy to find as if we just kept old
I-Ds, but a short note to the appropriate WG mailing with a URL
would keep a reference in the mail arhive, at least.
I've used the above system for a document that consensus dictated be
winnowed down from its original wonderfulness. :)
It's not necessarily a perfect solution, but neither is keeping
every I-D ever issued.
allman
---
http://roland.grc.nasa.gov/~mallman/