[rfc-i] errata maintenance

[Alex Rousskov]

On Tue, 13 Apr 2004, Bob Braden wrote:

> We are puzzled about what you are concerned about here.  What do you
> mean by "maintenance"?

Pretty much everything including errata solicitations, submissions,
review, publication, and search. As usual, there is hardly a single
problem that can be easily described (not to mention fixed) in
isolation:

0) publication/marketing

  In my experience, only a negligible number of folks who read and
  implement RFCs know that there might be an errata for those RFCs.
  In other words, errata does not exist for any practical purpose.
  Side effect: folks do not submit new errors.

1) solicitation

  There is no meaningful solicitation of errata from RFC readers.
  While the "Status of this Memo" section says that the document
  "requests discussion and suggestions for improvements", there
  is no obvious way to discuss and suggest short to spamming
  authors which most readers will not do for many reasons.

2) submission

  If a user is lucky to find submission instructions[1], submission
  itself is easy but informal: just send an e-mail to rfc-editor.
  Side effect: RFC Editor is unlikely to be able to cope with
  significant errata volumes, even if we are lucky enough to get them.

3) review

> Entries are reviewed by the authors, which seems to work
> satisfactorily.

  According to submission instructions[1], author or IESG review is
  required _before_ technical errata is submitted (i.e., before there
  is an errata entry). Thus, at least from a procedural point of view,
  the RFC Editor does not see entries rejected or ignored by authors
  and, hence, I am not sure you can claim satisfactorily performance
  of such reviews unless you know that most folks ignore your
  instructions and submit unreviewed errors that RFC Editor then
  forwards to authors for a review.

  Contacting authors (not to mention IESG!) directly creates an entry
  barrier that most (IMO) folks who spot an error will not overcome,
  for time, cultural, and technical reasons. I can create a more
  detailed list, but I am sure you know what those reasons are.

  AFAIK, there is no good documented mechanism to review/appeal an
  errata entry once it is published (not to mention rejected).

4) publication and search

  AFAIK, there is no good mechanism to track the status of an errata.
  My guess is that most submitted entries eventually show up on the
  page and some might be removed later. I do not know if RFC authors
  and WG are informed that there is new errata, and I am not sure
  there is a way to track what errors were addressed in later RFCs.

The above is not an exhaustive list, of course. For example, I have
not mentioned errata pages maintained by folks other than RFC Editor
but linked from RFC Editor errata page. Those pages are governed by
their own set of rules.

> It is not obvious that advertisement is useful/necessary. Their
> existence is now shown on the search engine for new readers.

Which, I bet, the vast majority of RFC readers do not use.

> If you had previously used some document for implementation, you
> would presumably have noticed any substantive typographic error
> yourself, and telling you later that someone else has noticed it and
> taken the trouble to report it seems gratuitous.

Typographical errors do not bother my much. There are more than enough
technical errors in core protocols such as HTTP that create real
problems. I bet that thousands of hours are wasted globally on [not]
fixing problems already known as errata to some. I thought it would be
obvious, but apparently I was wrong!

Alex.

[1] http://www.rfc-editor.org/errata.html