[rfc-i] Author names in RFC references

Dave Crocker dhc at dcrocker.net
Mon Mar 24 13:33:04 PDT 2014


On 3/24/2014 8:45 AM, Andrew Sullivan wrote:
> On Mon, Mar 24, 2014 at 08:41:04AM -0700, Dave Crocker wrote:
>>
>> 2.  What actual problems are caused by the current scheme?
>
> It's completely unique to RFCs,

No it is not. (As Elwyn D. has also noted.)

My own, stray example:  Trust Management, 2005, published by Springer. 
This is a compendium of conference research a papers.  So if the 
convention caused problems, it would have been a problem for an entire 
suite of authors.  (I'm making the large assumption that Springer is a 
credible publisher of technical material...)


>      so if you are in the habit of writing
> anything else anywhere you get it wrong all the time.

So it's really a miracle that anyone ever gets an RFC published with the 
names done properly?  And no one ever has to adjust to the 
idiosynchracies of each publisher's conventions?

My original question was not meant to ask whether anyone ever has a
problem with the current scheme, but whether there is a /pattern/ of
problems that have been demonstrated, warranting changing a 40+ year
convention.  If there is, where is it documented?


On 3/24/2014 10:38 AM, Paul Hoffman wrote:
 > On Mar 24, 2014, at 8:41 AM, Dave Crocker <dhc at dcrocker.net> wrote:
 >
>> On 3/24/2014 8:38 AM, Paul Hoffman wrote:
>> 1. It makes the list shorter
>
> It's 2014. We don't need to be saving bytes on the wire in our RFCs.

I didn't say anything about bandwidth issues.[*]

Even in 2014, humans prefer shorter lists and conventions that make them 
easier to parse.  Having the first/middle always be initials and the 
family name always be the full string makes for much easier /human/ 
parsing.  And it makes the total list shorts.


>> 2. What actual problems are caused by the current scheme?
>
> - It doesn't work well for authors whose names don't have clear
 > initials, particularly those in Asian countries who prefer to put
 > their family name first

So the proposal is to have authors using whatever ordering of names they 
prefer, relying on the reader to guess which is family name?  This is a 
far bigger change and I don't recall seeing it stated and agreed to, 
explicitly.


> - It makes reading our references more difficult for people who
> don't  understand the format

That borders on tautology. It certainly is universally true, for all 
publishers.

In any event, the question is:  where are the reports of this causing 
/actual/ problems, over the last 40+ years?


> - It is error-prone in a format that is supposed to be canonical

I don't understand what that means.


> On 3/24/2014 10:45 AM, Ted Lemon wrote:
> I find the authors' initials thing really weird and inconvenient—I absolutely_never_  care what the authors' initials are, and I_do_  care what their first names are

Whereas I find /that/ preference really weird...

However it nicely demonstrates the problem with using ourselves 
individually as if we were a representative sample.  While our 
individual anecdotes can be useful to highlight issues to consider, a 
sample of 1 is never very good at representing a much larger and more 
diverse population.

If we were to take this one example as essential, it means that author 
lists really should be changed to Ted L., Dave, C., Paul H., etc... 
That would be almost fun, if the IETF were still a tiny club of 40 or so 
folk.  Not so much for a global standards organization read by many more 
thousands than those who attend the meetings...
d/





[*] But since bandwidth has been mentioned, this discussion ought to 
take note of the discussion about revising the IETF web site that is 
noting how much of the world accesses the Internet only on cell phones 
with very limited bandswidth.  In general, we are treating ourselves as 
representing the total market, far more than is appropriate.  While it's 
fine to pay attention to our own use, with our own excellent platforms 
and bandwidth, ignoring those less fortunate is extremely UNfortunate.

-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net


More information about the rfc-interest mailing list