Re: [xml2rfc] RFC Bibtex format doi numbering incorrect

[Henrik Levkowetz <[e-mail]>]

Hi Matthew,

On 2017-05-30 01:27, Matthew Kerwin wrote:
> On 30 May 2017 at 08:58, Henrik Levkowetz <[e-mail]> wrote:
> 
>> Hi John,
>>
>> On 2017-05-30 00:26, John R Levine wrote:
>> >> Whatever was used, once you wrote the code, wouldn't it make sense to
>> >> publish where/how lookups should happen?  Has this been done?
>> >
>> > Of course.  It's in RFC 7669.
>> >
>> >> And no matter how opaque the DOI series identifiers are in general,
>> there's
>> >> still nothing that prevents the RFC Editor from _having_ a defined
>> mapping
>> >> from RFC numbers to DOI and vice versa, if a decision is made that this
>> >> would be helpful, right?
>> >
>> > No.  I really, really, wish that people would stop trying to relitigate
>> > this settled fact.  We had the exact same argument before the IAB
>> > published RFC 7669 which says:
>>
>> So you are saying that even if the RFC-Editor wished to define and publish
>> an algorithmic mapping between RFC numbers and DOI identifiers, they would
>> not be permitted to do so?
>>
>>
> This whole discussion is rather absurd, and I'm a bit disappointed to see
> it happening. (Not you specifically, Henrik, I'm just replying to the last
> message in the thread.)
> 
> The original bug report is legitimate, and should be corrected.  I.e. the
> tool that generated this document:
> https://datatracker.ietf.org/doc/rfc826/bibtex/ put the wrong value in the
> 'doi' field.

Agreed.

My question above was prompted by what seemed to be increasingly bombastic
declarations, which didn't seem to lead towards actually resolving issues.

> Presumably, that tool attempted to guess the value, rather
> than fetching it from a reliable source (contrary to RFC 7669, questions of
> which was published first notwithstanding), and guessed wrong.  Speaking as
> a general programmery type person, but not someone familiar with the tools
> and resources in question, I would assume that the tool should be able to
> retrieve the value for 'doi' the same way it was able to retrieve the value
> for 'title', from the same or a similar source, and if it were changed to
> do so the bug would be fixed.  Please correct me if that assumption is
> wrong.

I think that's right.  The code which generates the references on
tools.ietf.org is old TCL code which I very very happily leave to
Tony Hansen to poke at when necessary.  I'll have a look at the
datatracker code to see what happens there.

> Now, because I enjoy a good rabbit hole as much as the next person: Henrik,
> the RFC Editor can define and publish such an algorithmic mapping, of
> course.  If the goal is just to document the way in which they generate
> DOIs, it's potentially interesting, but not really useful to anyone outside
> the editors.  If, by publishing it, the RFC Editor wishes to make it
> possible for consumers to correctly "guess" a DOI, then the document would
> have to update RFC 7669 (with all the process and consensus that that would
> involve.)

Which is what I'd expect, and a very sensible response.

Thank you :-)

	Henrik