Re: [rfc-i] Fixing names in references to old RFCs

[Julian Reschke <[e-mail]>]

Am 18.02.2021 um 08:58 schrieb Carsten Bormann:
> On 2021-02-18, at 07:57, Julian Reschke <[e-mail]> wrote:
>>
>> The problem with *any* change to the references resources is that if you
>> re-generate older specs that use these resources you get a different
>> result. (We have a similar issue with other changes to <reference>, such
>> as changes to seriesInfo, or even the insertion of DOIs).
>
> Yes, we have that “problem” massively with unnumbered references to I-Ds.
> We actually like that problem so much that we have this mechanism in the first place.

Referencing unspecific ID versions only works well if you do not rely on
anything *specific* in it (such as a particular section).

> If something gets fixed in a shared reference resource, everybody benefits.
>
> It’s a feature, not a bug.

But it's bug if it breaks existing uses.

There are two questions here:

1) Should we "fix" references to contain something that actually is not
in the referenced document (such as a non-ASCII character in an author
name)?

2) Should we change the published reference files to use v3 features
(such as extensions to the content model and support for non-latin
characters in names and titles)?

For 2), a simple answer that would still give you the benefit of shared
resources is "opt-in" (here: just use a different URI for the v3 resources)

For 1), I can see good reasons on both sides...

Best regards, Julian
_______________________________________________
rfc-interest mailing list
[e-mail]
https://www.rfc-editor.org/mailman/listinfo/rfc-interest