Hello Carsten, others, (taking iana@iana.org out) On 2023-12-19 16:54, Carsten Bormann wrote:
Hi John,
On 2023-12-18, at 19:07, John C Klensin <john-ietf@jck.com> wrote:
Carsten,
(Top post... and I'm out of time for this) (1) My primary objective was to try to explain how we got here, in part because my involvement with the relevant decisions goes back further than Martin's tenure as charset reviewer and Ned is not here to defend himself.
Much appreciated! I was not aware that there actually was a discussion that rejected “ASCII” as a label for ASCII.
Although I haven't been part of it, I'm quite sure there was a discussion about avoiding "ASCII" as the recommended label; "US-ASCII" always was and is clearly labeled as preferred for MIME. On the other hand, I'm quite sure the discussion did NOT reject "ASCII" as a label for ASCII. Otherwise, it would not have been listed as an alias in the registry until 2013. The first instance I have found in the wayback machine is from December 1998, at http://web.archive.org/web/19981202100649/http://www.isi.edu/in-notes/iana/a....
My argument is that it is detrimental to call ASCII something else than ASCII in conversation, because there are no other ASCIIs (except for the historical ASCII 63 etc.). Imagine always saying “XML with angle brackets” when we mean XML...
That sounds fine, but communication is always context-dependent.
What labels we should use for interoperability is a separate matter.
Yes indeed.
I’m also intrigued by Martin’s research about how we arrived at the current state of the registry…
It was Steffen who found the difference. I just expanded on his "local copy" with something more generally accessible (the Web Archive).
If there really was a process error, that would sway me a bit to actually aligning the registry with the (feeble) reality of charset=ascii being in use.
I don't think there was a process error (which would to me mean that a process was not followed, or followed the wrong way). There may have been a processing error (to the extent that the conversion from the old text file in or around 2013 was done e.g. by a script) and/or a human clerical error (to the extent that the conversion was done by a human).
(But then errors that have been in place for a decade develop their own normative reality.)
My current understanding of reality would be that within a margin of error, the label "ASCII" was used about as rarely during the period it was listed in the registry (i.e. ~2013) as during the period it wasn't listed (2013~2023). If there's not much reality, it's difficult to speak of normative reality :-). Regards, Martin.