- From: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Date: Fri, 17 Jun 2022 06:20:01 +0000
- To: Mark Nottingham <mnot@mnot.net>
- cc: HTTP Working Group <ietf-http-wg@w3.org>
-------- Mark Nottingham writes: > If dates are their own distinct data type in SF, the underlying data > model would still be an integer or decimal, but it would self-identify > as a date, so that tooling could identify it as such automatically. If we did, we would just progress to the next level where the question becomes: What about time-intervals like "age", should they also be so tagged so they can be presented as "45m32s" ? We did not make SF context-free, you need to know what you are looking for to parse it correctly, so needing to know that some sf-decimals are timestamps is totally in character. (If we settle on sf-decimal to support fractional seconds, the precense of the period and the magnitude of the integer part will be a very solid heuristic for opportunistic pretty-printing.) Of course if almost nobody actually uses fractions, the extra ".0" is just wasted overhead, 11 bits in HPACK, but even then, a sf-decimal is still shorter and compresses considerably better than ISO-8601 timestamps would do. Therefore I prefer sf-decimal, but I can live with sf-integer. Introducing ISO 8601 as yet another textual timestamp format, offering no upside over the three already codified textual timestamp formats, would be not only pointless, it would be a regression, for instance in terms of HPACK efficiency. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Received on Friday, 17 June 2022 06:20:10 UTC