[rfc-i] GitHub references
Martin J. Dürst
duerst at it.aoyama.ac.jp
Thu Mar 29 00:11:07 PDT 2018
On 2018/03/27 22:45, Adam Roach wrote:
> Really, we want to point to a specific version — that’s what a hash gets us. Lacking a specific unique identifier, the best we can hope for is time to the greatest precision available. Day, if you have it. Minutes, if available. Seconds, too. If the system provides sub-second resolution, include it. Anything that reduces that chances of the reference ambiguously referencing more than one version.
If we want to point to a precise version, then that's certainly the way
to go. But there are cases where one may NOT want a precise version.
Examples include stuff like "there's code over there that you might be
able to use" or "use the newest version because it may have less (known)
bugs than older ones", and so on.
At the W3C (long time ago), we went through this with respect to Unicode
(which was at that time a book). In some cases, we wanted to cite a
specific version. In other cases, we wanted to the newest version (with
all the fancy new characters, which didn't included emoji at that time)
to be usable.
More information about the rfc-interest