[AFS3-std] review comments for draft-deason-afs3-type-time-03

[Benjamin Kaduk]

I read this draft, and have no objection to the technical protocol 
described therein.

However, the document itself has a couple of omissions/errors.  Section 
3.3 does not explicitly mention that the resolution field is in increments 
of 100 ns (the same as the other types/fields), though this is mentioned a 
little later in the document.  Also, the text in Section 7 for converting 
between FILETIME and AFSAbsTime64 (and vice versa) reverses the sense of 
addition and subtraction necessary for effecting the different 
conversions.  (The sample code in Section 7.2 appears to be correct.)

Also, since it was written, we approved rxgk, which has a 64-bit time 
representation, rendering the claim in the introduction of "All extant" 
representations no longer accurate.


Now, on to editorial comments.  I have a couple of general comments, and a 
pile of tiny nits which I will send to Andrew unicast.

Throughout the document, there are a lot of references to representations 
with "more" or "higher" or "less" granularity than others, or "at least" 
some granularity.  This is potentially confusing, since higher (finer) 
granularity would actually mean a smaller "granular difference" (smallest 
representable difference).  Most of the document does use the "finer 
granularity" and "coarser granularity" that I would prefer, but some 
places do not.

In Section 5, I believe the "for" in "accommodate for" is not necessary.
There is also a "make take" when just "take" is intended, I believe.

In Section 6, I would s/worthy to note/worth noting/ .

-Ben