2518bis-14 feedback

The new locking stuff is really nicely done; it covers everything I 
wanted 2518 to cover.  The only general comment is that the use of the 
phrases "write lock" and "types of lock" seems slightly confusing 
because the spec now defines no other type.

Some random nits in stuff I've read so far:

6.6: s/cleaned up/deleted - can infer that "cleaned up" means deleted 
but it's a bit vague; when I clean up my kitchen I aim to still have a 
kitchen afterwards ;)

7.5 para two has an uncapitalized "if header" in the third sentence.

7.7 para two; the MUST requirement is confusing; is it a requirement on 
client or server? The "use" of the request is something which is 
controlled by the client, but this is a requirement concerning the 
*interpretation* of the request, which is controlled by the server?

The 7.7 para one MUST requirement seems redundant too really.  If a 
client chooses to submit an identical LOCK request twice, then the model 
already specifies server behaviour.

8.7 is missing a trailing period on the last sentence.

10.4.1; bullet 2 sentence 1 s/that is has/that it has/.  The second 
sentence reads awkwardly, maybe replace it with something like "The 
semantics of the "submission" of a state token depend on its type (for 
lock tokens, etc)."

Received on Friday, 24 February 2006 15:54:47 UTC