Hi Jouni,

>>> * The intoduction mentions that the block transfwer can be used for
>>> random access. This is not separately described anywhere in the
>>> documents. Although the operation is more or less trivial I would
>>> appreciate an example or some text there because the Block2 overloads
>>> the NUM value 0 with special meaning (for the cases where the first ever
>>> accessed block is not the block number 0).
>>
>> The random access operation is not described in this draft, but it is
>> indeed quite obvious how one would make use of such a capability.
>> Its use would require some previous understanding between client and
>> server (possibly gained by the client simply trying out random access).
>> I'm not sure how NUM=0 is special in this case.
> 
> NUM=0 is used in some cases to indicate the desired block size in the
> initial request (as I read and understand from the draft.. if that is
> not the case then it needs to stated more clearly that it is not the
> case). I mean the text in 2.4 and 2.8 for example.

There is definitely a special case for NUM=0 in 2.8 (multicast), but
that is because, for discovery, the client will switch to unicast for
all other blocks.  It doesn't sound reasonable to "discover" a resource
while performing random access to it.

The text in 2.4 is more about the outcome of using Block2 options
properly; here we prefer the wording "first request"/"first response"
because indeed that might be for a block with NUM≠0.

(The special case for NUM≠0 in the last paragraph of 2.10 is about not
creating "holes".)

Grüße, Carsten

_______________________________________________
Gen-art mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/gen-art

Reply via email to