Re: "draft-wilde-sms-uri-19" available

On Mon, Aug 24, 2009 at 7:33 PM, Erik Wilde<dret@berkeley.edu> wrote:
> hello dan.
>
> Dan Brickley wrote:
>>
>> Looking at the abstract and examples, the scope isn't very clear to me.
>> http://dret.net/netdret/docs/draft-wilde-sms-uri-19.html#uri-examples
>>  "for specifying one or more recipients for an SMS messages"
>> - grammar seems funny; "recipients for an SMS message" or "recipients
>> for SMS messages"
>
> i don't see where you copied that text, it seems to me there are two
> locations saying "one or more recipients for an SMS message", but i could
> not find the text you pasted here. maybe you looked at an older version?

I can't find it anywhere, and therefore must blush and admit I must
have introduced the error myself somehow while copy/pasting the text.
Apologies!

>> It might help if you list the classes of thing that this scheme
>> provides identifiers for? ie. an sms: URI could be an identifier for
>>  - a single sms-capable account
>>  - a list of sms-capable accounts
>>  - one or more accounts plus a draft message
>
> the difficulty may be that the intent of the scheme is to allow the sending
> of one message logically speaking, but since SMS does not have the
> capabilities to send one message to more than one recipient, on the SMS
> level this maps to a number of messages being sent, one to each individual
> recipient. since there are individual messages, they could be different, but
> the intent is that if multiple messages are being sent, they all have the
> same content. this is stated in section 2.3:
>
> '5. If the URI consists of a comma-separated list of recipients (i.e.,
> contains multiple <sms-recipient> parts), all of them are processed in this
> manner. Exactly the same message SHOULD be sent to all of the listed
> recipients, which means that the message resulting from the message
> composition step for the first recipient is used unaltered for all other
> recipients as well.'
>
> it seems to me that this is sufficiently clear, and the syntax as well as
> the examples make it clear (it seems to me) that there can be more than one
> recipient, and that an (optional) body can be supplied as the initial
> message contents. do you think the whole draft is sort of unclear at that,
> or just the abstract? the body part is not mentioned in the abstract because
> it is optional, and because more often that not, it will not be used.

I went in mainly via the abstract and examples, like 90% of your
readers will probably do. My main agenda when reading a new URI scheme
spec is to ask "ok, what kinds of thing does this spec let me
*identify*". So I tend to skip past the "verby" parts, ie. the actions
made possible. My understanding of the work of a URI scheme definition
(rather than a protocol that uses it) is that the thing that's being
standardised is a set of identifiers. So I am not completely sure how
action-related SHOULDs fit in. Did you consider making a separate
protocol for that? Would it make sense to separate out those aspects?

cheers,

Dan

Received on Monday, 24 August 2009 18:37:39 UTC