Re: Use case: tiger map/census data: have it your way

On Mar 18, 2004, at 17:33, ext Seaborne, Andy wrote:

> Dan:
>> But I agree with Patrick: a good way to get a sense of it is to tell
>> each other stories about it and see if those stories line up.
>
> Expressability in technical terms shouldn't be a gating criteria for a 
> use
> case.  If "expressed in terms of" is not restrictive, then fine.  But 
> if it
> is restrictive, if use cases are not considered, or not considered as
> significant, if they aren't expressed in technical terms then I think 
> we
> loose out.

I don't mean it to be restrictive, only to reflect our expectations
of what the DAWG recommendation might address.

It's a bit like the blind men and the elephant. Each is talking about
their personal experience (use cases) in terms of what they all think
is a shared point of reference (the DAWG recommendation). At some
point, after a good bit of confusion, they ultimately reach a general
concensus about what that thing is.

So, my use cases may reflect a possible realization of the DAWG
rec that your use cases may reflect a slightly different realization,
and through the process of understanding each others use cases, and
how those use cases are presumably facilitated by the DAWG 
recommendation,
we are able to come to a concensus about some intersection of 
realizations,
and then get to work on making an actual rec.

So, just because one of my use cases may suggest that the rec would
govern a particular characteristic of a conformant application, does
not mean that you have to agree and make your use cases reflect the
same coverage.

Eh?

Patrick

>
> 	Andy
>
> ----Original Message----
> From: Dan Connolly [mailto:connolly@w3.org]
> Sent: 18 March 2004 15:13
> To: Seaborne, Andy
> Cc: Patrick Stickler; RDF Data Access Working Group
> Subject: RE: Use case: tiger map/census data: have it your way
>
>> On Thu, 2004-03-18 at 09:03, Seaborne, Andy wrote:
>> [...]
>>> Patrick:
>>>> I think it's useful if our use cases are expressed in terms of
>>>> the anticipated DAWG recommendation,
>>>
>>> While I have sympathy with having use cases be framed in terms of the
>>> future rec (i.e. diving into tehtechnical), I am also aware that, as
>>> a group, we do not have a sense of what that recommendation is.
>>
>> But I agree with Patrick: a good way to get a sense of it is to tell
>> each other stories about it and see if those stories line up.
>>
>> I think what he's suggesting echoes another one of the qualities
>> of the ideal use case that I gave earlier:
>>
>> "
>> The ideal use cases will
>>
>>   -- clarify one or more requitements
>> "
>>
>> My hasty use case description fell short of ideal on that count.
>>
>>>   I like the qualities Dan
>>> provided as they are outwardedly focused, not technology focused.
>>> Getting engagement with the wider audience means talking about the
>>> value provided and less about the "how".
>>
>> Yes, the outward focus is critical. But the "how" needs to be in
>> there somewhere, eventually.
>>
>>>
>>> So I see use cases serving as input to a refinement step.  Let's not
>>> jump too early and only make use cases a technical description.  The
>>> need for a technical feature needs to be backed with an illustrative
>>> use otherwise it is a requirement with unclear value.
>>>
>>> If we find that many use cases are covered by one technical aspect,
>>> then that's good.  We will come out of the first phase with concise
>>> requirments that cover a range application/user needs.
>>>
>>> 	Andy
>
>

--

Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com

Received on Friday, 19 March 2004 04:35:08 UTC