Devin Kouts wrote:
>
> The approach I used to build a construct similar to the following:
>
> <SurveyTeam>
> <Member>
> <Person/>
> <Role/>
> </Member>
> </SurveyTeam>
>
> has been termed a "fat protocol" in some XML discussion groups.
Thanks for the discussion about fat protocols. I guess I would
have named the difference "deep" (all elements) vs. "shallow"
(lots of attributes), but fat vs. thin works.
It would seem that both <Person> in <Instrument> and <Person> in
<Member> in <Survey> are fat.
So, I'd ask what would the design goal be that lead to this particular
separation of <Person> and <Instrument>. I can think of only one,
so that I can list those 'extra folks' on the survey team. Role=Tourist
(to
say it in a thin manner, but <Role> would do as well. Or maybe
<Role>Instrument Trainee</Role>, but otherwise I'm fishing for
discussion of other interesting reasons for this particular breakout,
regardless, if the result is
<Person role="Tourist">Paul Hill</Person>
or
<Insturment>
...
<Person> Paul Hill </Person>
</Instrument>
I wasn't worried so much about the total depth and element vs.
atttribute as the reason for the parallel branching of
<SurveyTeam> vs. <Instrument> instead of simply (?) a deeper/fatter
<Instrument> structure.
cheers,
-Paul
This archive was generated by hypermail 2b30 : Mon Apr 02 2001 - 18:00:00 CEST