Re: [PyWrapper-devel] WG: tapir: capabilities

[<m.d...@BG...>]

Renato,
I agree with your remarks. And I think we should stress the =
dynamic/static meaning more than the standard/custom one.

If everyone agrees I would opt for the dynamic/static terminology then =
and modify the capabilities section as renato has outlined?. Especially =
as we have been refering to dynamic models ever since.=20

-- Markus
 =20

> -----Urspr=FCngliche Nachricht-----
> Von: Copp, Charles=20
> Gesendet: Montag, 17. Juli 2006 15:29
> An: Renato De Giovanni
> Cc: D=F6ring, Markus; Javier privat
> Betreff: Re: WG: tapir: capabilities
>=20
> I think - standard/custom or static/dynamic is a choice of=20
> terms that either would suit depending on how you want to=20
> emphasise what they do.
>=20
> standard - meaning it comes provided for you as part of the=20
> service you are contacting and custom means that it can be=20
> changed and the user declares it.  static/dynamic emphasise=20
> what the service is doing in processing terms.
>=20
> I think your Renato's comment on <structure> makes sense.
>=20
> Charles
>=20
>=20
>=20

Hi,

I think I'm fine with any of the options. It could be "standard" if you =
all liked. The only detail is that a standardOutputModel doesn't =
necessarily come from an external library of standard models. It can =
also be something specifically tailored by the data provider for some =
reason.

By the way, I'm looking at the schema now and it seems to be possible to =
have a "customOutputModels" element with attribute "accepted=3Dtrue"=20
but with no schema structure capabilities specified - which is =
inconsistent.

I think that the "customOutputModels" section is more related to the =
ability to dynamically process any output model specified in requests. =
In contrast, the "predefinedOutputModels" is just a list of "static" =
output models that are understood by the provider, regardless of having =
the response schema structure processed dynamically or not.

So another possibility would be:

<staticOutputModels>
 ...list of outputModels...
</staticOutputModels>
<dynamicOutputModels>
  <structure>...</structure>
<dynamicOutputModels>

In this case <structure> could become mandatory (minOccurs=3D1) and we =
could remove the attribute "accepted".=20
Just another idea...
--
Renato


On 17 Jul 2006 at 13:20, "D=F6ring, Markus" wrote:

> hi,
> a name for "predefined" model.
> charles suggests "standard" as oppoased to custom.
> I think that hits the nail on the head.
>=20
> other suggestions so far:
> - standard
> - shared
> - preconfigured
> - known
> - available
>=20
>=20
> lets pick!
>=20
>=20
> -- Markus
> =20