Re: [Digir-dev] Request for a DiGIR Teleconference

[Renato De G. <re...@cr...>]

On 18 Dec 2002 at 15:05, Hobern, Donald wrote:

> Please take the time to answer the following questions:
> 
> *   Would you be free to participate in a teleconference on Wednesday, 15 
>     January (same date for everyone) at 5am in Tsukuba, 7am in Canberra, 
>     9am in Christchurch, 12 midday in San Francisco, 2pm in Kansas, 6pm 
>     in Sao Paulo, 9pm in Paris/Berlin/Copenhagen, lasting for one or two 
>     hours?

Yes.

> *   Would you be prepared and able to travel to a 2 or 3 day meeting to 
>     work on the DiGIR protocol and associated software? I think that 
>     California might be a sensible venue if most of us were to attend, 
>     but alternative suggestions are welcome. 

Yes. The only constraint is travel funding.

> *   Do you expect to use the DiGIR protocol to exchange data within 
>     2003? 

Yes. We already have a working prototype for the speciesLink project
(initially covering 12 collections from Sao Paulo state, including 
ichthyology, botany, entomology and microbiology).

> *   What schemas do you expect to use to model requests and results? 

Latest Darwin Core version.

> *   Identify any critical dates for your projects. 

October 2003.

> *   Do you expect to develop any DiGIR software in 2003?

Yes.

> *   What components will you be developing (Provider, Portal 
>     (federation), Client)? 

Mainly other clients, but we'll also continue to contribute with 
the current DiGIR provider and portal code base.

> *   Will you be developing on the SourceForge code base or separately

We usually develop using our internal cvs repository, but we may 
happily switch to sourceforge if someone else demonstrates any
interest on working together in a particular project.
 
> *   Will your components be available for others to use? 

Yes, one of CRIA's missions is to disseminate technologies related to 
biodiversity information.

> *   Can you please comment on a proposal from the BioCASE team to make 
>     certain changes to the DiGIR protocol schema (see 
>     http://www.biocase.org/temp/protocol/BioCASEproposal.htm). The 
>     changes proposed are:
> *   Add a Capabilities request to query schemas supported by a Provider. 

This kind of information, if implemented, seems to fit better inside 
the metadata stuff.

> *   Replace substitution groups with XPath expressions. 

We don't think we should all lose part of the current DiGIR functionality
without being absolutely sure there's no other way to accomodate all
interests.

Doesn't a "true" xpath query already imply the response schema? So, by 
allowing different request/response schemas aren't we creating a situation
where potencial inconsistencies may arise? (e.g. using elements on a query
that are not available or even mappable in the response schema, and 
vice versa).

Regarding the BioCASE project, why not identify the main concepts inside
the ABCD schema that are not covered by Darwin Core, extend the former
to a "super darwin" (flat), define a custom record structure to be returned
using the "super darwin" concepts (somehow cloning the ABCD schema),
and maybe help making the necessary changes on code?
DiGIR software already supports nesting and repeating elements on
response record structures, but certainly it still needs more adjustments
to make all of this come true.

> *   Add a Transformation element to identify an XSL style sheet which the 
>     Provider should use to transform the output before returning it. 

It really looks like a client task, since one can already indicate a 
response schema during request. And clients will have to deal with xml,
anyway. (probably using xslt!)

> *   Replace the existing logical operators with AND, OR and NOT. 

Ok, we agree on this.

> *   Please suggest any agenda points you would like to have covered in a 
>     teleconference.

To reach a consensus on the future of the protocol...

Regards,
--
Renato De Giovanni
Mauro Munoz
Ricardo Scachetti Pereira

CRIA - Centro de Referencia em Informacao Ambiental
http://www.cria.org.br/

--
This message has been scanned for viruses and
dangerous content and is believed to be clean.