[Fiware-wpl] IMPORTANT: Approach for Architecture Specification Deliverable

[Juanjo Hierro]

Hi all,

  As you know, we have to start working on the Architecture Specifications, which is one of the deliverables that is due in month 9, that is, end of January.

  My intention is to adopt a pragmatic approach, trying to generate something that can be integrated as part of the GE Open Specifications that we have to produce by month 12.

  My vision is that all GE Open Specifications should start with a chapter where an overview of the envisioned Architecture for the GE is described.   While style would be narrative, the description should be enough concrete, i.e., formulated over actual data type, interface and operation names and elaborating on the base interaction scenarios involving the different entities exporting the defined interfaces.   Then, after that chapter, the detailed specification of all data types and interfaces with their operations, are provided (this including signature of operations and accurate description of expected behaviour linked to operations)

  Then, the Architecture Specification deliverable would be just the result of developing this overview chapter.

  But nothing better than an example, so let me use one taken from OMG's set of CORBA Services Specifications.  Along my many years involved in different standardization efforts, I have found that OMG CORBA Service specs are rather comprehensive and close to what programmers (our ultimate customers!) love to see.

  Please find enclosed the CORBA Event Service Specifications.   What I would then select as the Architecture description are the contents of the following sections:

 *   The whole chapter 1
 *   Section 2.2 and 2.4 (which some people may have argued should have been included in chapter 1 :-)

  Note that anyone who reads the mentioned sections would get a CLEAR picture of what is the Architecture of the service and what are going to be the entities and interfaces/operations that will be supported in a compliant implementation.   The conceptual and programming model would be also rather clear from a programmer's perspective.   Therefore, rather valuable information for an application developer perspective which is what really should matter to us.    A side benefit is that what will remain as pending regarding GE Open Specifications will be less, that is, the detailed specification of data types and interfaces/operations.

  One thing that we will have to define is the set of conventions that we should all follow whenever we need to draw any figure, like figures 1-x or 2-x.   For this, Thomas and me will come soon to you with a proposal in short time.

  Please take your time to analyze this carefully and formulate any question you may have.

  From now on, I assume that you will start planning activities in Sprint 2 and 3 dealing with development of these specifications within each of your chapters.   This may well take the form of Work Items in the tracker.

  Best regards,

-- Juanjo


________________________________
Este mensaje se dirige exclusivamente a su destinatario. Puede consultar nuestra política de envío y recepción de correo electrónico en el enlace situado más abajo.
This message is intended exclusively for its addressee. We only send and receive email on the basis of the terms set out at.
http://www.tid.es/ES/PAGINAS/disclaimer.aspx
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.fiware.org/private/fiware-wpl/attachments/20111130/9182f227/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Event Service Specification 04-10-02.pdf
Type: application/pdf
Size: 515254 bytes
Desc: not available
URL: <https://lists.fiware.org/private/fiware-wpl/attachments/20111130/9182f227/attachment.pdf>