[cddlm] phone conference notes, may 25 2005
Ayla Debora Dantas de Souza - Projeto Ourgrid
ayla at dsc.ufcg.edu.br
Fri May 27 07:58:11 CDT 2005
Hi,
sorry, maybe it is too late to give some comments, but maybe they could
be useful and could help to clarify some points in the document, even
for other versions. I would like to say that these comments are based on
the view point of someone new in the group.
* I think the text should more clearly define the role of the
deployment API, or may be more homogeneous. Sometimes it is said to
deploy "jobs" (as in the introduction) and sometimes
"applications"(section 2.1). In section 2.1( wrongly numbered as 1.1.1),
it's role is just to deploy applications to one or more target
computers. In Introduction, it is said to enable programs to submit jobs
into the system for deployment, terminating existing jobs, and probing
the state of the system. The term "such tasks" that is used in the
second paragraph of section 2.1 is too general, because in the previous
paragraph, these tasks are not clear. In the same paragraph, the term
"this service" refers to the service provided by the API? Maybe this
sentence should be rewritten.
* Maybe it would be better to remove the title "1.1.1
ConfigurationDescription Language", or it should be included a title for
the 3 points described in the section referring to each document.
* Section 2.2 - will the document be updated to refer to new drafts
of some documents , such as in the WSRF-RP and WSRF-RL or
implementations should consider old drafts? Maybe there are not so many
differences.
* Last paragraph of section 2.2: "the referenced to...": I think the
verb is missing.
* In section 3.1, it would be good to put a reference to the document
that should indicate what you consider an OGSA-compliant Grid Fabric. Is
it based on the first version or the
"draft-ggf-ogsa-wsrf-basic-profile.doc" document?
* The term "cluster" in section 3.1 is used to refer to any set of
machines or to the cluster definition used for High Performance
Computing? If you refer to the first case, wouldn't it be better to
remove the term cluster?
* Second paragraph of section 4.1: "...deploy to one or more..."
instead of "..deploy to ore more...." At the end of this paragraph, it
is said that the process for obtaining a Portal EPR is not in the scope
of the document. Does this mean that the EPR can obtained in any way or
that there is another document specifying that?
* Section 4.2, 5th paragraph, first sentence: The configuration of
the component properties may be made before an "initialize" call on a
component or configurations may be done just during initialization?
* Section 4.4 - Subsection numbers are wrong, as previously noted. In
section 4.4.1 (that is now 3.1.1), I think a ".." is missing between "0"
and "9" when specifying the characters range.
* Section 3.1.2 that should be 4.4.2: replace "provide" by "provided"
* Section 4.5: replace "exports" by "export".
Best regards and thanks in advance,
Ayla
Steve Loughran wrote:
> Summary
>
> Discussion of the deployment API. The doc is going to be updated with
> feedback from the group and
> Bryan Murray; a new version will come out friday afternoon. This will
> be the last call document;
> Steve is away the following week.
>
> For a friday release, all comments must be in by friday am 0900 UTC.
>
> There may or may not be a phone call next week.
>
> -------------------------------
>
> Dejan
>
> -structure.
> consolidate fault tolerance
> consolidate architecutre
>
> -section numbering broken (word problems)
>
> s2.1.1 should be broken into language and s.2.1.2 comp model.
>
> Cover materials
> -needs to put stuff in context with the other specifications. This
> should go on the front page.
> -relationship to other working groups. JSDL, ACS, DRAMAA
>
> security: should vs must in recommendations about filesystem and
> visibility
>
> References
> -add Jun as author of XML CDL
> -dejan has verified that references appear in doc; all looks good.
>
>
> stuart email comment:
> -put sequence diagram example at back.
>
> (dejan: how are (relatively high level) use cases realized?
> steve: sequence diagrams show how
>
> stuart email comment:
> -consistency with pictures vs schema fragments.
>
> dejan: Pictures are good - how about both?
>
> steve: its hard work keeping everything consistent. The XSD will go in
> an appendix.
>
>
> Jun
>
> -put XSD at end.
>
> -s5.3 use prefix api for typing in XML namespace.
>
> -will have lots of feedback at implementation time.
>
> steve: we can spend some time talking about testing @GGF.
>
> Kojo
> -feedback by friday morning UTC.
>
More information about the cddlm-wg
mailing list