RE: SUO: Getting Ahead - Procedures
Dear Pierre,
Thank you for the comments. Hopefully the will lead to an improved
document.
See responses below.
Matthew West
Principal Consultant
Shell Information Technology International Limited
Shell Centre, London SE1 7NA, United Kingdom
Tel: +44 20 7934 4490 Other Tel: +44 7796 336538
Email: matthew.west@shell.com
Internet: http://www.shell.com
> -----Original Message-----
> From: Pierre Grenon [mailto:pierre.grenon@ifomis.uni-leipzig.de]
> Sent: 06 September 2003 17:00
> To: West, Matthew R SITI-ITPSIE
> Cc: Standard-Upper-Ontology (E-mail)
> Subject: Re: SUO: Getting Ahead - Procedures
>
>
> Matthew,
>
> Well, for all it's worth I'm thankful for these efforts. My
> French part is a
> little put off by the formality of the procedure.
MW: But at least you would know where you are.
> What would
> really be useful -
> and what is missing, but you probably can't be blamed - are
> guidelines for
> _content_development_.
MW: I agree. Are you offering to develop a deliverable?
> Other than that, I guess the procedure you offer is standard
> ISO community
> procedure.
MW: Actually it is highly simplified, but is based on standard
quality improvement principles.
> Assuming that we need thinking about that kind of
> procedural
> guidelines, a few comments which I consider as supporting the
> first two general
> ones. (I regard most of these comments as major issues.)
>
> Best,
> Pierre
>
> Genreal comment: It seems that the process you describe here
> is Starter and not
> Standard development. (see comment 9. below)
MW: Well we differ on what is a starter document. I would have
said that Draft Material was my idea of starter document, and
perhaps Draft Standard would be your idea. But if you are
saying that what I am suggesting as a standard would only
be starter material for you, I have to ask what standard you
set for something that would satisfy you as a standard!!
MW: Perhaps you have in mind that the only standard is the
final ONE SUO standard? This is not in practice how standards
are developed. Almost all but the most trivial standards are
developed as a series of parts. Each part is developed as a
standard, but of course there may be dependencies and
relationships between them.
MW: As an extreme example ISO 10303, developed in ISO TC 184/SC4
has several hundred parts, some of which run to several
thousand pages ...
>
> Genreal comment: What is the status of the group working on
> the proposal?
MW: Good question. I guess we could have an SUO recognised
Project Team. However, whilst in ISO work groups and teams
are recognised together with their leadership, membership
is quite informal in practice (you are a member by
participating). I would nto be inclined to propose something
more formal than ISO ...
> How
> is membership decided?
MW: Those with resources to contribute volunteer to do work.
The project leader organises the resouces and distributes
tasks - obviously in consultation with those volunteering.
(You can't tell people what to do unless you are paying them).
> at the discretion of this all powerful
> project manager?
MW: The Project Manager has more responsibility than power.
He obviously can't make anyone do anything.
> Why is there a project manager and no project (actually not
> until 2.3) nor
> project members?
MW: The Work Programme is the Project. I guess I shouls have
called it a Programme Manager.
>
>
> 1. Section 2.1
> I still think that this notion of a starting item is too weak
> (objective and
> outline).
MW: We clearly disagree, and I think this is at the principle
level rather than detail. The Group will have to decide this.
>
> 3. Section 2.1, first paragraph, last sentence.
> I'm leery about including the nomination of a project manager
> in a motion for a
> starter. Project manager should be a role. It should be
> optional. Even if the
> proposal mention the need for a project manager, the mention
> of the name should
> be optional only. Project editors, if any, should be
> removable as a matter of
> principle.
MW: Unless you have a Project Manager, I don't think you have
a project. Project Managers or editors can be changed at any
time.
>
> 4. Section 2.1, 3rd p.
> A compulsory role is that of an editor (removable as well). I
> have no intuition
> as to whether nominal mention should be compulsory.
MW: I have been trying to be minimal, clearly each deliverable
will need an editor, but finding one from volunteer resources
(or doing it himself) is the Project Managers job. I'm not convinced
we need to approve editors. It is the results that we need to
approve, not who edited them. I would want to minimise beaurocracy.
>
> 5. Section 2.1
> Would any of these roles have to be filled by voting members
> of the SUO only?
MW: Good question. I would say desirable but not essential
(given that there is a time constraint on becoming a voting
member).
>
> 6. Section 2.2, 2nd sentence. "This may be circulated for
> comment at the
> discretion of the Project Manager."
>
> What is that? Is a project manager a dictator or something?
MW: In a more developed version of this we would have the
project leader having the responsibility to ensure that the
document met certain quality requirements before wasting the
time of the Group on a review or vote. It should ideally be
someone independent of document production. In ISO there can
be 3 or 4 levels of these Quality checks.
> Remove or change to
> 'This may be circulated upon the discretion of the authors.'
MW: Now the authors are dictators!
MW: Firstly, you do not have authors, but editors. In the
end the whole group are authors of the document, so if you want
authors to decide, you would have to have a vote of the Group
to decide whether to circulate the document. I think a project
arrangement is more appropriate. Especially given that this is
about checking quality before distribution for comment.
MW: I will try to make the responsibilities clearer.
>
> 7. Section 2.2, 3rd sentence.
> No formal vote is requested where? In the SUO WG or in the ad
> hoc subgroup?
MW: I always mean in the SUO WG when I say formal vote. I expect
communication within a project team to be very informal. I would
expect an internal review by a project team before asking for
comments from the whole group. You only waste peoples tiem and
patience with poor quality deliverables.
>
> 8. How issues can be raised and sent to the SUO list (Lord
> Almighty have
> mercy!) if the Draft Material is not public? By whom will
> these issues be
> raised? The project manager?
MW: You may note that the material is to be circulated for comment.
The document should make clear that this is to the SUO list. I will
fix that.
>
> 9. Section 2.3 on the vote on the draft standard proposal.
> That is completely
> abstract. I assume the vote is taken in the SUO WG group, not
> in the project.
MW: Indeed.
> But what is voted on? How can this achieve any level of
> completion by itself if
> this is intended to be incoprorated in a modular framework.
MW: Well the whole point of a modular framework is that there
are parts that can be developed independently. This is normal
for standards development as I pointed out above.
> Why, you are
> proposing to develop many standards?
MW: This is the normal approach. A single monolithic standard
is impractical to develop.
> It seems that the process you describe here is Starter and
> not Standard
> development.
MW: That suggests you equate starter with what I would call
a part. If that is the case, none of the documents we have
had presented - even the IFF stuff comes anywhere near the
quality I would expect.
>
> 10. Editorial: Project Leader reads Project Manager.
MW: Thanks.
>
> 11. Remove mention of Project Manager.
> This decision should be the outcome of a vote among the
> project members
> (modality to be elucidated). There should be an arbitrating
> instance in case of
> indecision or major disagreement to which any party should be
> able to appeal.
> SUO WG Chair? SUO WG itself?
MW: I disagree. If there is an action some person in some role
has to be responsible or it doesn't get done. Basic rule of
management. If you would prefer this to be the SUO Chair, I
have no objection, but I think someone close enough to the
work to appreciate the issues is better.
>
> 12. Triffle: Pro Forma should include an entry 'Project name'
> or some other
> identifier.
MW: Good point.
> Also, an entry, 'issue number'. You can conflate
> and have one entry
> formatted as PROJECT_NAME/ISSUE_NUMBER.
MW: I was expecting the Project Leader or whoever was collating
the issues to do this, rather than the issue raiser. But I will
add the field.
>
> > Dear Colleagues,
> >
> > Please find attached a PDF file of a document entitled:
> >
> > Standard Upper Ontology Development Guidelines Version 1.0 Draft 1
> >
> > This can be considered as a preview of a deliverable of the first
> > task in the proposed Work Programme.
> >
> > If it is applied recursively, it is draft material for review.
> >
> >
> > Matthew West
> > Principal Consultant
> > Shell Information Technology International Limited
> > Shell Centre, London SE1 7NA, United Kingdom
> >
> > Tel: +44 20 7934 4490 Other Tel: +44 7796 336538
> > Email: matthew.west@shell.com
> > Internet: http://www.shell.com
> >
>
>