[Modeling] FIPA Modeling Area document format

Radovan Cervenka rce@whitestein.com
Mon, 24 Feb 2003 17:03:56 +0100


Dear all,

> Here are some comments I thought when reading the document:
>
> * In Front matter, I think about adding a field keyword. Usually, they are
helpful to
> have a first glance of the content of the document if they are accurate
and meaningful
> * Even if it is more or less the same, I prefer version to date of the
document, you can
> have several versions for the same day
>
> Part II: in class summary, I think about adding semantics, formal one this
time, I know
> it is a painful job but we need that
> I don't understand mapping to UML Base classes
>
> Appendix: maybe we can add a part on changes to help designers understand
the
> modification of the specs between two versions of the document.


basically I agree with improvements of Marc-Philippe, even if I'm little bit
skeptic about formal defining of semantics. "Mapping to UML base classes"
you pointed out is not about semantics of newly defined elements. It is
mostly used to express syntactic, not semantic properties/constraints.
Semantics of UML profiles is usually expressed in a natural language. This
"mapping" is a common mechanism how to define UML profiles, and most of
"UML-oriented" software engineers understand it. My skepticism about formal
methods results from the fact that they are not used in practice (in the
software development) too much, they are not used in OMG standards, most
people do not understand them, etc. But if we will see them useful, we can
put them into FIPA specification. Marc-Philippe, I'm looking forward to your
examples.

> > * Goal & soft goals (Radovan Cervenka)
>
> If it is strongly related to Tropos, that's not really something I
consider but I can
> give feedback based on my own work

Yes, Tropos (GRL, i*) was considered to be one (important) source. Can you
please give me some ideas based on your work? Thank you.

One remark regarding to the Diagrams section of the document format:
sometimes it is difficult (and can be also misleading) to prescribe diagram
types for new UML profile. UML diagram types are different than
profile-specific (domain-related) diagram types. One UML diagram type can be
used for several profile-specific diagrams, e.g. class diagram can be used
for acquaintance graph, goal decomposition diagram, service specification
diagram, etc. Furthermore, modeling elements of one type can appear on
several profile-specific diagram types in different context (e.g. showing
relationships to another elements). Sometimes, specific conditions of
methodology, implementation environment, technologies used, etc. determine
specific/customized using of existing modeling elements and also creating of
specific diagram types may happen. Therefore I have two recommendations for
authors: 1. please refer to the profile-specific diagram types, not UML
diagram types, and 2. put into the Diagrams section a note (if you see it as
useful), that mentioned diagrams are just usual cases, and other diagram
types may be used if required by specific conditions of the
project/product/technologies/...

Remark II: if we agree upon the document structure, can somebody of us
(Jim?) define also the document template for the FIPA Modeling Area
Documents? I think some common FIPA document template can be re-used...

With kind regards,

Rado1.
--
Radovan Cervenka | rce@whitestein.com
Whitestein Technologies | www.whitestein.com
Panenska 28 | SK-81103 Bratislava | Slovak Republic
Tel +421(2)5443-5502 | Fax +421(2)5443-5512
--
If you are not the intended recipient of this email,
you are not authorized to make any use of it;
please delete it and notify us by return email.
Thank you.