search

gbuehler / 4

0 stars 0 forks source link

Adjust definition of 'realization' to match that in UML 2. #130

Closed gbuehler closed 3 years ago

gbuehler commented 3 years ago

Justification: ISO 19103:2015 includes a definition for realization, at 4.29. The claimed source is [UML 2], which the bibliography implies is the 2004 Rumbaugh & Gouch UML book. This doesn't give a formal definition.

In comparison to the definition in OMG UML Superstructure 2, 19103 contains three extra words, added near the end.

I have not bought ISO 19505-2 to check whether these words are introduced there. Proposal: Delete "of the latter" at clause 4.29

Ripple this change to those standards which import the 19103 definition, such as 19139-1

gbuehler commented 3 years ago

It is unclear which normative reference is meant by "[UML 2]"; it would appear to be a reference via the bibliography to "Rumbaugh J., Booch G., Jacobsson I. Unified Modeling Language Reference Manual. Addison-Wesley, Second Edition, 2004". That book provides a good description of realization, but not a formal definition.

It would probably be better to cite definitions from ISO 19505-2, and adjust the "[UML 2]" to be an internal reference to the normative reference, not the bibliography.

In OMG UML Superstructure 2 realization is described at Clause 7.3.45 (and 2.4.1, where it is clause 7.3.46): "of the latter". The 'descriptions' are the same words as 19103 uses for the definition. Presumably ISO 19505-2 has the same definition (perhaps as a formal ISO definition); I haven't bought it to check.

I made a comment against 19139-1 DIS that these words are confusing. I had missed it when involved in 19103:2015.

My preference would be to align the definition with that in OMG UML. If ISO UML is different, then also raise this comment against 19505-2. Having found a copy of ISO 19505-2:2012, at https://www.omg.org/cgi-bin/doc?formal/12-05-07.pdf, I find that it does include the extra three words "of the latter" (in description at 7.3.46)

I wonder what they were thinking? I guess I need to find a way to raise it there. Strangely, I've just re-checked, and even the OMG documents contain these words. Not sure what happened; I'll attempt to close/delete this request!

gbuehler commented 3 years ago

It is unclear which normative reference is meant by "[UML 2]"; it would appear to be a reference via the bibliography to "Rumbaugh J., Booch G., Jacobsson I. Unified Modeling Language Reference Manual. Addison-Wesley, Second Edition, 2004". That book provides a good description of realization, but not a formal definition.

It would probably be better to cite definitions from ISO 19505-2, and adjust the "[UML 2]" to be an internal reference to the normative reference, not the bibliography.

In OMG UML Superstructure 2 realization is described at Clause 7.3.45 (and 2.4.1, where it is clause 7.3.46): "of the latter". The 'descriptions' are the same words as 19103 uses for the definition. Presumably ISO 19505-2 has the same definition (perhaps as a formal ISO definition); I haven't bought it to check.

I made a comment against 19139-1 DIS that these words are confusing. I had missed it when involved in 19103:2015.

My preference would be to align the definition with that in OMG UML. If ISO UML is different, then also raise this comment against 19505-2. Having found a copy of ISO 19505-2:2012, at https://www.omg.org/cgi-bin/doc?formal/12-05-07.pdf, I find that it does include the extra three words "of the latter" (in description at 7.3.46)

I wonder what they were thinking? I guess I need to find a way to raise it there. Strangely, I've just re-checked, and even the OMG documents contain these words. Not sure what happened; I'll attempt to close/delete this request!