Generalizations - isSubstitutable

[Paolo F Cantoni]

Generalizations have a custom property: isSubstitutable.

What does it mean?

TIA,
Paolo

[Paolo F Cantoni]

Is there any formal documentation on the Custom Property sets?

Paolo

[Eve]

According to the spec isSubstitutable

Indicates whether the specific classifier can be used wherever the general classifier can be used. If true, the execution traces of the specific classifier will be a superset of the execution traces of the general classifier.

Generally these are direct properties out coming from the spec.

[«Midnight»]

I looked at this too Paolo. At the time I had misunderstood what "custom" meant and was wondering why I couldn't use them for my own purposes.

You'll have to dig into the spec to see what these indicate. They can be useful in specific situations, but are generally fairly niche oriented.

There does not seem to be much in the way of an overriding pattern for them either.

[Paolo F Cantoni]

I looked at this too Paolo. At the time I had misunderstood what "custom" meant and was wondering why I couldn't use them for my own purposes.

You'll have to dig into the spec to see what these indicate. They can be useful in specific situations, but are generally fairly niche oriented.

There does not seem to be much in the way of an overriding pattern for them either.

Oh.. I see...

"Custom property" means UML 2 Attribute, not already handled on the "normal" EA unique user interface ..  

So to find out what they mean, I need to got to the UML 2 [size=13]Superstructure[/size] Specification...

Paolo

[«Midnight»]

Actually, I think the UML 2.0 specification treats (at least some) of these in a different manner than the rest of the stuff.

IMHO, to the extent that there is (or we can discern) any patterns in these, we should petition Sparx to handle these (sets of attributes that are in each pattern) in a consistent manner, and expose them to the modeller and the automation interface.

At present the custom properties are read-only to the API.

What say?

[KP]

So to find out what they mean, I need to got to the UML 2 [size=13]Superstructure[/size] Specification...

There's a more recent version of UML 2.0 Superstructure linked here.

[«Midnight»]

Simon,

Some time ago I (think I) saw a note from somebody at Sparx about UML 2.0 versions. It said (as I perhaps remember it) something like "we're using the ... version" with the ellipsis being the OMG document date; the one pointed to was earlier than that referenced elsewhere in the thread.

All that said, is the "newer" one you refer to now the Sparx standard? If so, will Sparx be (over time, given realities) updating user guide references etc. to refer to the new version?

David

[sargasso]

UML2.0 versions are not that "hard".  The following are from OMG's site and may explain ...

Adoption of the UML 2.0 Superstructure is complete - No further technical work is being done; in fact the Superstructure specification has been stable since it took its adopted form in October, 2004.

 The later versions consists mainly of errata and clarifications.  August 2005 is a "final" which is a clean copy of the "convenience" versions published between 2004-10 and 2005-07.
Having said that,

A note about UML 2.0:   OMG is currently upgrading all of UML to Version 2.0. It's a large specification, made up of four parts (Superstructure, Infrastructure, Object Constraint Language (OCL), and Diagram Interchange).  Until all of the parts of the UML 2.0 specification become formal specifications, you can refer to the following documents that comprise these parts of the UML 2.0 specification.

So expect changes to the remaining specs (which could cause re-issuance of the Super spec) until they are adopted at some time.

In fact the only "standard", as in ISO standard, is

OMG is pleased to announce that the UML specification (version 1.4.2, OMG document:  formal/05-04-01) is now an accepted ISO specification.  It is available from ISO as
ISO/IEC 19501.  

The major problem we have, and have had in the forums, is that references provided to the differing versions make it difficult for each other to find the relevant bit!  For example, refering to a super spec diagram on page n causes confusion, sometimes even referrring to diagram m-n is indeterminant.  The super structure is pretty stable though  and referring to "Section 7.3.7 " for example will always  :-/ get you to the UML 2.0 class structure section.

-- "The last grape in the bunch" : --  
It helps when referring to the spec, to give a clear indication of the version you are using. There have been a few "discussions" in these forums that have arisen because of the microscopic differences between versions at a particular sentence.   (Would I get involved in these? ... Never   )

bruce