Obsolete UML within EA (from the new standard)

[alicecbrown]

I"m looking up 'thread' for a UML definition, and although it is used all through the standard, this is what I find last:
This is scary:UML Superstructure Specification, v2.1.1 695

Changes from previous UML
The following table lists predefined standard elements for UML 1.x that are now obsolete.
:oChanges from previous UML

Standard Element Name Applies to Base Element «access» Permission «appliedProfile» Package «association» AssociationEnd «copy» Flow «create» CallEvent «create» Usage «destroy» CallEvent «facade» Package «friend» Permission «invariant» Constraint «local» AssociationEnd «parameter» AssociationEnd «postcondition» Constraint «powertype» Class «precondition» Constraint «profile» Package «realize» Abstraction «requirement» Comment «self» AssociationEnd «signalflow» ObjectFlowState «stateInvariant» Constraint «stub» Package «table» Artifact «thread» Classifier «topLevel» Package

[thomaskilian]

I guess you're the first to make that comparison. You should submit a bug report to Sparx. I haven't tried (and can't do it right now), but the Aggregation connector has been also keft over as heritage (still there from the quick connector?)

[sargasso]

By way of interest, does anyone know if Sparx has an explicit UML compliance or compliance policy notice on the website - or elsewhere?

Now I do not see that the retention of UML 1 items in the tool is a bug.  I have several clients and many models that are UML 1.  Neither I, nor they, see any need to feel that these models are incorrect.  They still work and are still maintained as UML 1.

Also, as I have said before, UML isn't L.A.W. and AFAIC the support of both spec versions is a strength.


bruce

[«Midnight»]

And we should not forget that the ISO standard version of UML is based on 1.4.x.

[alicecbrown]

   so although the EA literature implies that it is UML Standard (updated) compliant, we should not believe that.

My further question, is whether the standard means that these terms are no longer constrained by their definitions, or are indeed 'obsolete', as stated.  The latter does not seem feasible, since 'thread' is a part of our software language, our way of communicating a certain behavior of the code.
That scary sentence preceding this table needs further amplification.    English is a horrendous language for communicating rigorous concepts. And that sentence needs a whole bunch of caveats!!!  

[«Midnight»]

Solution: standardize English.

I'll wait while you guys handle that part.