UAF MDG Request

[Richard L. Warren]

You really should take a look at the new Unifed Architecture Framework (UAF) that the OMG has come up with.  Given it's integration with UML and models that you also support, it would be a welcome addition to the Zachman Framework MDG you have currently.  I suggest you look at the specification at https://www.omg.org/spec/UAF/1.0/PDF and the appendix A at https://www.omg.org/cgi-bin/doc?formal/17-12-02.pdf for a real eye-opener.  Please add me to a list is you decide to move forward with this.  I'd pay real money on TOP of my ultimate subscription for an MDG that implements this framework!

P.S. As you did with BABOK and the IIBA, think about a relationship with OMG for this standard too.

Thanks,
Richard

[Eve]

Hi Richard,

EA already implements UAF and is a member of the OMG.

[aldr1c]

We are seeing support for standards slipping within the tooling - principally amongst the MDG elements.  UAF 1.2 is out, SysML is 1.6/1.7 with v2 in Beta, TOGAF is at V10, and so forth.  is there a published strategy for uplifting the support for these standards?  Having to do more and  more work in tools other than EA and as a consequence, having to direct clients to other tools for these capabilities, is a shame.
Things will be even more of an issue with the move of sysml (and soon uml and bpmn) to a KerML based representation, so I guess I am asking - should i/we 'keep the faith', or cut and run?

[Eve]

No, there isn't a published strategy.

I can't give any announcements, but future SysML2 support was discussed recently.

[skiwi]

Can I add that some of the analysis patterns, e.g.  Azure (Nov 2022), AWS etc. need to be frequently updated,
and must be able to be updated on the fly, rather than waiting for new releases.
(BTW could we please have consistent naming standards for these patterns. e.g. Nov 2022, Rel 1, Rel 5, v1.5 are not helpful when attempting to understand what the most recent version is (tip: date works well))

[bergtwvd]

I would like to use UAF 1.2 rather than 1.0.

I tried to import the UAF 1.2 profile file from the OMG site as a UML Profile in my project. But I got an error: "There was an error processing the supplied CML profile file". Why does this not work?

See https://www.omg.org/spec/UAF/1.2 for files.

[qwerty]

I'd guess the OMG published a metamodel of the profile and not an EA MDG. Eventually you can hammer tha MDG our of the profile. Not sure whether that's a) possible and b) if a) then how difficult that would be.

q.

[aldr1c]

This is becoming a problem.  Although we can create an MDG from the standard (NATO did this for NAFv4 I believe) it is problematic as there is no guarantee that other teams, orgs, etc will follow the same rules (compliance, compatibility, etc) and do not put their own juice into the patterns/stereotypes etc.  Not unless you are a NATO and as such can dictate or mandate use of a particular freely available MDG (or third party paid at worst).

With SysML going to V2, and UML, BPMN and UAF rolling into the same KerML basis, there is a real and growing movement towards tools that do support these up-to-date and changing standards, which 'traditionally' Sparx doesn't speak on, until a grand reveal.  Personally I have been in the position twice this year of including tools in recommendations to clients that would ordinarily not be there, and having to point out this shortcoming in standards support within the Sparx eco-system

its a real shame.

[Eve]

This is becoming a problem.  Although we can create an MDG from the standard (NATO did this for NAFv4 I believe) it is problematic as there is no guarantee that other teams, orgs, etc will follow the same rules (compliance, compatibility, etc) and do not put their own juice into the patterns/stereotypes etc.  Not unless you are a NATO and as such can dictate or mandate use of a particular freely available MDG (or third party paid at worst).

I promise that it's not as problematic as if we don't take the time to do the job properly. It's being worked on.

With SysML going to V2, and UML, BPMN and UAF rolling into the same KerML basis,

Just a warning, don't put all your hopes on KerML based versions of the other OMG languages. Yes, SysML2 is in development. It's not even finalized yet, and in the current form they have completely abandoned any form of diagram interchange. There is a lot of work to do on the KerML before it is a fit replacement for MOF based languages. That's not to say there aren't good things in the language, there are things that I'm excited for. We're working hard on SysML2. Even more so than an incremental update to UAF, it's important that we provide a fantastic base for developing future languages.

[aldr1c]

This is a useful response, thank you.  Having some sight of the roadmap, not of the tools development specifically, but of its support for changes such as standards is really key.

The point about KerML isnt really about hopes - it is about options as an architect.  Business modeller tools will have issues trying to adapt, as well as visually biased tool vendors as notations move towards textual biased systems.  To be honest, in analysis and governance, we rarely look at the diagrams anyway - running queries and reports against the model is usually a better option, as that is where the monsters lie.  EA is pretty good at enabling this of course.  That is the underlying concern as we dont want to 'throw the baby out with the bath water'.

If we get comms on where things are going (the OMG roadmaps etc let us know where we may require to prepare for change - essential in multinational, multi-partner efforts with less flexible tools than EA is).  we know that a couple of vendors are planning (for example) SysML v2 support in 2025, so we can plan against that notion.  Not looking for hard and fast drop dead dates, just something like a developer coffee pot or water cooler set of comms.  If that is already available, please do point us at it.