JSON seems to me a pretty poor alternative to EXPRESS. Are you sure it's not about using JSON instead of XML analogous to ISO 10303-28?

The authors of -243 use the JSON schema as an input to OpenAPI tools to build webservices. The Application Protocol model with definitions of what things mean is in SysML, this will also be included in the package for the standard.

A new XML exchange format for models defined in SysML is being defined in ISO 10303-15.

Just had a look at the new ISO 10303-1:2021 (second edition, replacing the frist edition from 1994) which appeared in March. EXPRESS is still the primary description language and will obviously remain with us for a long time to come. Section 6.3.3 explicates "EXPRESS models provide the basis for all specifications of product information in ISO 10303". SysML is not even mentioned (in contrast to e.g. UML). Section 6.3.6 states "ISO/TS 10303-25 does not map all EXPRESS constructs to the UML meta-model, because that meta-model does not support all the corresponding EXPRESS concepts. The specified mapping is a one-way mapping from EXPRESS into the UML Interchange Meta-model". There is no reason to assume that SysML is more expressive than UML in this regard. ISO 10303-11 was last time confirmed in December 2019.

ISO 10303-15 is apparently a technical specification on how to transform SysML XMI to XML Schema (XSD) format.

A decision was made to have ISO 10303-1e2 just reflect what has been published to date, then start work immediately on ISO 10303-1e3 that will also contain the description of how SysML will be used.

The main use case of ISO 10303-15 is to be able to validate XML exchange files in the new format.

I expect it will take many years until the mentioned technical specifications become regular standards. I would be very surprised to see a third edition of ISO 10303-1 within the next eight to ten years. From what I've seen up to now is that SysML might get a similar position as UML in the todays version. Even if there is a feasible mapping from SysML XMI to EXPRESS, the official description language will still be EXPRESS in future editions of the ISO 10303 series. Fads come and go; we saw this with UML, and we will see this with SysML; SysML is simply not suited to replace EXPRESS, even less than UML. I bet that in seven years at the latest, SysML will be eclipsed by a new fad.

Well, prepare to be surprised.

If I were actually surprised, ISO/TC 184/SC 4 would have done an extremely poor job in terms of reporting and public relations. But let's continue this discussion in five years, then we'll know more.

There has already been one successful ballot on ISO 10303-4000, you won't have to wait five years.

I don't really see that there was a need for public relations, we are trying to maintain backwards compatibility with existing software and models.

> I don't really see that there was a need for public relations, we are trying to maintain backwards compatibility with existing software and models.

Sounds like a statement from a state monopoly with a five-year plan.

Joking aside: You have just experienced a use case for this yourself. There is nothing at all on the web about what you claim. So I as a developer and user of the standards have no information whatsoever about what is bubbling up in the "black box ISO", if EXPRESS is really to be replaced by SysML (which I still do not believe). There is no mention of this in the presentations, strategies, and architectures I have access to.

Who is "we"? What's your role in ISO?

> Who is "we"?

TC184/SC4/{WG11,WG12} - a mix of people who have been developing STEP since the start and some new blood.

I'm one of the older ones, I wrote ISO 10303-45 and maintain an EXPRESS compiler used in the document production process.

> TC184/SC4/{WG11,WG12}

I see; this gives a new perspective to this discussion and increases my uneasiness; looks like the "holy wars of STEP" (NIST Special Publication 939) was going into a new round behind the scenes (or never stopped).

Does that mean we have to say goodbye to EXPRESS (i.e. it is optional as a description method)? Or is SysML just another method to specify Business object models (next to EXPRESS and UML)?

> an EXPRESS compiler used in the document production process

Do you mean the Eengine (written in Lisp)?

> Concerning "public relations":

e.g. the SysML v2 Submission Team (SST) did quite a good job in communicating their plans and the future direction of SysML; so even long before the standard is published I have a good grasp of what to expect, so also on the new text representation which will make MBSE much more efficient (like the advent of hardware description languages quickly made schematics obsolete for real world designs); this gives me enough time to prepare for the paradign shift (e.g. adapt my tools etc.). It is only to be suspected that STEP will still be based on the old SysML version with a different metamodel and no textual representation (i.e. the only "textual representation" is XMI or XSD).

I see no "holy war", certainly not like those in the past and I was there for them.

Business Object Models are deprecated in ISO 10303-1e3 and replaced with Domain Models specified in SysML (v1), the ISO 10303-242 one will look very similar to the current BOM and will have alternative EXPRESS and XSD representations. The SysML Canonical XMI files that are the specification of the standard will be provided with it, a user can use the tools I linked to in order to convert this to something that can be loaded into a SysML editor.

Someone using ISO 10303 EXPRESS models now can carry on using them in the future, there isn't some change they are forced to adopt. The only real user visible difference will be a more complete mapping between a Business Object Model/Domain Model and the traditional EXPRESS Application Protocols.

I am maintaining Express Engine and adding new features, I'm not the original author though.

Thanks. If there is any presentation, whitepaper, strategy or other document describing this new direction of STEP I would be very interested.