Provide a model organisation for Papyrus

[cecilebraun]

@UweSteinkeFromSiemens provides a model structure to do collaborative work with Papyrus see https://github.com/openETCS/modeling/tree/master/ModelSpace

[UweSteinkeFromSiemens]

The task is to prepare a model and file structure for collaborative modelling in WP3. The model structure is given by the SysML model "WP3-TaskForce-Architecture" at https://github.com/openETCS/modeling/tree/master/model. Exactly for this model the structure shall be prepared and stored at https://github.com/openETCS/modeling/tree/master/ModelSpace/SysML. The structure shall be expandable in parallel to the growing model.

Please take as an advice what I did already for Scade Suite with the following capabilities:

• Each of the subblocks of "WP3-TaskForce-Architecture" has its own submodel and subproject, so that it can be assigned to exactly one person for modelling without unneccessarily affecting others.
• Integration models and projects are prepared for the stepwise integration of the submodels.
• Type definitions to be shared between several submodels reside in separate models and can be included as libs into each submodel where needed.
• The "ETCS Language" type definitions (subset 026-7, 026-8) are available in libs and can be used for modelling.
• Several basic types (for length, speed, acceleration, gradient, ...) for consistent internal usage in the models have been defined additionally.
• Each subproject is based upon the same project template and sets all required options.

Cooperation between different persons on the model structure requires the exclusive assignment of a submodel to only one person at time. Since github does not have this capability, it requires manual care between the participants until the tool chain has solved it.

A similar approach should be used for SysML modelling too.

[UweSteinkeFromSiemens]

The collaborative model and file structure for SysML should rely onto the openETCS tool chain with the requirements management and data dictionary integrated.

[MatthieuPERIN]

Solution proposed during the WP3 meeting on 2014.02.26 @ Munich:

• [ ] A document with precise process to create and modify the SysML document will be provided (@MatthieuPERIN)
• [x] A test run will be held base on this document and the first WP3 action on core function (@MatthieuPERIN & @astante)

[MatthieuPERIN]

The test have not proving to be successful on WP3.

The document is under formatting and completion, a draft is present in the modeling repository.

[MatthieuPERIN]

On going, new issue found (see https://github.com/openETCS/modeling/issues/35)

[MatthieuPERIN]

Due to reduction on WP7 force, this task have been postponed.

[MatthieuPERIN]

As a work around have been found with WP3, the issue is closed.

(see openETCS/modeling#35 for more details)

[MariellePetitDoche]

Unfortunatelly, for those who start to model now, or letter, no documentation are available to start correctly a SysML model for openETCS and to organise it.

What is the concrete result of the "work around" ?

[MatthieuPERIN]

Yeah I know about that, but as I have shifted manpower to WP7 the document has not move from the draft present at https://github.com/openETCS/modeling/blob/master/ModelingRules/SysML_modeling_rules_draft.docx

To me this document may be really impacted by shifting Toolchain to Luna as some diagram placement may become invalid in the new release (no more IBD diagrams inside a Block are allowed)

@jastram , @cecilebraun , @UweSteinkeFromSiemens Do I have to work on that document ?

[UweSteinkeFromSiemens]

Dear Mathieu, the document update is necessary, really. But it’s date is not critical, since the WP3 practical modeling work is currently performed with SCADE Suite. Perhaps, the perspective of the WP3 architecture team (@Baseliyos ++) might be different …

Von: Matthieu PERIN [mailto:notifications@github.com] Gesendet: Montag, 13. Oktober 2014 16:16 An: openETCS/toolchain Cc: Steinke, Uwe Betreff: Re: [toolchain] Provide a model organisation for Papyrus (#294)

Yeah I know about that, but as I have shifted manpower to WP7 the document has not move from the draft present at https://github.com/openETCS/modeling/blob/master/ModelingRules/SysML_modeling_rules_draft.docx

To me this document may be really impacted by shifting Toolchain to Luna as some diagram placement may become invalid in the new release (no more IBD diagrams inside a Block are allowed)

@jastramhttps://github.com/jastram , @cecilebraunhttps://github.com/cecilebraun , @UweSteinkeFromSiemenshttps://github.com/UweSteinkeFromSiemens Do I have to work on that document ?

— Reply to this email directly or view it on GitHubhttps://github.com/openETCS/toolchain/issues/294#issuecomment-58897793.

[MariellePetitDoche]

Uwe, Matthieu,

What is missing currently is a document for a new modeler to describe how he can start to model : initially it was said with SysML and papyrus, Uwe you said that currently the work is done with Scade suite directly... This is not clear and it is really difficult to find the information in the different repository or issues tracker... if this information is present.

[UweSteinkeFromSiemens]

Dear Marielle, Besides the model organization and the modeling rules at https://github.com/openETCS/modeling/tree/master/ModelingRules starting with a new modeler requires some coordination with the ongoing work and preparation of the existing model structure. A good starting opportunity would be attendance to the WP3 GotoMeetings on Friday and Tuesday morning and creating an issue on the SRS-Analysis repository.

Which function do you plan to model, and how does it fit to the sub-models under construction?

[MariellePetitDoche]

I fear if the information exchanged during telco are not correctly tracked, divergence between the informations and lots of difficulties to integrate and validate the different parts of the model.

I have planed to create an issue on what I have started, as said in Bruxels, but with a top-down approach expessially to try to fit the sub-models under construction and to test the under development pluggin of the toolchain. Thus when I look on the current SysML model, it is really difficult to have a clear state of the model, and how to fit to it on the best, for the moment i think the SysML modelling rules is just a draft, and not yet sufficient.

[UweSteinkeFromSiemens]

Therefore, the live show currently takes place in SCADE Suite, while the SysML model is hibernating.

[jastram]

@UweSteinkeFromSiemens , considering your last comment: Do we still need this at all? If no, please close.

[UweSteinkeFromSiemens]

Since some colleagues at DB are using SysML based on SCADE Suite, I will forward this question to @BerndHekele

[BerndHekele]

ince the issue was opened we gained a bit of expierence in handling the models. I believe, we can close the issue with status quo.