Home » Modeling » Papyrus » Papyrus Community Census
| Papyrus Community Census [message #1061348] |
Fri, 31 May 2013 10:05  |
 Sébastien Gérard
Messages: 121
Registered: July 2009 |
Senior Member |
|
|
Dear all Papyrus Users,
I am Sébastien, the leader of the Papyrus project. We, the Papyrus developper team, would like have an idea of the landscape of Papyrus users.
So if you could post response to this new topic outlining who you are, what is of interest/usage (including the context) in Papyrus, what would be the nice feature to have in it, and any information that could be helpful for ues, we would appreciate this effort.
Moreover, if you have some willing to help us in making Papyrus more and more useful, please do not hesitate to make suggestion here also.
Finally, if you have some success story that can be cited on the Papyrus web site, please do not hesitate provide us th einfo or to directly contact me: sebastien.gerard@cea.fr
Thanks,
Best... Sébastien.
[Updated on: Fri, 31 May 2013 10:08] Report message to a moderator |
|
|
| Re: Papyrus Community Census [message #1061681 is a reply to message #1061348] |
Mon, 03 June 2013 13:29  |
Vincent Hemery
Messages: 7
Registered: April 2013 |
Junior Member |
|
|
I'm Vincent Hémery, working at C-S. I used to be a Papyrus committer, but since I left my previous company, I haven't had an occasion for developping on Papyrus again.
Nevertheless, I keep using Papyrus, in a TOPCASED environment, whenever I need to create UML/SysML diagrams, mainly for specification or illustration purpose. In this context, the diagrams I use the most are Class diagram, Sequence diagram and Component diagram.
|
|
| |
| Re: Papyrus Community Census [message #1062737 is a reply to message #1062585] |
Mon, 10 June 2013 14:53 |
Klaas Gadeyne
Messages: 165
Registered: July 2009 |
Senior Member |
|
|
Hi, I'm working at the Flanders' Mechatronics Technology Centre. I'm mainly [interested in]/[trying to use] Papyrus as
- SysML modeling tool
- DSML creation (and usage) tool, with a focus on formal (read: OCL enhanced) DSMLs
The main "nice features" I'm interested in (to make it a little concrete I tried to refer to forum topics or bugzilla reports):
- SysML modeling.
- DSML creation
* Mainly documentation on how to create a true DSML environment (using the 'lightweight or 'profiling UML' approach): <http://www.eclipse.org/papyrus/custumizersCorner/custumizationExamples/custumizationExamplesIndex.php> is now "coming soon" for a couple of years already  . I know there is some documentation on defining profiles, and customizing \
the palette, but that is only part of the story. Documentation on how to customize the model explorer, how to include validation (see for instance http://www.eclipse.org/forums/index.php/t/478570/), how to customize the properties view, ... Whereas SysML support is available from multiple modeling tools, I think this DSML creation could really be a (the) "selling point" for Papyrus.
In general, more documentation would be very welcome, both from a user point of view (http://www.eclipse.org/forums/index.php/t/478570/), anyone knows how to use the "advanced transition textual editor?" apart from reverse engineering it), and from a developer point of view (in order to also get some patches from the community instead of just bug reports).
But, apart from all those feature request, my impression is that, in general, the stability of papyrus has improved a lot lately [*], as has been the reactivity of some of the developers on this forum and the bugzilla (thank you, Camille and Ed! and others). Also, there were some nice efforts to improve the useability (for instance the simpleUML customization). Also, the CDO integration seems a nice feature.
[*] a unit test for every feature/bug would still be better
[**] And as really nice-to-have: https://bugs.eclipse.org/bugs/show_bug.cgi?id=330199 seems to be "dead" for a while?
|
|
|
| Re: Papyrus Community Census [message #1062758 is a reply to message #1062737] |
Mon, 10 June 2013 15:57 |
Camille Letavernier
Messages: 952
Registered: February 2011 |
Senior Member |
|
|
Hi Klaas,
Thanks for the feedback!
I'll probably be a little bit off-topic, but here are a few comments:
- User and Developer documents: Basic user documentation will be included in Kepler, on the following topics:
- Model search: Done
- Diagram Stylesheets (CSS): Done
- OCL Constraints support (Definition + Validation)
- Profile definition and usage
- General Papyrus tips (Diagram manipulation)
I also plan to write a little bit of developer documentation this summer, because it is highly time to have a reference document for all the new developers and contributors.
Regarding customization tool, unfortunately, we don't have much more to propose yet. The CSS and Palette customization documentations are already available. For the ModelExplorer, we're fully based on EMF Facet 0.1 (Which is still available in EMF Facet 0.3, although deprecated, but I'm pretty sure it will remain for a while). EMF Facet 0.1 was initially derived from MoDisco, so you can find information here:
http://download.eclipse.org/modeling/mdt/modisco/nightly/doc/org.eclipse.modisco.doc/
(The only difference between MoDisco and EMF Facet 0.1 is the NsURI of the Query/Facet/uiCustom metamodels). We should probably embed this document within Papyrus, or at least add an hyperlink to the MoDisco documentation (Which is already embedded)
The EMF Facet documentation describes a newer component, on which Papyrus does not rely.
For the properties view customization, well... It's a little bit more complicated, but hopefully we'll be able to have a usable documentation for SR1.
Quote:[**] And as really nice-to-have: https://bugs.eclipse.org/bugs/show_bug.cgi?id=330199 seems to be "dead" for a while?
Not exactly dead. But it takes quite a while, and nothing has been published yet (AFAIK).
Regards,
Camille
Camille Letavernier
|
|
| |
| Re: Papyrus Community Census [message #1065898 is a reply to message #1061348] |
Fri, 28 June 2013 07:59 |
Marc-Florian Wendland
Messages: 83
Registered: January 2013 |
Member |
|
|
Dear Sebastien,
I'm developing our Model-Based Tetsting Environment on top of Papyrus and,
thus, have learned much about the internal architecture of Papyrus. We use
the UML Testing Profile as modeling notation on top of UML. We do not make
use of most of the offered modeling tooling (e.g., search dialogs etc.) by
Papyrus, but want to provide a more tailored and customized interface to the
users. As a general modeling language environment (for UML), Papyrus suffers
from the same characteristics as any other UML modeling environment, i.e.,
in does only cope with specification compliance, whereas in our opinion, a
test modeling environment must offer methodology-specific support. But as I
said, this is not a problem of Papyrus.
I really appreciate the infrastructure and the extension capabilities of
Papyrus. In particular the customization of the palette with pre and post
actions are very useful to offer dedicated and tailored variants of UML
standard diagrams to the users. We heavily use this concept (with some
proprietary technical augmentation) for providing test-specific diagrams
based on Papyrus' UML GMF implementations.
Regarding customization, Papyrus should be partialy more flexible. For
example, it is not possible to switch off the predefined 'Create Child'
context menu entries. This is contraproductive if you want to allow just one
particular way to model things. As an example, UML does not prescribe what
ValueSpecification shall be used as Message argument. In our tool, we only
allow InstanceValues pointing to an InstanceSpecification. Without being
able to switch off (parts of or entirely) the 'Create Child' entries, user
can still make use of the functionality to produce invalid models. A simple
property tester and an additional extension point could be used for allowing
developers to explicitly switch off entries in certain situations.
Papyrus has to be refactored regarding its proprietary models concept.
Currently, all deployed models must be part of the modelset which disallows
using Papyrus for any other DSLs, because Papyrus tries to create each and
every deployed model king when opening a Papyrus resource. This is really a
hard and unneccessary limitation. A simple solution could be to add the ID
of the model kind into the DI file (i.e., describe the modelset in the DI
file as well). By doing so, it would be even possible to have more than one
semantic models pointing to different physical resources in the ModelSet.
For example, if you want to develop a DSL with Papyrus, the semantic models
could be the UML/MOF file (that will later on be imported by UML2 EMF
importer) and also an instance of the DI metamodel describing the graphical
elements. The DI model is not a UML model, but could be deployed as Papyrus
semantic model as well. This would even allow the Model Explorer to switch
between different semantic models in the ModelSet.
I think the Papyrus development team should, from now on, concentrate more
intensely on stabilization and usability instead of offering more and more
orthogonal concepts like shared model editing and so forth. This is not the
primer of Papyrus, in my opinion, but of other initiatives like CDO or
ModelBus. However, Papyrus need to offer interfaces for these kind of
services. In addition, I don't think that each and every part of Papyrus
extension functionality needs to be realized in a model-driven, generative
manner. For example, the extension of the properties view is a great idea,
but, personally, if I am implementing an extension, I do want to tailor this
to my particular needs and might not be interested in having these parts
generated with the offered widgets by Papyrus. Because these widgets serve
commonly a general purpose. So, my comment here would be to offer the
interface, a comprehensible documentation, but not necessarily a dedciated
generative tooling that is tricky to implement and to maintain.
Finally, I do think Papyrus should try to integrate with a decent document
generator like GenDoc2. It is really hard to find anything useful for
generating documents (.docx, .odt) out of a Papyrus model with all the
diagrams and model information.
I already suggest improvements to you and your team. I would like to
collaborate closer with your UML Sequence Diagrams team since I do have some
nice ideas how this could be made more useful.
Papyrus is a great tool that steadily mature. Let me commend you and your
team for this. And as we discussed several times, I think it is really time
for the community to come together on a regular basis (in terms of Papyrus
Days) in order to share experiences etc.
Hope this comment will be appreciated.
Marc-Florian
"Sebastien Gerard" wrote in message news:ko9sl7$e67$1@xxxxxxxxe.org...
Dear all Papyrus Users,
I am Sébastien, the leader of the Papyrus project. We, the Papyrus
developper team, would like have an idea of the landscape of Papyrus users.
So if you could post response to this new topic outlining who you are, what
is of interest/usage (including the context) in Papyrus, what would be the
nice feature to have in it, and any information that could be helpful for
ues, we would appreciate this effort.
Moreover, if you have some willing to help us in making Papyrus more and
more useful, please do not hesitate to make suggestion here also.
Thanks,
Best... Sébastien.
|
|
|
Goto Forum:
Current Time: Fri Apr 26 16:25:15 GMT 2024
Powered by FUDForum. Page generated in 0.03604 seconds |
] 
Search
Help
Register
Login
Home
