Essentially, yes. But we have agreed a specific version (recognizing past sensitivities when moving from version to version (even minor versions)).
If you take a look at the guidelines referenced above, or the corresponding video fragment, we identify Papyrus version 3.2.0 RC4 as the version to use. The video shows you how to get this version.This should be run on Eclipse version 4.7.2 “Oxygen”. We should also use Gendoc version 0.6.0 as noted in the guidelines, but that has not yet been agreed.
I should also mention that the RC4 load happens to also be the formal release of 3.2.0 (we set the version before they formally released).
I am not aware of any differences but I have not run the tool on Mac. The only capability we have used to thsi point is gendoc (I am not aware of another capability), was there any reason why you felt it inappropriate?
About gendoc - if there is no alternative to Gendoc, I was wondering what the hold-up in agreeing to Gendoc v 0.6.0 is? (I don't see any advantage to Gendoc v 0.5.0).
It depends upon the question. We have use the Papyrus forums for some questions but in the standards work there are a few of us who tend to use each other's knowledge built up from struggling with the tool. Did you have specific questions (I guess they are not covered by the basic tutorial that I gave)?
Probably a template issue (although I under the impression that the templates were OK in NFV - I have worked closely with Marc Flauw who did the originals and I used some of his work in the ONF templates).
Another question based on experience with Sparx Enterprise Architect: Sparx gives the capability of creating "linked documents" to packages, classes and other UML elements. These documents become part of the model; that a linked document exists is visible by a little icon in the class or package icon on the diagram; a linked document transports with the model, is editable and printable with the model.
Example: the highlighted text in the Word file here is from a linked document.
Does Papyrus provide a similar facility? If not, what is the suggested way to keep documentation co-resident and co-transportable with the model? In practice, in my opinion how to use a model, what the model intends to represent is, etc., is subject to widely varying interpretation without supporting documentation, including examples. In such a wide community as ONAP, this cannot be left to folklore.
I am not aware of such functionality in Papyrus, but you can ask the question directly in the Papyrus Eclipse forum.
With Gendoc, it is rather the reverse, including model in doc. For ETSI NFV, we generate the complete model as 1 document. ONF is creating more specific documents using only a part of the model and merging with manually entered information in Word.
The Draft_IISOMI-515_Papyrus_Guidelines_v1.3.04.docx proposes a default Package structure, with packages titled Associations, Diagrams, Imports, Interfaces, Notifications, ObjectClasses, Rules, TypeDefinitions. However, what is intended by this structure and the benefits of this structure are not explained. Some information about that would be appreciated.
An attempt to put the VNF Descriptor Design Time Model Clean Version into Papyrus. Includes a gendoc template and the generated document. Maybe it can serve as a conversation starter on how we should organize the ONAP data model(s).
I think the diagram below is shows what is intended semantically in the VES event, but I'm not sure it will map to the intended YAML or JSON, in which there will be no eventBody, but a substitution of one of Fault, Measurement, Heartbeat or StateChange.
19 Comments
Michela Bevilacqua
Is Oxygen the Papyrus version to start to use in ONAP ?
Nigel Davis
Essentially, yes. But we have agreed a specific version (recognizing past sensitivities when moving from version to version (even minor versions)).
If you take a look at the guidelines referenced above, or the corresponding video fragment, we identify Papyrus version 3.2.0 RC4 as the version to use. The video shows you how to get this version.This should be run on Eclipse version 4.7.2 “Oxygen”. We should also use Gendoc version 0.6.0 as noted in the guidelines, but that has not yet been agreed.
I should also mention that the RC4 load happens to also be the formal release of 3.2.0 (we set the version before they formally released).
Arun Gupta
Nigel, is there an alternative to Gendoc for generating documentation from Papyrus?
Thanks!
PS: also - are there any meaningful differences between Papyrus on Windows and Papyrus on MacOS X?
Nigel Davis
I am not aware of any differences but I have not run the tool on Mac. The only capability we have used to thsi point is gendoc (I am not aware of another capability), was there any reason why you felt it inappropriate?
Nigel
Arun Gupta
About gendoc - if there is no alternative to Gendoc, I was wondering what the hold-up in agreeing to Gendoc v 0.6.0 is? (I don't see any advantage to Gendoc v 0.5.0).
Nigel Davis
No holdup, just we have not yet agreed. I have no issue. If I own the decision, then it is 0.6.0.
Arun Gupta
What is the best place to get newbie questions answered? Is it the Eclipse Papyrus forums?
Thanks!
-Arun
Nigel Davis
It depends upon the question. We have use the Papyrus forums for some questions but in the standards work there are a few of us who tend to use each other's knowledge built up from struggling with the tool. Did you have specific questions (I guess they are not covered by the basic tutorial that I gave)?
Arun Gupta
I'm trying to generate documentation using the template available with ETSI IFA-015 gdNfvTemplate.docx.
Documentation is generated, but the Microsoft Word document always requires repair. The message is:
"Ambiguous cell mapping encountered. Possible missing paragraph element. <p> elements are required before every </tc>.
Location: Part: /word/document.xml, Line:1 Column: 0"
Is this a gendoc bug or a bug in the ETSI template?
Thanks!
Marc Flauw
Hi Arun,
This is a known bug in Gendoc that occurs when there is an empty cell in Word:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=478883
I fixed the cases I found in ETSI NFV IFA015, but there might be others depending on the model used.
The way to debug is to open the /word/document.xml (open word as a zip) and find the </tc. without a <p>
Best regards,
Marc
Arun Gupta
Thanks, Marc!
Nigel Davis
Probably a template issue (although I under the impression that the templates were OK in NFV - I have worked closely with Marc Flauw who did the originals and I used some of his work in the ONF templates).
Arun Gupta
Another question based on experience with Sparx Enterprise Architect: Sparx gives the capability of creating "linked documents" to packages, classes and other UML elements. These documents become part of the model; that a linked document exists is visible by a little icon in the class or package icon on the diagram; a linked document transports with the model, is editable and printable with the model.
Example: the highlighted text in the Word file here is from a linked document.
Does Papyrus provide a similar facility? If not, what is the suggested way to keep documentation co-resident and co-transportable with the model? In practice, in my opinion how to use a model, what the model intends to represent is, etc., is subject to widely varying interpretation without supporting documentation, including examples. In such a wide community as ONAP, this cannot be left to folklore.
Thanks!
Marc Flauw
Hi Arun,
I am not aware of such functionality in Papyrus, but you can ask the question directly in the Papyrus Eclipse forum.
With Gendoc, it is rather the reverse, including model in doc. For ETSI NFV, we generate the complete model as 1 document. ONF is creating more specific documents using only a part of the model and merging with manually entered information in Word.
Best regards,
Marc
Arun Gupta
The Draft_IISOMI-515_Papyrus_Guidelines_v1.3.04.docx proposes a default Package structure, with packages titled Associations, Diagrams, Imports, Interfaces, Notifications, ObjectClasses, Rules, TypeDefinitions. However, what is intended by this structure and the benefits of this structure are not explained. Some information about that would be appreciated.
Thanks!
-Arun
Arun Gupta
VNFDescriptorModel.zip
An attempt to put the VNF Descriptor Design Time Model Clean Version into Papyrus. Includes a gendoc template and the generated document.
Maybe it can serve as a conversation starter on how we should organize the ONAP data model(s).
Thanks!
-Arun
Arun Gupta
Hi,
Is there an IISOMI guideline for how to handle in UML the XML Schema "choice" element?
Thanks in advance!
-Arun
Marc Flauw
Hi Arun,
In IISOMI-514, ML Modeling Guidelines, look at clause 7.5.2 "Choice" and the use of the "Choice" stereotype.
Best regards,
Marc
Arun Gupta
Thanks, Marc!
I think the diagram below is shows what is intended semantically in the VES event, but I'm not sure it will map to the intended YAML or JSON, in which there will be no eventBody, but a substitution of one of Fault, Measurement, Heartbeat or StateChange.