Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - PeterHeintz

Pages: 1 ... 3 4 [5] 6 7 ... 58
In general this example diagram looks a bit strange to me, because the IDB seem to have blocks rather than parts.

Ignoring that fact, SysML allow connecting parts without ports. Not using ports to connect part1 with part 2 means part1 is connected somehow to part2 but I do not tell you how or I do not know how.

If the two parts are connect with ports having classifiers, it means part1 is connected with part1 port and the details are defined in the part1 port classifier to part2 with the part2 port and the details are defined in the part2 port classifier.

If they had added the counterpart ports to the CAN_Bus as well, it would make clear that there is a one to one relationship between ports.

Without those ports, you are allowed to implement that one to one relationship, or you are allowed to implement a one to one relationship ports within a super port, or you are allowed to implement a super port which can handle all ports connected to CAN_Bus in common.

In other words, yes there is a semantical difference.

General Board / Re: Naming conventions for interfaces between SysML blocks
« on: September 05, 2019, 03:04:04 am »
In such a scenario, I typically give no port a name, but only a classifier only.
The drawback in EA is, that within some EA feature, the names are relevant to distinguish because; classifiers are not shown in those features.

However, by doing so, you get just rid of your problem of thing about the port names. I use port names manly to distinguish, when I have several ports with the same classifier.
If you use the interface block for both sides you have to set the isConjugated (as you have done), because in fact an interface has a provider and a consumer somehow. In this sense the interface block is a kind of short cut.

Your port names are for sure legal, but at the end, those names define somehow what ports of the OS should be connected which ports of the middleware. Once the connection is established in the IBD it is just done.

In principle you could do reverse engineering to get modelling elements representing you code. This, I assume works fine for OO languages (I did already for Java). If reverse engineering of c-code would work as well, you need to find out somehow.

Hi Fabio,
Here you might get some feedback or you might get not. To rise a real bug report you should go to Sparx homepage under „Support/Contact Us“.

Hi Fabio,
page 148 is just graphics and do not tell so much.

But when looking on page 151
“An Action appearing in an “AllocateActivityPartition” will be the /client (from) end of an “allocate” dependency. The element that represents the “AllocateActivityPartition” will be the /supplier (to) end of the same “allocate” dependency. In the «AllocateActivityPartition» name field, Properties are designated by the use of a fully qualified name (including colon, e.g., “part_name:Block_Name”), and Classifiers are designated by a simple name (no colons, e.g., “Block_Name”).”

I think you are right, but it seems to be ignored by EA. Maybe you issue a bug or a feature request.

if you want to use the relationship matrix you need to establish a real relation by an sysml <<allocate>> relationship.
What you are doing here, is not much more than a visualization of relations you do not have in your model.
Of cause it would be nice if EA would establish that relation for you in background, but it does not and the omg standards are very inconsistent in how to handle “containment”.

Automation Interface, Add-Ins and Tools / Re: report templates
« on: July 15, 2019, 10:39:44 pm »
I am only 99.99% sure, because sometimes I find some tricks to make impossible things possible.
However I remember that I tried it again, but gave it up. What is missed, is that you can not use the template selector for packages as well.

Oh, yes! This is what I was looking for. Thank you KP!

In V15 when you double-click on a diagram element, e.g. to open a composite diagram, the current diagram is closed. By an arrow you can move back to the original diagram.
In former version all diagrams are kept open.
If you navigate/hop to different parts of the "diagram hierarchy", before V15 you might end up with lots of diagrams opened, what is not so nice. Now only one diagram is open, what is nice if you used this hopping just to get to that diagram.

However I use this hopping most frequently to open at least two diagrams. One diagram I want to change and the others to look what I have to do. This more frequent use case (at least more frequent use case for me) worked fine for any version before V15.
To address both use cases, I would prefer a simple way to decide if I want to keep the current diagram open or not before I hop to the next.
If this is not possible for whatever reason, I would very much like if the V15 change would be cancelled.

What is the opinion of other users?

I assume stirring the menus each time around, is not intended to be an improvement of EA, but to be a improvement of us, by finding features you have never missed when searching the features you need.  ;)

Yes, I try to persuade my users to use the GUID as identification of anything (not only requirements). But I admit with no success so far.

When I ask why a requirements number or even different number groups are needed, I always get the answers “to have identification within e.g. PDF reports” and “to provide identification to which part the requirement belongs to”.

However these statement are valid to any element, not only requirements and can e.g. be addressed by the GUID and packages or links.

The GUID is not accepted because it is not "readable". And I admit “Component_Req_0001” is more readable than {EA39C8DE-572B-43b2-80DF-BA76BF19EDFB}.
But the message “This is Requirement 1 of the Component “Component” does not say so much as well. And finally if the requirement number is even used, I only see Copy/Past either to reference it in something else or the do a search.

For the Copy/Past actions it is quiet the same if you copy/paste “Component_Req_0001” or “{EA39C8DE-572B-43b2-80DF-BA76BF19EDFB}.
But we have to have “Component_Req_0001” just because we have to have. ::) ::)

No, this is not supported by EA. You might do that by writing a script.

In build 1505 you can only browse one diagram not more. This is hopefully a bug and not a new feature.

This effect is only there, if you try opening diagrams by double click on diagram elements having composite diagrams assigned to.

Bugs and Issues / Stereotype Dialog is Faulty
« on: June 18, 2019, 09:40:52 pm »
The new stereotype dialog introduced in V14 seem to intend to address a better filtering of many stereotypes. However now setting stereotypes is more often a complex or even impossible search.
Example: In V14 in tried to change an object in an SysML activity diagram to a <<datastore>>, but I had to go to UML instead of SysML1.5 to do that. In build 1503 even in UML the <<datastore>> has gone. I have no idea how I can set my object to be a <<datastore>> in build 1503.

Please validate what is shown in the stereotype list associated to a Perspective/Profile.
Currently it is more or less useless.

Pages: 1 ... 3 4 [5] 6 7 ... 58