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 - Uffe

Pages: 1 ... 55 56 [57] 58 59 ... 85
General Board / Re: Hyperlink to a specific element of a diagram?
« on: November 30, 2015, 10:33:48 pm »
Hi Gergely,

You can select and deselect objects in a diagram by adding them to / removing them from the Diagram.SelectedObjects collection.

Whether this also re-pans / re-zooms the diagram to ensure the selected objects are in the field of view I don't know, but my guess is it doesn't.

I agree that this anchor-type hyperlink function would be useful, so perhaps you should write it up in the Suggestions and Requests forum and see if anyone else agrees.


General Board / Re: Distributed team deployment and EA LITE Best c
« on: November 19, 2015, 07:51:10 pm »
Hi again,

An HTML dump is definitely worth looking at, if the intended recipient only needs to look at models, not update them.

The HTML version of a model is easier for an EA novice to navigate, but it obviously can't be used for things like document generation or other stuff you need the EA client for.

The default HTML template is OK, but of course it dumps everything. You can tweak it and remove some of the stuff you're not using. Note that what comes out is all JavaScript, so that needs to be enabled in the browser. It does not need a web server, it's a self-contained file package. How nicely it plays with SharePoint I couldn't say.

the scenario - disprate working teams, we will be purchasing approx 20 corporate floating, areas of development are segregated, and checked out to stop two projects working on the same elements at the same time as this would be a contractual issue, I am after LITE to allow the customer to view the current baseline...
In a DBMS setup (with security enabled) you lock the packages you're working on, so they are protected from concurrent editing. However, if a diagram in team A's model contains an element from team B's model, and team B changes that element, the changes will be reflected in team A's diagram. This is true if you use file-based repositories and version control too, but in that scenario team A decides when to "get latest" version of team B's model.


General Board / Re: Distributed team deployment and EA LITE Best c
« on: November 18, 2015, 11:57:11 pm »
Hi Robin,

I tend to recommend against using EA Lite at all, since it's very limited in its abilities. I'm pretty sure it can't access version-controlled models, nor can it do baselines. I also believe it won't run Add-Ins, which might be a factor in your case.

Another option for your deployment scenario is to use EA's cloud facilities. This does not mean you store your models on a Sparx server, but that you upload them to, and retrieve them from, a server of your own. Again, I don't know whether EA Lite can do this but I think not.

One reason not to use EA's integrated version control is that EA's set of permissions does not differentiate checking in from checking out. In other words, if you have permission to check packages out of the version control system, you also have permission to check them back in -- not what you want if you have any concerns about data integrity and the EA skills of your customers.

With the cloud storage model, you can specify read and write access separately.

Version control, which I wouldn't recommend for your scenario, requires the Desktop edition (the cheapest). Baselines and cloud connections require the Corporate edition (two steps up).

If you want to squeeze every cent, you can set it up so that only one user requires the Corporate license by making it their job to retrieve the models into a shared repository, to which the other users connect with their cheaper licenses. But you should also take into consideration the costs involved with managing different types of license and/or different types of client (if you do use EA Lite).

I'd say a DBMS repository for the development team, and for the customer either a replicated repository as Geert suggested or a cloud server. I wouldn't bother with version control at all, but use baselines instead.


General Board / Re: Logical model from Reversed engineered model
« on: November 20, 2015, 06:58:48 pm »
Hello Yves,

In any diagram, you can deselect the attributes in the Element tab of the diagram properties. This will hide database columns as well.



General Board / Re: Link to Element Feature
« on: November 17, 2015, 01:27:30 am »
Hi Michel,

The context menu item is disabled on self-referencing connectors.

I tried to hack my way around by creating a copy of the class, drawing the connector between them, linking to two different attributes, then moving one end of the connector to refer back to the original class -- no go. The end that was linked to an attribute becomes unlinked.

It might be possible to achieve with a more involved hack which bypasses the GUI and updates the database directly, but I'm thinking that if they've gone to the trouble of implementing the unlink-on-connector-move there's probably a good reason.



General Board / Re: Trace requirements to implementation code?
« on: November 16, 2015, 08:52:22 pm »

Yes, you can link code to requirements in EA.

What you need is to get both the code and the requirements into your model. The code can be reverse-engineered, and you can trace requirements to individual attributes and operations, although this is an EA-specific function which is in violation of the UML standard and thus unlikely to be supported in any other tool.

Pulling the requirements in from another source is also possible, but unless you're using DOORS there's no ready-made solution. If you've got them in a simple format which you can Excel-ify, you can turn that into CSV and import the requirements using EA's CSV import.

You will need to draw the relationships yourself. EA does not have a function to import code + requirements.

As to the testing side, EA has two different tracks. You can either represent your tests with model elements (there's a Test Case element, or you can use sub-typed requirements), or you can use EA's Testing facility.

The latter is geared towards actually running your tests from within EA, while the former is better if you just want to see a visual representation of your tests, trace them to your requirements and/or code, and pull some reports out of the model.

Again, there isn't much support for pulling test cases into the model from another source, but if you can get them into a CSV format you can import them that way.

You can do all the above without using SysML if you wish.



General Board / Re: Constrain model elements available
« on: November 09, 2015, 07:47:09 pm »
And just to pile on,

EA isn't the tool to use if metamodel constraints are critical. In EA, modellers will always be able to work around such constraints with very little effort.

In addition to the above advice, you can restrict users' permissions in a repository (referred to as "project security" in the help), but that restricts functionality (eg document generation), not the metamodel.

You can also restrict whole sections of the GUI using Workspaces and Commands (under the View menu); again, this restricts functionality rather than modelling language.



General Board / Re: UML Links in EA not supported?
« on: November 06, 2015, 02:37:34 am »

Well, yes... ish.

There's no "Link" in the toolbox, but the toolbox does not always show all permissible connectors, just the most common ones. If you use the quick link to draw a connector between two objects in a component diagram (or an object diagram), the pop-up menu will have "Link" as the first item.

However, once the "Link" connector has been drawn, it appears that it is in fact an Association (it's referred to as such in its properties dialog). So there doesn't appear to be a cleanly defined "Link" connector type.

This is in EA 11.1. It may have changed in 12.



General Board / Re: Security users - windows authentication
« on: November 03, 2015, 07:45:44 pm »
Hi again Heiko,

What I did (and missed to mention, sorry) was to use the "Login as another user..." entry in the Security submenu.  This does not seem to work at all with Windows authentication.
No, it doesn't and that's by design.
If you're using Windows authentication, EA uses the current Windows user to authenticate access to the project without a password check. As long as you're still logged in in Windows as the same user, you can't change to a different Windows-synched user in the EA project.

However, you can still use "Login as Another User" to switch to non-Windows-synched users, eg admin. But not back again. (Which is annoying, and would make a good feature request.)

The name of the model shows on the top status bar.
This is a common misconception. An EA project has no name.
  • If you're using a file-based repository, the top bar shows the file name.
  • If you're using a DBMS connection, the bar shows whatever name you have specified -- which might be completely different from what your team mates have specified in their connections.
    (Some people think the name of the database -- "Initial Catalog" in the connection string -- is shown, but this is incorrect. It's good practice to specify that as the name, but it's not automatic.)
  • If you're using an EA shortcut (an .EAP file containing a connection string), its file name is shown, even if the connection string contains a different name.
In a deployment where multiple EA projects are used, and especially if the same users accesses several of them, I always set up a shared directory with .EAP shortcut files and tell people to use those instead of rolling their own DBMS connections, simply because this way everyone sees the same name in the top bar.

This directory sits neatly alongside the shared directory for custom MDG Technologies. :)


General Board / Re: Security users - windows authentication
« on: October 29, 2015, 07:14:29 pm »

When using Windows authentication, EA checks whether the logged-on Windows user has an account in the project (t_secuser, the user name includes the domain).

I don't know whether or not EA actually requests authentication from the Windows domain controller, but the authentication for EA project access is in any case completely separate from the underlying database access.

You can, if you want, set it up so that everyone connects to the database using the same credentials (although this is obviously poor security), but that will not affect the EA project authentication.

Other problems you may run into:
  • The database user does not have proper permissions in the database (all database users need write permission, even if their EA project user group does not have write access).
  • If you use Windows authentication in a file-based repository and move/copy the file, the users are locked out. The solution is to disable and re-enable Windows authentication. (Not relevant here, included for future searches.)


General Board / Re: Document template in several projects
« on: October 30, 2015, 02:09:22 am »
Hi Ronald,

Yes, there are two basic ways to do that.

The simple way is to export/import the template(s) as reference data (Project menu -- Model Import/Export).

The more elaborate way is to create an MDG Technology (EA's way to package adaptations) to hold the template, and use that MDG Technology in both your projects.

If you're looking for a simple one-off, do the import/export. If you want easier maintenance down the line, set up an MDG Technology.



General Board / Re: Linked element feature
« on: October 27, 2015, 09:28:06 pm »

"Link to element feature" is an EA feature that has no basis in UML, where connectors can only go between elements. So its implementation is by necessity a bit of a hack.

You can't find the linked features in regular RTF templates, but you can create a script that finds them and then use that script in a template fragment.

The information is stored in Connector.StyleEx, which is a string of key=value pairs. The keys to look for are LFEP and LFSP, whose values correspond to the GUIDs of the linked features.

Note that there is nothing to signify what type of feature the connector is linked to, so you'll simply have to find out by trying Repository.GetAttributeByGuid(), Repository.GetMethodByGuid() etc.



General Board / Re: Export EA diagram to Visio
« on: October 20, 2015, 11:15:49 pm »
Hi Tehila,

Not with any Sparx-supported method.
Possibly if some third part has built a Visio exporter, or an in-Visio EA importer, but I haven't heard of one (not that I've looked all that hard).


General Board / Re: Cannot get nested classes in component to disp
« on: October 16, 2015, 02:57:37 am »

No, you're right -- classes are not structural elements. You have to drag the classes onto the diagram separately.

Once they're placed there (you may need to increase the component's size) they will follow the component around when you drag it, but there's no way that I know of to make EA display them automatically.

If you go to Features & Properties -- Feature and Compartment visibility (Ctrl-Shift-Y) and select Structural Compartment, this will prevent you from dragging the classes out of the component once you've placed them on the diagram. But the classes still won't be displayed automatically.

There is another way, however.

If you create a composite diagram in your component, and place your classes on that, you can display that diagram either inside the component ("compartment", which I prefer) or change the component display to look like a diagram ("frame"); both options are under New Child Diagram.

With this, you can't interact with the individual classes in the "higher" diagram because they're not actually there, it's just an image of another diagram. But if you've got relationships between them those will show up.



General Board / Re: Inserting a Document
« on: October 09, 2015, 06:44:34 pm »
Hi Guillaume,

Since all element types have "file" links, I agree that the default EA behaviour should be to open the regular properties dialog. But what you're after is pretty simple to implement in an Add-In. I've done so in the past.

If you want to get a little fancy, you can create an "external" stereotype which applies to Artifact, add an "L", or an "X" maybe, to the shape script and icon (in place of the "A" shown in linked documents), and only launch on double-click for those artifacts. I also think you should launch all files, if the user has added more than one.

Of course, if Sparx were to implement that, they could create a custom properties dialog for it and include it in the Artifacts toolbox page. But a home-grown solution is a matter of hours to implement.



Pages: 1 ... 55 56 [57] 58 59 ... 85