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.

Topics - Ian Mitchell

Pages: 1 [2] 3 4
General Board / EA13 Sweeper and 'HV Lock'
« on: June 13, 2016, 10:15:47 pm »
A few new menu options have appeared in V13, which don't seem to have any help - odd, as they are in the 'Helpers' bit of the 'Layout' ribbon.
  • What does 'sweeper' do (and please don't say 'sweep'
  • And 'HV Lock'

I'm trying to figure out how to create an EA chart (bar chart etc) using the EA API, and there doesn't seem to be any obvious support for it.
Nor can I even find where in the EA database the definition of the example charts is saved.
Does anyone have any experience of creating charts this way ?

General Board / Archimate2 Metatypes and SQL
« on: March 24, 2016, 07:45:00 pm »
My customer is using Archimate2, and has a large model, developed over many EA releases.
They are now seeing some pain from how Archimate_Location elements are stored, as their model has saved them in two ways, and we don't know why.
Looking in the database, some Archimate_Location elements have Object_type="Class", some have Object_type="Location".
Has something changed across EA releases? This seems possible because when opening the properties of a 'Class' style element, then closing it (without changing any data), it got saved back in the database with object_type='Location', and a new 'modified date'.
This would normally not be a problem, because the UI makes everything look OK, but our model-quality analyser is flagging errors because we have the same 'thing' stored in two ways.
Has anyone seen this before?
Also, when I run a query to find these things in a test repository with a few brand-new Archimate_Location instances, the SQL seems to vary the results I get: for some queries, the object_type='Class' , for others object_type='location'. Sometimes both IN THE SAME QUERY. The Object_Type just seems to randomly say 'Class' or 'Location'.

Automation Interface, Add-Ins and Tools / Another error using SQL
« on: March 14, 2016, 10:24:25 pm »
Trying to use the the SQL:
" Select, Object_Type, Stereotype, COUNT(*) from T_OBJECT  "
 group by, Object_Type, Stereotype "
(which finds duplicate elements in a model)
  • Works fine in the model search with EAP file
  • Works fine called via Repository.SQLQuery to EAP file
  • Works fine in a model search with mySQL
  • Doesn't work with Repository.SQLQuery in mySQL
..and the error is the one so often mentioned in this forum:
Error: Code = 0x0 Source Line: 0; Char : 0, Error Description = (null).

This is an especially nasty error because (a) it's very unhelpful and (b) keeps repeating until the use crashes EA.
Other forum items suggest changing the SQL a bit, but this really is what I need the query to look like. About the only change possible is to change the select(*) to a select(Name), but that doesn't work either.
Does anyone have an ideas either (a) how to get around this or (b) how to get more helpful error message from EA ?

General Board / When is an instance not an Object?
« on: February 20, 2016, 02:04:07 am »
When we create EA Instances of EA things (Class, Use case, Actor etc) then EA creates a new entry in t_object, with Object_Type = "Object", and an instance classifier which points to the relevant 'thing'.
But when I create an instance of a 'DeploymentSpecification', I get the correct instance classifier pointer, but the instance 'thing' has an object_type="DeploymentSpecification", not "Object".
Are there other EA meta-types which create instances which aren't "Objects"?
Seems like the only way to distinguish between an instance and a 'thing' is the presence of the instance classifier, but that doesn't seem quite right....
If this a bug?
Or a feature?

Suggestions and Requests / Improve speed of forum searches
« on: November 07, 2012, 02:50:20 am »
Could Sparx possibly buy a bigger server for the forum searches? Mine usually specify 'all time periods' and each search seems to take several minutes to run - not a great user experience, but one which should be relatively easy to fix.

General Board / Schema Composer - Editing Restrictions
« on: June 22, 2015, 11:41:07 pm »
I'm a bit of a newbie to the XML Schema Composer, but think I get the basic function - which is clear and usable.
I'm now trying some of the value-add which I would expect to be doing as I generate a schema from a class model, by adding and editing Restrictions - the multiplicity/optionality of an attribute.
I right-click the attribute to see the restriction, but I can't edit it. Is there something simple I'm doing wrong?
(EA V12 build 1215, Ultimate Edition)

General Board / EA and Use Case 2.0 'Slices'
« on: April 23, 2013, 06:27:40 pm »
I'm just starting to use the UseCase -> TestCase generation feature, as it makes EA REALLY useful for those who are using the new "Use Case 2.0" idea of 'Use Case Slices'. (see Jacobsen et. al white paper at

What I noticed is that the test case scenarios which get generated aren't all the possible paths through the use case, as described in Jacobsen's paper.

There seems to be one per Alternate, but you don't handle combinations of alternates.
So, if my Basic Path has alternates A1 and A2, I get test case scenarios for :
- BP
- BP, including A1
- BP, including A2,
- the A1 fragment
- the A2 fragment

but, I don't get BP including A1 and A2 (assuming that the logic of the use case allows the user to follow both paths in the same run of the use case).
I totally get why this would be: it's computationally nasty to cope with combinations of multiple alternates, and the user can quite easily construct these complex cases from the generated fragments, but I wanted to make sure that this was a deliberate omission. If so, then I'll explain to my users how they can make the complex cases themselves.

@Sparx - is this how it is ?

Also, it would be good, at some point, for EA to explicitly support the Use Case Slice idea, as it looks like you've done all the hard work, and it just needs a skin and a few graphics changes. UC2.0 looks a really practical way to use 'use case' ideas in the Agile world, so is likely to get more popular over time.

..and no, I don't want to write ANOTHER extension....:-)

General Board / Additive Attributes or TVs
« on: October 20, 2014, 09:51:54 pm »
Has anyone got a way to make some EA Attributes additive? For example, if I give each Use Case a 'points' (tagged) value, I'd like to give the 'Work Package' element which implements those use cases another Tagged Value, which is the simple sum of all the points values of the Use Cases it realizes?
A bit of a fiddle to setup, but seems really useful: then EA could make sure that the derived value is always up-to-date, and it become easily printable. Should be good for lots of other values, especially in the SysML world, where you're dealing with numeric values quite a lot.
Probably best done with a Script, but I can't quite see what the trigger would be? Has anyone done this already ?

General Board / Attributes of a Scenario
« on: July 23, 2014, 02:30:27 am »
I'd like to add some attributes to Use Case Scenarios, mainly 'Story Points' to help with estimating and managing my project.
Does anyone know a work-around for this, because sadly, there's no way to attach anything to a Scenario - tagged value, or connector, 'cos it's not a proper element. But is there any way to make it look a bit more like one?
Current - rather poor - work-around is to run a little script to create new <<scenario>>Use Case elements from the proper use case, and attach attributes to this, but it's clunky, and I have to re-run it each time I create a new Scenario.

General Board / EA11 Themes - how to edit ?
« on: March 06, 2014, 10:28:05 pm »
These Diagram Themes in EA11 really seem to make diagrams more readable: my BPMN process flows suddenly look much more interesting, and now I want to edit the default colours from the built-in Theme. The default 'EA11' theme seems to choose yellow for activities, green for Gateways etc, but where do I change and save these colours? I get that the built-in ones aren't editable, but when I save theirs with a new name,  it's not clear how to associate element types with fill colours.

General Board / Transforming into locked packages
« on: March 10, 2014, 09:29:30 pm »
Apologies if I'm doing this wrong, but my customer has just asked me what I thought was a silly question.
They have a locked part of their model, for master versions of stuff only.
A user is trying to create a Communication diagram from a Sequence Diagram in this locked area, and to save it in their own area, for which they have permissions.
So, they choose Tools | Model Transformations, and select Communication.
EA correctly doesn't let them choose the locked, master package to put the resulting diagram, so they choose their own package.
...but when they do the transformation, the resulting diagram is created, but in the Locked area!!
Are they doing this wrong? Is there some other option? This seems to break the EA security model, so they are naturally concerned.

General Board / Display of Connector types and Stereotypes
« on: May 24, 2013, 12:12:35 am »
I'm hoping there's a simple explanation for this, but....
I've been noticin several oddities in the EA diagrammer, when displaying the Types and Stereotypes of connectors.
  • Link two elements with a <<trace>>Dependency, and the connector displays the <<trace>> stereotype text on the link. OK
  • Link two elements with a <<important>>Realization, and the <<important>> doesn't appear. Yes, I have got the diagram set to display connector stereotypes.
  • Double-click a <<trace>>Dependency (or <<trace>>Abstraction connector in EA10 - the one seems to have replaced the other as the default behaviour) and the properties window is called 'Trace Properties' - the name of the Stereotype, not the connector type
  • Double click <<important>>Realization, and the properties window is called 'Realization Properties': the Type not the Stereotype
  • Information Flow connectors get a stereotype of <<flow>> in the diagram, even though in the Properties page, they don't have one
  • ..but the properties is called 'Information Flow Properties'
  • Pick a stereotype for the Information flow (e.g ItemFlow), and the properties window uses the name of the stereotype...(Item Flow Properties)
  • ..except if I choose a 'RequirementRelated' stereotype (?), which means I just get the regular 'Information Flow Properties' title.
Now I'm not turning into an EA Diagram Geek, but people do ask me about this stuff occasionally, especially EA newbies, and I just can't create a simple explanation for what I see.

Bug? Feature? User Error?

General Board / BPMN 2.0 Simulation of sub-processes
« on: August 13, 2013, 01:52:40 am »
I'm trying to figure out how to simulate a BPMN sub-process.
For simple sub-processes, which just have one end event, everything is fine. The SIM dives into the sub-process, then comes back up again, and carries on. Fine.

Now I want my sub-process to end on one of two end events: lets call them 'OK' and 'Not OK'.
The flow out of the parent process clearly depends on whether the result of the sub-process was 'OK' or 'Not OK'. Somehow I need to hook these end Events to the activities which follow them.
Does anyone know how to get this to work with the simulator? It's a common use cases,  as it follows BPMN 'best practice' from lots of the BPMN books, and it would be great to show this all working dynamically in EA.

General Board / EAP files getting larger
« on: July 15, 2013, 11:35:00 pm »
I've had a couple of customers recently who have created unexpectedly large EAP models. In one case, a few newbie BAs created a 200+Mb model in just a few weeks, but <5,000 rows in t_object.
Compacting the model took it down to 125 Mb, but that still seemed a lot. Finally, I exported all the content to XMI, and all the ref data, then re-imported everything. Result: a 12Mb model, which is about what I was expecting.
  • My another customer is not quite so extreme, but they see the same effect when exporting/re-importing: model size is reduced by 80%.
  • Just tried the same thing on what I thought was a well looked-after model of my own: this time just a 50% reduction in size!
Question is, what's happening here? Have  I missed-out in exporting some data, such that this isn't a fair comparison? Ticked all the boxes in the Ref data, and the whole model went into the XMI.
Normally I recommend that model manager runs the Project Integrity check once a week, and compacts their EAP, and that's enough to maintain a good model. But is there something else we should be doing?

Why am I bothered about file size? Just makes everything take longer, especially when models are being moved about, versioned, backed-up etc.

Also, I can't make any of these models available for anyone else to experiment with.

Pages: 1 [2] 3 4