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 - Richard Freggi

Pages: [1] 2 3 4

I am curious what would total participation look like in Crow's Feet (I.E.) notation?  Or what would a SQL DDL code for total participation look like?

If it's 1:1 as suggested in previous post, it's certainly possible to do with EA but 1:1 is usually a sign of a problem in the data model... probably semantic mistakes or treating an attribute as if it was an entity, or using 2 entities where 1 would be enough.  I'm curious to know... thanks!

General Board / Re: Document a whole Solution
« on: March 02, 2018, 11:39:25 am »
Agreed!  TOGAF content metamodel really good, Zachman framework is similar and probably a little easier to use.

Geert, if you use UML for everything you will have a fully integrated model, with consistent elements and artifacts from contextual down to physical and out of context... easier, more efficient and effective than mixing multiple notations.  IMHO UML works even better at contextual-conceptual-logical than at physical level (UML was originally developed to talk to people, and a failed attempt to shoehorn it into a programming language has damaged it quite a lot...)

Willeygi, Sparx has plug-ins for TOGAF and Zachman; you can find full TOGAF documentation online at the Open Group website and there's plenty of Zachman info and examples in the web. 

As for the model organisation, I prefer abstraction layer before domain

The TOGAF content meta-model is by far the best part of TOGAF and is very useful for structuring "content".

Uml Process / Re: Modeling complex component associations and interfaces
« on: February 13, 2018, 02:26:05 am »
Hello Miktaylor,

a good UML diagram shown only the minimum required to convey clearly and unambiguously the concept it wants to express.  For example, what is a functional architecture, who needs it and what for?

You need to ask yourself exactly what is the purpose of your diagram, who is the audience and what they will get out of your diagram.  This will help you choose the right diagram for your idea.

Experienced modelers break down large complex diagrams into several simple ones, each with a specific and clear message.  A very short book "The elements of UML 2.0 stlye" is a great reference.

Suggestions and Requests / Re: Relationship between/to Attributes
« on: January 30, 2018, 05:43:14 am »
Thank you Paolo and qwerty for the info!

Automation Interface, Add-Ins and Tools / Re: Project Template in MDG
« on: January 28, 2018, 07:25:29 pm »
If the template refers to the structure of the model, with maybe some strawman diagrams, maybe a UML profile would be a good way to do it?  Then the profile could be included in the MDG.

Suggestions and Requests / Re: Relationship between/to Attributes
« on: January 26, 2018, 07:44:21 am »
Thank you for your insightful post Simon.
For my own education: about this
"UML doesn't define a notation for rendering relationships between the attributes of a class. (Although, it's not incorrect to render it as a Port or Part to show the relationships."

Understood re. diagramming.  In practical terms, my assumption would be that if some element has a real association with something else, then both elements are classes (i.e. the attributes having the relations are in fact classes on their own right).   Would there be a realistic practical situation where such a relationship exists between non-class elements?  In some programming language or some practical application. 

This isn't an EA issue, but the way UML works: Attributes Do Not Have Relationships.
Not quite true. Dependency, some of its specializations and Information Flow all go between Named Elements, which includes Attributes.

On the other hand, UML doesn't define a notation for rendering relationships between the attributes of a class. (Although, it's not incorrect to render it as a Port or Part to show the relationships.)

General Board / Re: Importing ERwin models
« on: January 20, 2018, 06:40:37 pm »
I am maybe not the most qualified person to post because my imports from ERWin (circa 2015... I forgot which ERWin version but it was the current one) did not go well.  I followed the instructions in ERWin to export as Sparx EA compatible XMI; all models imported but there were missing relationships so I had to manually correct the imported model, very time consuming.  Tried several times with different formats, still not good.

For me the safest way is to forward engineer the source model, create a database with it (in mySQL, SQLite, Oracle express etc), then connect Sparx to that database and reverse engineer it.

Good luck!

General Board / Re: XMI Round Trip - Data Loss?  Your help needed!
« on: January 18, 2018, 09:39:39 pm »
qwerty and vkn, many thanks and this helps.

"Create placeholders for missing External References during XMI 1.1/2.1 Import" locks the placeholders (I have 13.10 professional desktop) and I can't unlock them to change colors or names (I can change fonts though).

Anyways that's already a big step forward, many thanks for your info!

General Board / Re: XMI Round Trip - Data Loss?  Your help needed!
« on: January 18, 2018, 03:11:34 am »
Thanks qwerty!  I'll look into EA user guide for more info

If anyone has more pointers on how to make this work correctly please chime in... for example if I want to send the exported XMI to a 3rd party and need them to be aware of the dependencies while maintaining encapsulation of the dependent package. 

For example in StarUML (another UML CASE tool) the tool added a dummy element to the diagram (with an 'M' mark to show it was just a placeholder) so the 3rd party knew there was a dependency and what was the name and type of the element at the other end of the dependency.

General Board / Re: XMI Round Trip - Data Loss?  Your help needed!
« on: January 18, 2018, 02:02:35 am »
I am resuscitating this old thread because I have the same problem with v 13.10 and have not  been able to find a solution by searching. 
My problem is that if I export XMI from a package in a model, all information about external dependencies of that package and its elements are lost.  These external dependencies are critical design parameters so there must be a way to preserve them.  Thanks for any help!!!

Either way - strong request to Sparx team: please keep EA fully UML compliant - no 'dialects' or pidgin UML. 

Profiles are the right way to extend UML to a local 'dialect' if it is beneficial for  users... 

Hello Drew
you are making a very common error - trying to make a flow chart (a functional decomposition that includes some logic) with use cases.
It's not going to work - and it's not going to be useful for your team.
Use cases should only show WHO (actors) want to achieve WHAT goal (use case).  They should show the scope and requirements for your project.  For example 'Going outside' is unlikely to be a goal (unless you want to break out of prison, for example); 'Buy Xmas gifts' is a goal and going outside is how you achieve the goal if you need to get to the shopping mall to achieve it.  Use cases should not say anything about HOW the goals are achieved: sequence diagrams are used to show HOW.

I think your team may want to think about whether you want to use UML as the notation for your functional documentation; for example BPMN (not UML but supported by a free plugin in EA) is a good functional decomposition notation.

If you want to keep using UML, could I recommend some UML books such as 'Use case modeling' and 'Applying use cases' (Addison-Wesley).  Cheers!

Uml Process / Re: Importing XMI from Papyrus to EA
« on: December 18, 2017, 11:38:53 pm »
Hmm interesting!  But no prices no downloads... I'll stick to manually recreating models for now.  But thanks for the info!

Uml Process / Re: Importing XMI from Papyrus to EA
« on: December 14, 2017, 01:16:18 pm »
Arun, if you find a way to import FROM Sparx INTO Papyrus please post!  Urgently needed for me (I use Sparx for work, Papyrus for teaching).  No dice on the Papyrus forum - nobody seems to have done it.

General Board / Re: Linking a Logical data model to database builder model
« on: December 09, 2017, 01:33:34 am »
I don't thins there is / should be an automatic linking of data models at conceptual - logical - physical (schema) level.  Going from one level to another you make a number of decisions about the model structure based on circumstances, your style and habits etc. There are many different correct ways to develop a physical schema from a logical one and vice versa.  Don't know how it could be automated to have changes propagated correctly.
This particular client has all their transformation rules defined.
The add-in transforms the logical model into a physical model following these transformation rules.
Then the DBA's review the proposed structure and overrule certain rules if needed.
The tool marks the database objects that have been overruled in order to not propose a change according to the standard transformation rules next time the logical model is being transformed.

In any case, this save a lot of mindless grunt work and helps with the standardization (naming conventions etc..)


Geert: very impressive and also extra props to you for releasing it as open source!!!

Pages: [1] 2 3 4