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
1
General Board / Re: attribute range and unit; best practice?
« on: November 21, 2017, 11:13:24 pm »
I suspect that the best practice in UML would be to not display this information unless it was critical to the design concept, in which case Comments (Notes) would be a good way to display this information in a diagram.  See "the elements of UML style", simplicity guidelines: show only what you have to show and Prefer notes over OCL to indicate contstraints.

From memory: if must you can show the element property 'Constraints' page as a note in the diagram I think.  The user manual explains how to do it.

2
Thanks you Geert this is wonderful, I'll try this during Xmas holidays!

3
General Board / Re: How does EA treats Note, Comments and Text box entries?
« on: November 15, 2017, 12:14:40 pm »
Thank you qwerty it's clear!  Basically just pick one and stick with it.

4
General Board / How does EA treats Note, Comments and Text box entries?
« on: November 14, 2017, 12:04:32 pm »
I've searched the user guide, this forum and googled a bit but could not find a description of how are they different in EA and how they are handled differently by EA.  So far seems like Comments and Notes are the same thing although they have different icons on toolbox; and text box cannot be linked to an element and cannot set line color for the box itself. 

For now my goal is to make an informed decision about which I should use for which purpose.  Does anyone have any info on what is the semantic difference and how EA handles each (including reporting, export etc.) thanks!

5
Geert - thanks.  Spent another few hours today trying to clean and making progress with the regex to get rid of junk.  I'll post when I have something that works reliably.
I still have problems with CSV export handling of new lines in 'Memo' fields, I guess it's an intrinsic limitation of the tool as you say, thanks for pointing it out. 

I'll study up on SQL queries when I have time - should be better than the CSV export utility.

6
I am working on a large project, and I need to copy/paste text from several sources into element notes and tagged values.  Many sources are MS Word and Powerpoint documents.  This text contains a lot of formatting characters; I tried to clean them out by pasting into MS Notepad before pasting onto EA; using Excel CLEAN function, and using Notepad++ to find and replace any weird characters.  Still a lot of junk got thru.  The Sparx spell checker ignores these characters.  Main issue for now is that CSV export displays wrong then imported to MS excel: notes and tagged values fields go to the wrong cells and screw up everything else.  I need top quality CSV export because I use it to make tables/slides for presentations to management; also this is a large long term project.
(I export from EA as tab delimited as that is a character that just should not appear in any of the text.)

My preferred cleanup strategy would be to:
1. Export the whole model to XMI
2. Open the XMI with Notepad++
3. Use the 'search and replace' with regex to replace all this junk with empty strings. 
4. Save the cleaned XMI and import into EA in a new project and delete the old project

I googled around but cannot find the regex expression that will safely leave all XMI that EA needs safely untouched, but take out all the junk characters from notes and tagged values.  I tried several expressions but they either leave several junk characters in or they destroy some of the tabs and line feeds that were legitimate.  Example (hope this displays correctly):
Quote
  –  â
and many more

If anyone can provide to me I would be very grateful!  Thanks!!

p.s. Google also brought up an approach to use unix tr command as: tr -cd '\11\12\15\40-\176' < file-with-binary-chars > clean-file.  Anyone has an idea if this would be a better strategy?

7
Hello, here's a few ideas:

Data modeling:
1. Add true crow's feet notation, where keys are underlined instead of having the key icon as stereotype (following the bar across relationships helps following identifiers)

2. If it's not patented, how about adding automatic highlighting of the keys when you hover mouse over a relationship... this is something that Mysqlworkbench does and I find it useful to plan queries and joins

3. You can drag attributes (non-keys) from a table (or Class) to another in Project Explorer, but not in the diagram... it would be useful to be able to move attributes around in diagram as well (like ERWin etc.)

4. Editing attribute names in the diagram by click+F2 or click + slight move cursor is buggy - sometimes changes are not saved

5. Add ability to reverse engineer SQL DDL code (from DDL file to ERD diagram, instead of from database connection to ERD diagram)... useful when database connection is not available and to share schemas via DDL.  MySQLWorkbench and ERWin can do this

6. No CSV export for attributes is a major feature gap... I have read up on EA limitation to CSV export classifiers only and not their attributes, and I understand the difficulty in implementing this across all UML classifiers, but a table-attribute specific script should be made available by Sparx.  It's very valuable for quick editing-fine tuning-reporting when under time pressure (ERWin is excellent in this respect)

Other:
7. If I add a nesting relationship to a package in package diagram, it is not reflected in project explorer... in project explorer the nested package is NOT moved to inside the parent package.  Shouldn't it be?

8. In sequence diagram there should be a checkbox or setting to have default every activation is new... so when using message numbering, the numbering is 1,2,3,4 by default not 1.1, 1.2, 1.3.  It's very time consuming to right click on each message and set new activation for all messages.  I think activations should be new by default

9. When using search function (CTRL-F) several times in a row, the last text typed in the search bar is still there and unselected after pressing CTRL-F again.  So I have to move the mouse to the text bar, delete the text, then type your new search item.  It would be better if the text in search bar was highlighted-selected after each CTRL-F so if I type a new text it will be immediately overwritten; and if I want to keep or add to the old text I can just click on the bar to deselect it.  Like search function works in WIndows.

Pls PM me if any of these suggestions are implemented in future versions.  Thank you!


8
This is really a UML question, not a tool question.  I don't think it makes sense in UML to have multiple instances of the same classifier in a diagram, and if absolutely-positively need to, the correct way to do it is with a generalization (duplicate element IS-A original element).  EA implements the UML standard correctly by not allowing 2 of same classifiers in a diagram. 

Since version 12 or 13 there is something called virtual connector ends that sortof kinda does this in a bit weird and Sparxian way.

See http://sparxsystems.com/enterprise_architect_user_guide/13.5/modeling_tools/virtual_routing.html for more info.

Geert

9
Thank you Geert and Maxym.  Yes I agree with your posts - although to be fair Sparx EA is not marketed as a data modeling tool, the ERD is a useful add-on but it will never be at the same level as dedicated data modeling tools.  Same for other EA add-ons (BPMN comes to mind), I think it's great that Sparx provides some of them for free and they are very valuable in a pinch but they were clearly not meant to replace the dedicated tools.  Another thing is that I think it's perfectly appropriate to model data and design databases in 100% UML and in this case Sparx EA would do a fine job.

I'll submit the feature recommendations (sans-DDL reverse engineering), we'll see what happens.

10
Great feedback Sunshine thanks!
Yes reverse engineering DB works fine for me too - but I'm referring to reverse engineering a DDL text file (CREATE statements...).  Open the SQL file and the tables+attributes+associations+diagram are generated for you
I use CSV export of tables and attributes to go over action items for each table or attribute (check semantics with someone, validate domain, etc.) , sometimes several times a day.  I don't really do this for relationships.  I edit the schema on Project Explorer.  Html or rtf document generation is an OK workaround but not as clean and fast as having an Excel file.

Anyone else?

11
I am thinking about submitting a feature request, but first I would like to sound it off here, since I'm still learning about EA (previous ERWin user)

1. What do you think about adding a true crow's feet notation, where keys are underlined instead of having the key icon as stereotype
1.1 Also if it's not patented, how about adding automatic highlighting of the keys when you hover mouse over a relationship... this is something that Mysqlworkbench does and I find it useful to plan queries and joins
2. You can drag attributes (non-keys) from a table (or Class) to another in Project Explorer, but not in the diagram... it would be useful to be able to move attributes around in diagram as well
2.1 Editing attributes in the diagram by double clicking is buggy too... changes are not saved
3. EA (at least Professional) cannot reverse engineer SQL DDL code... I find it useful to move models around and update schemas.  MySQLWorkbench and ERWin can do this
4. No CSV export for attributes is a major feature gap... I have read up on EA limitation to CSV export classifiers only and not their attributes, and I understand the difficulty in implementing this across all UML classifiers, but a table-attribute specific script should be made available by Sparx.  It's very valuable for quick editing-fine tuning-reporting when under time pressure (ERWin is excellent in this respect)

Any feedback or ideas?  I'll submit feature request in a few days


12
General Board / Re: What is Sparx and for what purposes it should be used
« on: September 28, 2017, 12:01:35 pm »
Hello Alex,

the most widely accepted definition of Enterprise Architecture is provided by the Open Group via the TOGAF standard.  TOGAF is 100% free and all of it is available at
http://pubs.opengroup.org/architecture/togaf9-doc/arch/

UML is widely accepted as one of the better ways to generate the artifacts described in TOGAF. 

Sparx provides a TOGAF plug-in module with all artifacts templates ready for use.

It doesn't get any easier or better than this.  All the hard work is done for you... it's just a matter of rolling up your sleeves and getting to work.

13
Suggestions and Requests / Re: In-house training in HK
« on: September 09, 2017, 01:21:37 pm »
I may be able to help depending on the kind of training needed.  I'm based in Taipei but will be presenting at the Open Group - Enterprise Architecture Digital Transformation conference in Shenzhen on Sept. 21-22 (using EA! Yay!), so we may be able to meet there on en route (flying via HK).  I speak Mandarin but not Cantonese. Send me a personal message if interested. 

14
Uml Process / Re: Importing XMI from Papyrus to EA
« on: August 23, 2017, 10:19:45 am »
EA version 13.0.1310; importing a model from Papyrus Neon (version 2.0.2).  Couldn't do it without the kludge: I was getting an EA error message on the lines of "unknown XMI version".

15
Uml Process / [Solved]: Importing XMI from Papyrus to EA
« on: August 23, 2017, 01:48:00 am »
Hello I resurrected this old topic because I was just now in this situation and could not find a solution online, but this kludge worked for me: I manually change the Papyrus *.uml file on the second line as follows:
From "uml:Model xmi:version="20131001"  to uml:Model xmi:version="2.1"

Then all model elements imported into EA correctly (even notes loaded correctly).  Obviously no diagrams since the *.di file is a Papyrus specific thing, but everything else seems fine!

I don't know why all my Papyrus UML files have xmi:version="20131001" in my *.uml file header.  I'll ask on the Eclipse Papyrus forum when I have time.

Pages: [1] 2 3