Messages

1

Suggestions and Requests / Re: Hiding the stereotype on an artifact

« on: June 26, 2010, 10:48:19 am »

Hi Geert,

You know I had to look that up, and it's pretty explicit:

In UML 2.0, a Component is notated by a classifier symbol that no longer has two protruding rectangles. These were cumbersome to draw and did not scale well in all circumstances. Also, they interfered with any interface symbols on the edge of the Component. Instead, a «component» keyword notation is used in UML 2.0. Optionally, a component icon that is similar to the UML 1.4 icon can still be used in the upper right-hand corner of the component symbol. For backward
compatibility reasons, the UML 1.4 notation with protruding rectangles can still be used.

But strangely, I can't remember that bit about the keyword being mandatory in "previous" 2.x's???

Anyway, another day another thing lerrnned.

cheers
bruce

(Quote from Super 10-05-05 section 8.3.1 p159 "Changes from previous UML")

2

Suggestions and Requests / Sub-pane synchronization

« on: July 03, 2010, 12:13:32 pm »

This on is a bit more complex.

Some panes synchronize with other visible sub-panes, some don't.  It would be   r e a l l y  nice if there was some consistency here or at least some indication in the sub-panes as to which *&$%^&^ element/feature it was currently looking at.

For example, if we have the project browser, the Maintenance/Tasks pans, the notes pane and the tagged values pane visible, then selecting an element or feature in the project browser will immediately update the other panes.  This works for both elements and features.

However, if I use the Element Browser instead of the Project Browser
the Notes Pane and Tagged Values pane don't change when a feature is selected.
BUT!  the Maintenance/Task Pane does.

Does this irk others.  If so I'll report it, but do you reckon its a bug or a suggestion.

C-C-C grrrrrr
bruce

3

Suggestions and Requests / A few other useabilities

« on: July 03, 2010, 11:55:59 am »

Refresh buttons

• on the element browser.  Changes to a feature name or signature are not immediately reflected in the browser.  This should be a simple enhancement.
  
• on the traceability pane.  This appears to suffer the same problem. However, see the DUMB BUG below.
  

DUMB BUG

• Traceability white outs. Changing the name of a related element (say a supertype) using the "View Related Element Properties" results in the traceability panel suffering from a complete white out.
  

All reported, please vote early and often.
bruce

4

Suggestions and Requests / Re: Viewing more things 'side by side'

« on: January 25, 2010, 09:26:07 pm »

Me too! (Even having a classifier attributes, methods and the classifier popup "activatable" at the same time!

bruce

5

Suggestions and Requests / Re: Create hyperlinks to elements within element n

« on: December 30, 2009, 06:55:42 pm »

And further, this would open up a channel that could either

• confuse the reader unless you stringently control the contents of your links

or

• add a truckload of extra work.

Not voting against it, just being   c a r e f u l !

b

6

Suggestions and Requests / Re: Aggregation in requirement diagram

« on: October 09, 2009, 11:31:36 am »

+1 for what they said.  It still confuses the heck out of me and I still have to manually reverse the association every time.  By "it" I mean the UML aggregation standard and the EA options and the congruence of the two.

bruce

7

Suggestions and Requests / Re: Move Interfaces

« on: January 25, 2010, 09:23:02 pm »

My fading memory says I wrote something about this some years ago.  IIRC, they (exposed, dependant and realized) are not "exactly the same thing" with a fairly good reason.

Maybe things have changed, maybe not.  Sorry, I can't even come up with a search topic. It might have something to do with components though.

b

8

Suggestions and Requests / Re: Use Case Metrics ineffective

« on: January 02, 2010, 10:52:47 pm »

Interesting, but you gave no reference to Kirsten Ribu's extension of Karner's original proposal.  I believe that the EA calcs are based on this work, which as she says is only useful for estimating the size of a project in the early stages.
Certainly as development of a model proceeds, it is possible to go back to the more formal (although still fraughful) function point method.  I look forward to re-reading your paper once it is translated (as my French is possibly worse than my coding ability    )
cheers
bruce

9

Suggestions and Requests / Re: Add more Use Case example templates

« on: December 30, 2009, 06:44:46 pm »

Actually marcelloh, actors are not part of a use case.  The use case exists whether or not the actors do.

All a use case knows about the external environment is how to respond to an initiating stimulus and how to follow a structured "conversation" with whatever provided that stimulus.

An actor is a role, it is also (if required) a classifier.  If so, then the use case may be able to interrogate its environment to find out something about this role.

Use cases (IMO) are NOT actor dependent.  Then again, I have seen people trying to build systems that do try to implement this dependency.  

By way of an example, a constraint on a use case like "only sentient beings from the planet Xxurrkk are allowed to use this use case"  may temp a young inexperienced modeller/designer/builder into trying to assume that the use case "automatically" knows the role being played by the external influence that has perturbed its wellbeing.  I seriously doubt it.  The system which may know something about its' environment may be able to provide such information, but the use case ... no.

To summarise, from the point of view of the use case, the "actor" is just a ??something?? that has initiated a perturbation in its state.  Nothing more.

hth
bruce

10

Suggestions and Requests / Re: Half baked idea re reallly big classes

« on: October 29, 2009, 09:44:15 pm »

I thought you were an Aussie and not a Kiwi

 BAH!@#$& I'm a tester remember, all classes are bug classes!

11

Suggestions and Requests / Half baked idea re reallly big classes

« on: October 28, 2009, 08:24:02 pm »

This is just me thinking again.  I don't even know if its possible, maybe a Sparxian can comment.

You know those really bug classes, like MyFancyControlThatInheritsEverythingAndThenAdds903OddNewAttributes, it would be kinda nice to be able to set its height in a diagram and have scrollbars to skim up and down in it.

I am going to emphasize here, <emphasis> I only mean on the screen</emphasis>.When the diagram is printed it should be layed out according to the set visibilties.

Currently when I add "that class" to a diagram or change the visibilities, it expands the diagram to about 30gazillion pages and disturbs all the associations .  

Comments?

bruce

12

Suggestions and Requests / Re: List of tabs, switch between them (back functi

« on: November 13, 2009, 09:39:05 pm »

formally defined voting mechanism

Just searched both the super and infra structures.  Couldn't find one either  

But yes +1 here too.

13

Suggestions and Requests / Re: GUID integrity

« on: November 28, 2009, 07:03:41 pm »

but I think that is as likely as me winning the lottery two weeks in a row.

Actually, theoretically, if the 122 "random" digits in the GUID are truly random then I think you'd have to win a lottery with 50,000 tickets about 8 times in a row.

However, the 122 digits are not truly random and if they use an algorithm that includes the MAC address of the local computer and a single timestamp, then the likelihood of the collision is a lot higher.

But, then again, my problem is winning the lottery one time in a row.  

bruce

14

Suggestions and Requests / Re: Better rendering of crossing connectors

« on: November 27, 2009, 12:01:02 am »

Lines, aka edges never join. Joins can only occur at vertices aka elements.  While I agree that crossing lines are a visual problem I don't really think extra adornments will contribute to alleviating the situation.

As Thelonius says, if the diagram is problematic simplify it.

hth
bruce

15

Suggestions and Requests / Re: Alter Table (SQL DDL)

« on: November 24, 2009, 03:42:31 pm »

Getting back to the revivalist post.  

I've been going through this for years, not only in EA but in other RDBMS modelling tools including whats-its-name from CA.

The restrictions on ALTER vary so much between database types and even within a database type depending on whether or not a relationship is populated or not are such that the useability of this is questionable.

The disruption of indices, side effects on views etc etc has meant that I have often found that reloading the data into a new schema is a lot easier than trying to use and fix what ALTER has done.

Not actively voting against it, JM20cW
bruce