Heather Wallace

[Heather Wallace]

I’m very sad 🙁

[Heather Wallace]

Flattery will get you everywhere! But this is more EA’s way of doing things than UML.

I drew an association, then a separate information flow. Then I used right-click > Advanced > Information Flows Realized…. to associate the two. Its much easier to manage than lots of separate lines. Also works for control flows, very useful.

I didn’t realize the xmi hadn’t uploaded. I’ve zipped it this time.

[Heather Wallace]

A noddy example

Heather

[Heather Wallace]

That’s great news,

I like playing the role of clue finder – especially when it results in a fix!

Heather

[Heather Wallace]

I think I’ve found the source of this, though it still doesn’t make much sense to me.

1a. If I output the children of an element using the sub-elements menu each child is correctly assigned the next heading level.
1b. If I add a title to the Children instruction, each child is correctly pushed another level down

2a. If I output the children of an element as a RelationshipElement (with no “title”), each child is assigned the heading level 2 levels down – as in 1b

For the current document I have replaced the 2a approach with 1a. But for some documents I need to output children of specific types, so must use the 2a approach. As for my previous posting I have been working in draft mode.

Unfortunately my examples have confidential content. I would need to recreate something representative. Can you let me know if these clues are sufficient, or if I should devote effort to recreating the problem?

Thanks,
Heather

[Heather Wallace]

I’ve set up a bunch of styles in my eaDocX Word 2007 document called CLAS_Heading_1, CLAS_Heading_2 …(see attached document – for styles only). Font sizes are:
CLAS_Heading_1 – 16pt
CLAS_Heading_2 – 14pt
CLAS_Heading_3 – 12pt
CLAS_Heading_4 – 11pt
CLAS_Heading_5 – 10pt
CLAS_Heading_6 – 10pt

In EA10. I use Tools, Options and Settings, inline text settings to select CLAS_Heading_1 for Heading 1 style, CLAS_Heading_2 for Heading 2 style etc..

I generate my document using the draft option (because of the SystemAccessViolationException bug I reported in June #979). Perhaps the sizes would have been corrected in the final pass that doesn’t occur working in draft mode?

Cheers,
Heather

[Heather Wallace]

Another thought for you / clue for eaDocX
(extra info added)

On my current task I observe the following:
1 Tables that are simply output as a table of relationship attributes are complying with their column width definitions and text alignment in cell.
2 Tables that are defined in the same way, but are navigated to via RelationshipElement (to stereotype defined as table) are automatically resizing columns and using centred cell alignment.
Example structure:
Package
– FeatureGrouping (Object Type = Class) IS PARENT OF SystemFunction (Object Type = Activity)
– FeatureGrouping (Object Type = Class) IS PARENT OF Capability (Object Type = Class)
– SystemFunction (Object Type = Activity)IS PARENT OF SysUseCase (ObjectType = UseCase)

My FeatureGrouping profile outputs:
1. A table of the Name and Description of its SystemFunction children (WELL BEHAVED TABLE)
2. A relationship element (SystemFunction), which is defined to output inline only a table of the SysUseCase objects that have a trace relationship (mostly those in the tree but also some others) (BADLY BEHAVED TABLE)
3. A relationship element (Capoability), which is defined to output as a table (i.e. not inline). (WELL BEHAVED TABLE)

I will no doubt get further clues over the next few days, but thought this worth mentioning,

Heather

• This reply was modified 11 years, 2 months ago by Heather Wallace. Reason: Added exta info and clarification

[Heather Wallace]

Ah yes,

I also find the Word table style is only partly applied, for example my offsets within the cells are not applied. I’d forgotten about “manual multi-select” of tables, I did use that at some time in the past. The trouble is that now we typically have circa 200 tables per document. So it may be appropriate for a final baseline publication, but wouldn’t want to do it every time.
Good luck with your experiments.

Heather

[Heather Wallace]

The column widths has been driving me to distraction too. There seems to be no rational on when Word tries to help. I have tables constructed in the same way in eaDocX on different stereotypes in the same document, same styles and table styles, same column widths. On one stereotype (a custom stereotype of Class) the widths are always output consistently and the requested cell alignment is observed (top), on other stereotypes (custom stereotypes of Activity) the column widths change for the contents and the cell text is vertically centred. Makes no sense to me at all.

I’m learning to ignore the detrimental impact it has on the documents. There are just too many tables to change them all individually, and Word doesn’t seem to offer the option of selecting all tables or all instances of a table style. If anyone knows how to multi-select tables in Word 2007 I’ll be forever in your debt!

Regarding table width, on my recent documents I’ve standardised on a total of 575 pixels across all columns, using a table style that centres the whole table on the page. Thus far all my tables are outputting at the same width (15.21 cm) in perfect alignment. It’s just slighly less than my margins, but looks fine.

[Heather Wallace]

Fingers-crossed for a forum-watcher in the know!

[Heather Wallace]

OK. I’ll look at that when I get the next version installed by IT (hopefully today). It will provide a solution of sorts, but ideally we prefer to have the same set of subheadings for each element, rather than some that are standardise and others that are element-specific (specific diagram name)

Heather

[Heather Wallace]

I’ve always got around this problem by outputting the conficting sections as separate eaDocX documents and manually fitting them together in MS Word. Not ideal. I believe this is what the eaDocX element report is designed for (I plan to try it out shortly). I think the profile for a stereotype within an element report can be different from the profile that applies to the rest of the document.

HTH (keen to see if a moderator agrees!)

Heather

[Heather Wallace]

Oh dear,

This fudging, as you put it, puts EA way behind some competitors in the way it handles data flows. I was hoping we could simply address this via the output, but it seems that a more complex solution will be required.

Thanks for giving it some thought.

Heather

[Heather Wallace]

Hi,
I thought it would be useful to mock up the anticipated output. As I did so I noticed an error in use of activations in the sequence diagram, so I have made two minor changes. Please find attached the replacement XMI and a Word file with the target output for each of the packages.

Heather

[Heather Wallace]

Hi,
It is not possible to show all the relationships on one diagram, so I’ve built and exported a mini package containing a sequence diagram, a relationships diagram and all the involved elements. The note in yellow on the relationship diagram explains what I am hoping to achieve in terms of hops.

The document output would be as follows:

Stating at (e.g.) the Sensor Infrastructure Functions Package
1. [Package Name]
1.1 SEFunction: [SEFunction Name]
[SE function description]
Logical data elements consumed*:(list of names of data elements consumed)
Logical data elements produced*:(list of names of data elements produced)

repeat for other SE Functions, as 1.2, 1.3 etc
* See yellow note on relationships diagram for expected hop sequence.

I hope this is clearer,

Heather

• This reply was modified 11 years, 4 months ago by Heather Wallace. Reason: Formatting when uploaded to site looked confusing

[Author]

Posts