Messages

31

Suggestions and Requests / Re: Rotate Diagram Objects

« on: May 30, 2007, 12:52:03 pm »

Second this too.  Would you want the name to rotate with the object (or keep horizontal?).

Reason to rotate the name is it reduces the "footprint" the object takes up and may allow for connectors to fight snuggly (rather than draw over text).

Cheers,
David.

32

Suggestions and Requests / Re: cut/copy/paste in project browser

« on: September 16, 2007, 11:47:52 am »

Multiple project browsers - I agree.

Control drag to copy would be nice to - like Windows explorer (rather than explicit copy/paste)

I've also suggested a favourites / bookmarks / shortcuts in the past for common or frequently accessed packages

Cheers,
David.

33

Suggestions and Requests / Re: Custom Toolbox / profile for common used eleme

« on: August 08, 2007, 08:18:25 pm »

Yes - a favourites browser is better.  Best not to overload usage of the toolbox (almost as bad as overloading a table column for several purposes

Gets another vote from me.

34

Suggestions and Requests / Custom Toolbox / profile for common used elements

« on: August 08, 2007, 02:10:49 pm »

How about being able to create a toolbox to contain key actors, components, nodes etc... often used on diagrams.

To be able to be configured per user.

This saves having to find in a large project hierarchy the right components.  Relocating frequently used elements to a "common" package wouldn't work since different project roles would have different elements (and some might be shared) and would also "break" the logical structure in the model moving things about.

What do you think guys?

Cheers,
David.

35

Suggestions and Requests / RTF generation - Connector source / target element

« on: July 29, 2007, 03:19:30 pm »

I need to be able to list all the connectors under my diagrams in my RTF documents and be able to include the source / target element names for each.

Currently - you can only include source / target element names for a connector, when you list the element at a package level (not within the diagram section)

Sparx - can you please extend the connector under the diagram section in the RTF generator to be able to include source / target element information (to be consistent with the connector setting already under the element section?).

Thanks,
David.

36

Suggestions and Requests / Sparx - how do you prioritise new features?

« on: May 31, 2007, 02:38:07 pm »

Dear Sparx,

Can you please tell me how you prioritise new features for major releases?

Do you solely go by formally logged requests?
Based on discussions with your largest customers?
Do you use much input from the eccentric (ie. me) on these forums?
Competitor features?

It would be good to know what use our posting does here to help Sparx decide on what features to introduce.

Would appreciate your response.  If there is a way we can better structure our suggestions to make it easier to filter / group, then please let us know.

I also suggest you run a poll on this site for features you wish to introduce for a new version (obviously prior to feature freeze) so that us, the users can help you prioritise.

Look forward to your response.

Cheers,
David.

37

Suggestions and Requests / Re: What EA needs for large organizations?

« on: May 31, 2007, 02:42:06 pm »

Chris - have you tried to clear your cookies and re-login?

Sparx need to address model scalability - my 6,000 object SQL Server database model is getting slower.  ie. adding new elements to a diagram can take 1 - 2 seconds.  This may not sound much - but if you are doing that operation hundreds of times per day, it can add up (I am very productive you know!

David.

38

Suggestions and Requests / Re: What EA needs for large organizations?

« on: May 02, 2007, 04:54:56 pm »

My advice:

1. Ditch EAP files.  You need to use a database
2. Create 2 top level root models (in the same DB) - one called "Enterprise Architecture" the other "Projects"
3. When a project adds new artefacts and is signed off - move the artefacts you want to re-use (signficant ones) back into the Enterprise Architecture model.

You should adopt Zachman or another framework to give a project indepedent structure to your models - (ie. for TOGAF an application, data, business, infrastructure architecture views)

In any case (and I think this applies to all modeling tools), you need to have a process for moving in artifacts added for a project back into the enterprise model.  This is something you could automate if you use stereotypes or some way of "flagging" the components you want to move back (ie. use EA automation or SQL directly on the database).

If you have a geographically dispersed site - either use Citrix for the EA client or replicate the database between sites for performance.  If you have remote users - get them a network connection to always connect to the central database (it will be cheaper than people's time synching remote models, occasional stuff ups and frustrations - trust me).

Can you say how other tools achieve this better?  Suggesting to Sparx might be useful here.

Cheers,
David.

39

Suggestions and Requests / Re: XSD Schema generation with choices

« on: October 16, 2006, 03:19:42 pm »

Hi there.

It seems to work for me in EA (apart from the extra root element EA puts in when you generate the XSD back out - which you just delete).  

Which version of EA are you using? (I am using the latest - 6.5.798)

I don't know why your EA collapses into only one complex type.

1. I imported your XSD (your original version at the top).  I got the 2 distinct complex types generated in EA
2. I then generated the following XSD out of EA.

<?xml version="1.0"?>
<xs:schema targetNamespace="http://www.testStuff.com/Handler" xmlns:ce="http://www.testStuff.com/CanonicalEvents" xmlns:hnd="http://www.testStuff.com/InHand" xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:import namespace="http://www.testStuff.com/CanonicalEvents"/>
<xs:element name="ArrayofEvents" type="ArrayofEvents"/>
<xs:complexType name="ArrayofEvents">
<xs:choice>
<xs:element name="BalanceEvent" type="ce:BalanceEvent"/>
<xs:element name="CloseEvent" type="ce:CloseEvent"/>
<xs:element name="OpenEvent" type="ce:OpenEvent"/>
<xs:element name="ReturnEvent" type="ce:ReturnEvent"/>
<xs:element name="TimeoutEvent" type="ce:TimeoutEvent"/>
<xs:element name="VoidEvent" type="ce:VoidEvent"/>
<xs:element name="WarningEvent" type="ce:WarningEvent"/>
</xs:choice>
</xs:complexType>
<xs:element name="InHand"/>
<xs:element name="InHand" type="InHand"/>
<xs:complexType name="InHand">
<xs:sequence>
<xs:element name="inHandDateTime" type="xs:dateTime"/>
<xs:element name="inHandEvents" type="hnd:ArrayofEvents" minOccurs="0" maxOccurs="unbounded"/>
<xs:element name="inHandId" type="xs:unsignedLong"/>
</xs:sequence>
</xs:complexType>
</xs:schema>

Have I missed something?

Cheers,
David.

40

Suggestions and Requests / Re: Suggestion: Merge on XMI import

« on: November 29, 2006, 12:19:06 pm »

Hi Simon,

I realise this could be dangerous (but hey - a modeler's life isn't worth living without a little danger   ...

I like professor's idea of a merge window (you could re-use your diff logic here).  This would essentially display clashing items and allow you to select (say CTRL-select) the items to bring in vs. not.

The practical reason I'm asking this, is we create a "template EAP project" for projects to use - with some re-usable elements.  Problem is users:
a) often move stuff around (so importing just by package is too limited - you might miss something, or duplicate something)
b) Don't find an element - and recreate it - duplicating in the model

When we import (merge) back in the completed artifacts into our central project database - I have to clean things up manually.  Would be great if EA allowed me to avoid importing duplicate elements, or if I decided to import, bring in elements into their original package (even though it isn't the one I'm importing under) and update with the new content.

I know version control would place huge implications on this.  I think what I'm asking is to have a "package-less" import option. - where elements can be imported (refreshed) into whatever packages they currently reside.  After all - in many cases, all the package offers is a convenient container to make model navigation easier.

Hey this is a tough one... but I know Sparx is up to it!  

Cheers,
David.

41

Suggestions and Requests / Suggestion: Merge on XMI import

« on: November 13, 2006, 05:55:19 pm »

This would avoid duplicating elements already moved or relocated in the target (being imported into) model.

Suggest on the XMI import page, some options are added to allow merging by object type, stereotype and whether case in-sensitive merges are performed.

David.

42

Suggestions and Requests / Request: Quick link to support stereotypes

« on: November 28, 2006, 02:33:44 pm »

For example - when using quick link to create a dependency (or any other type of relationship for that matter), use a cascaded menu item to select the stereotype.

The user could still simply click on the dependency menu item as per now to create a simple dependency (without a stereotype), or display a sub-menu with the valid list of dependency stereotypes (and click on that).

Saves having to edit link properties after creating a relationship to set the stereotype.

Cheers,
David.

43

Suggestions and Requests / Re: Suggestion: Layer support in diagrams

« on: February 15, 2007, 03:34:35 pm »

Any thoughts on this one Sparx?

I've had another solution.... more simple and in fact probably more powerful / re-usable in other contexts.

Have a "copy diagram layout" option.  This would re-position elements on the current diagram to where they are located on the diagram you select.

Better still - have a "sync layout" toggle - so that EA will keep the diagram sync'd on changes.

That way, you could have multiple diagrams.  Each showing a different "layer" - with common components in the same place.

PS: Sparx, if you do the above, don't forget swimlane syncrhonisation too!

Cheers,
DAvid.

44

Suggestions and Requests / Suggestion: Layer support in diagrams

« on: November 30, 2006, 05:32:55 pm »

Be able to create layers and assign which layers an element or connector appears on.

This would save having to create multiple diagrams then aligning element positions to be the same on each.

Suggest a user is able to define per diagram layers (arbitrary names / numbers) and you can right click a connector / element and checkbox which layers it will appear in.

Then a right click option on the diagram to toggle layer visibility.

Not sure how this would work on web publishing.  Suggest "show all elements / connectors" would be the default.

What do you think?

Cheers,
David.

45

Suggestions and Requests / Re: Label Hiding, En Masse

« on: December 10, 2006, 06:01:48 pm »

This posting in the general forum using shape script might be your answer...

http://www.sparxsystems.com.au/cgi-bin/yabb/YaBB.pl?board=general;action=display;num=1165490013