Sparx Systems Forum
Enterprise Architect => Suggestions and Requests => Topic started by: fprog26 on April 10, 2006, 09:50:24 am
-
I would like to make a small suggestions would it be possible to add a small vote sections for new features. :)
I think it would be a nice way of asking us ;D,
if a feature is needed, really needed or not at all, instead of searching the forums. ;)
You can add me for:
- Doxygen comment style parsing ;)
- Unlimited character length template parameters, since it gets automatically truncated ???
- Better control over RTF document generations
- Being able to import/export EAP into XML format for scripting purposes.
For what is worth, I've been using Rational Rose since the beginning,
but since I used EA for a while now, I can say that it is definitively
better, easier and the RTF document generation is just awesome! ;D
-
Something like this has been the subject of lively discussion on this forum recently, but for the life of me I cannot think of a search term that will find the thread. [Some people might suggest that's fairly useless information; those people would be correct.]
As to your specific suggestions:- Doxygen style various stuff has been brought up a lot, and may well be on Sparx' radar.
- Several 'things' (i.e. types of strings) get truncated at 255. This is (usually) the result of Access field lengths in the Jet model (anything > 255 needs a memo, and imposes other limitiations on what can be done). We certainly need to keep this in sight, but until a different (better?) default repository is found we're probably stuck with this. Given that the Jet drivers remain part of Win32 this could be a while. [Still, MS seems (sometimes) to be allowing Jet development to gradually decline, so who knows...]
- Sparx continues to improve this capability since about EA 4.5 (or 5.0, I cannot remember). Mostly this is evolutionary rather than revolutionary (i.e. adding or improving features, rather than the sudden addition of the template-driven mode as a whole). IMHO we would do well to make specific suggestions and requests in the hope that these get traction at Sparx, rather then just asking for 'more' or 'better' capability.
- Both XML (as well as XMI) and Scripting are pervasive in EA, and both could continue to be extended. That said, I'm not sure what you mean. This is an area of great interest in my community, so please (please, pretty please) elaborate on what it is you want, and let's see if we can raise some interest in the community and at Sparx.
I too came out of the shadows of (several) other UML and modelling products to work with EA. Quite a revelation overall. Many of my clients - and more recently several of my associates and their clients - have agreed; EA has started to appear all over the place around here.
David
-
BTW, look into the EA User Group. You'll always find the current link as a 'sticky' topic at the top of the Automation forum.
[Yes, I know I could have pasted the link here, but I wanted to get you to look at the other forum, and the link there should stay current.]
The user group is the brainchild of one of the gurus here, and could be a useful place to build up suggestions to a point where they're ready to go to Sparx. Some people have also contributed some nifty stuff for the community.
David
-
Doxygen comments such as: //!< and ///< and /**< */ can easily be patched, just truncate the extra "<".
Unfortunately, Doxygen and EA with HTML generations are the only tools so far which can deals with C, C++, Java and similar languages.
However, on the good side, Doxygen jams on large and very complicated template "code", while EA doesn't show everything that doxygen produce, but at least "works" somehow, but doesn't show "meaningful" associations, which must be manually added. The work around for doxygen being of reverse engineering small digestible subsets.
The example template in question << 255
Try this:
LargeClassName<LARGE_PARAMETER_NO_1,LARGE_PARAMETER_NO_2,LARGE_PARAMETER_NO3,LARGE_PARAMETER4,LARGE_PARAMETER_N5,LARGE_PARAMETER_N6,PARAMETER7,PARAM_NO8>
or with params:
LargeClassName<LARGE_PARAMETER_NO_1 : size_t =0,LARGE_PARAMETER_NO_2 : size_t =0,LARGE_PARAMETER_NO3 : size_t =0,LARGE_PARAMETER4 : size_t =0,LARGE_PARAMETER_N5 : size_t =0,LARGE_PARAMETER_N6 : class,PARAMETER7 : class,PARAM_NO8 : class = XClass>
I get 141 and 251 bytes respectively including spaces and class name, in the graph it gets truncated around LAR after Pameter 4. :-/
Also, each template parameter should "normally" be on a database in seperate table rows/cols, right?!
So, in theory each template parameter name/type/value could be varchar(255)... like this?
CREATE TABLE template_params (
uiid int NOT NULL,
parent_id int NOT NULL,
no int NOT NULL,
name VARCHAR( 255 ) NOT NULL,
type VARCHAR( 255 ),
value VARCHAR( 255 ),
);
RTF generation, I'm not sure how I'm still figuring out, I try to have "everything" in EA... However, I'm still playing with tags. For instance, having the Software Design Document (IEEE SDD), having the introduction in plain text in EA... glossary... definitions, assumptions, risks, assessments, dependancies, etc. Samething for the Software Requirement Specification Document (IEEE SRS) or Software Architecture Document (SAD).
Such that when I regenerate the document from EA,
I don't have to copy-paste around or hard-code things into a RTF template.
I'm also not sure how to create a Revision History of the document in the RTF template, with date, version, author, description of changes... inside a table such that when I change the document in EA, the generated RTF document shows the change being done.
Also, right-clicking and go to "notes" for every class, functions, attributes is not really "efficient".
Editing in plain-text is however, so having an XML file would be handy, or having proper comments in the generated C++ source code, that could be afterwards be reparsed and reintegrated easily would be handy.
Also, it seems that enum values are not properly reverse engineered into "tagged values", don't know why.
Finally, what I meant was that an EA project (EAP extension) can be exported/imported into a single XML database flatfile. It doesn't have to be efficient just 1:1 mapping would be sufficient, so we can add other scripts and tools around it more easily. ;D
Have a nice day!
-
I'll have to digest most of this over time.
However, on the last point (project to XML) what about the following feature of EA.
When you right-click the top-level element (i.e. the 'model' element) in the Project View the context menu includes two choices: Import Model from XMI and Export Model to XMI. (These have keyboard shortcuts of ctrl-alt-I and ctrl-alt-E respectively) Each has a dialog with a few meaningful options.
This might accomplish what you are looking for. The XMI is version 1.x (see the Tools / Options / XML Specifications dialog for more). If you require XMI 2.x (as I do) there is good news here too - this appears to be slated for version 6.5 which looks like it will arrive in the near future.
BTW, there was some discussion (in the Sparx forum) about Doxygen recently, and I think this concerned (among other things) comment syntax. Not being a Doxygen user this has little meaning for me, but you might do well to search on Doxygen. I cannot remember which area(s) the posts appeared in, so try them all.
I'll post more later, when I've digested things, if I think I can move you forwards. In the meantime let's see if the other members can address some of this. The forum can be pretty active.
Finally, please remember to post any successes you have with this, so others can share the benefits.
David
-
Hi,
still then, we have the 255-length problem.
What keeps Sparx away from porting the Access db over to MSDE, if not SqlServer Express?
Regards
Hans