EJB and UML modelling

[javelin5]

I have been using EA for awhile and a friend was telling me today that EA could not be used to model EJB (enterprise java beans).  Is this true?  Or is the person mistaken?

I have never developed EJB, I might be doing that in the future.  I prefer EA over Rational Rose and want to be able to keep using the tool.  Also, if somebody could give me the basics in how to model EJB I'd appreciate that.

[PhilR]

The friend is either misguided or has some underlying motive for saying this.

There is a UML profile for EJB developed by Sun.  EA has a neat facility for defining profiles.  Why not post to this board and ask if anyone has defined a profile or define one yourself based on the UML profile (not difficult).

Try the folowing link for more details
http://salmosa.kaist.ac.kr/~course/DrBae/cs650_2001/LectureNotes/UMLProfileForEJB.pdf

[javelin5]

That's what I thought, they are a big proponent of Rational.  I will look at the info and try to make my first profile.  I assumed that any UML modeller could do EJB (not knowing EJB I assume it conforms to Javas OO concepts).  I would recommend that Sparx put a EJB profile together and post it to the web along with a blurb that they have an EJB profile for use, or someone else give them an EJB profile.  My friend was using that issue to try to sway people at my company away from recommending EA for use in UML modelling because it could not do EJB.  I am not very happy with this kind of attitude I run into, especially from a friend who obviously does not understand UML and OO design as well as I had thought.

Tim

[PhilR]

Wierd isn't it?  I have seen a lot of this during my career and think its a resistance to change - not a good thing in our industry.

Good luck with the profile.

Phil.

[Steve_Straley]

Phil,

What is worse (or better if you are from Rational) is that most companies tightly link the concept of RUP to the product line so when they engage in best practices or the RUP mindset (which they still cut corners), the automatically think that only the Rational set can be used.   That is the worst up-hill battle of all in my opinion.

Of course, these are the same "types" of people who ask you to develop USE CASES when in fact what that really are meaning to say develop full UML documentation and RUP documentation... but "use case" seems to be the catch-all for both...

Steve

[javelin5]

I am glad to know that this kind of attitude doesn't just happen where I work.  I think the best tool for a given job should be used.  I agree that there is a stigma that you can only use Rational Rose with RUP.  People don't grasp the concept of UML being a standard and any tool that fully implements the UML should be able to bu used for RUP.

Tim

[Steve_Straley]

Tim,

Hey there....

The trouble is many fold, as I see it.

First, companies SAY they use RUP when in fact they always short circuit RUP because of various reasons including the notion that they don't really understand RUP.

Second, companies SAY they use RUP when in fact what they are REALLY saying is they use Rose.

On the UML side, it is amazing that people in companies who say RUP don't understand that there is a difference between USE CASE diagrams and USE CASE documents.   Moreover, you ask these same people to define the following...

USE CASE
USE CASE DIAGRAM
USE CASE DOCUMENT

they seem to think they all mean the same thing.   So, for example here, I am pushing to have ACTIVITY, SEQUENCE, and STATE diagrams done as well, they don't see the need since it is all "handled by the USE CASE".   I then mention that they don't want UML people and they certainly don't want RUP if they don't have the complete picture and I'm replied with that glassy-eyed stare that comes with "what on earth are you talking about?" <ROFL>

On another note, one of these days I would love to write (a colleague of mine is thinking the same) a "USE CASE Cookbook" with various types of diagrams... system to system, integration, passive integration (e.g. Web Logic) complete with diagrams and documentation.   Too often I would like to see some example of an odd and/or unique process and I can't find it.   Oh well... another day....

LOL

Cheers,

[PhilR]

Drives you mad doesn't it?  As well as resitance to change, I blame the way that people acquire knowledge in this industry.  First they hear a buzz word or acronym.  Then they read the minimum to see if it is relevant to them.  The they start talking about it as if they are experts.

Try this on people...

The UML is a standard notation similar to the one used for electrical circuit diagrams.  The standard tells you nothing about how to do analysis/design etc  Any more that a notation for electrical circuits tells you how to design satellite receivers!

The original designers of the UML (known as the three Amigos) also developed a process called the Unified Process (note the complete lack of an 'R' for Rational).  UP tells you how to apply the UML to the analysis and design of software.  It is in the public domain and there is a text book which fully describes it.

UP is not a complete software development process - it lacks advice on things such as project management, change control, quality assurance and configuration management.  

By strange coincidence, all of the Amigos work for Rational (and are probably major shareholders as well).  RUP is Rational's proprietary version of the UP.  It adds the missing bits to UP to make it a full-blown methodology.

Tools such as EA, Rose, Together J or Poseiden serve two functions:  they provide a drawing tool that conform to the UML notation and a repository which is linked to objects on diagrams created with the drawing tool.

In a perfect world, the tool would have no influence on the process being followed (hard to acheieve in practice because tool designers love to impose their own restrictions and interpretations).

Ideally you have three quite independent things: notation (UML), process (UP) and tool (EA rules).  You should be able to make a rational (pun intended), and cost-justified decision on what to use in each area, independently of what you use in the other areas.

If you are a marketing manager.  You try to build a strong link between the three so that you fill everyone's mindspace with the UML/Rose/RUP trinity.

...now watch their eyes glass over and label you a subversive, wierdo or whatever.

Final point.  You can quite nicely combine combine UP with other public domain stuff such as the PRINCE2, or PMBOK project management methodologies.  You can get lots of process guidance from the CMM.  Then buy a copy of EA and you pretty well have the entire Rational circus.  Total cost US$150  

I have some more references on this if anyone is interested.

Phil.

[kelly_sumrall]

Phil

You say you have more references.  Please share.  I work for a company where there is the potential of purchasing software without fully exploring the options due to time constraints.  The decision is based on buzz words, magazine advertising, and perceived value.  Help me get some ammo to present to my leaders to help convince them of the true value of EA.  I have been using this tool for about 6 months and my world is a brighter place.

[Steve_Straley]

Phil,

Exactly right.  But added to this is corporate mismanagement in making the RIGHT choice.   Often, too often, price warps the mindset.   The trouble with great tools like EA is that big corporations think it is not worthy of their time if it is not priced high enough.   The trouble with sub-standard tools like Rose is that they are priced TOO HIGH, come with a complete infrastructure of high-pressured sales people, mediocre tech people, and developers disconnected from the public and as a result, sub-standard software.   And corporations LOVE this approach (being sarcastic at the moment).   It amazes me to this day.

I remember once I wrote an accounting program.  I had three versions: Gold, Silver, and Bronze.   The Bronze was priced at $99.00.... the Gold was priced at $1000.00.   Everything was hunky dorey until I was asked once "What's the difference between the Gold and the Bronze?"   My reply: "The Price... that's it."   Needless to say, people were a bit upset purchasing the Gold when they could have purchased the Bronze.... and when I said "Hey, you wouldn't have considered the software unless it was priced high enough to hit your radar screen... and now you complain?  How absured."

But the point remains... we perceive "value" based on price and marketing hype and large corporations (I'm consulting for GE at the moment and it is true here) perceive quality based on:

a) Price Points
b) Sales People woo-ing them to lunch constantly
c) Marketed integrated solutions (that are not integrated and too often closed)

Thankfully, Geoff and Paul keep hope alive that there is a better way to do things...

Sorry for the rambling...

Steve

[PhilR]

Kelly, I will post some references under a new thread so that more people can use them.
Phil.

[flego]

Has anybody a EJB profile for Enterprise Architect ?

[RichK]

Hello Everyone!

I've just read (in a manner of speaking) the EJB profile, and it scared the life out of me!

I model EJB's extensively using EA; all I require is a <<session>> or <<entity>> stereotype, that works for me. I don't need any more complexity on my models.

And of course, EA can do this just as well (better...) than Rose can.

I let a tech manager evaluate EA against Rose the other week; I studiously avoided telling him the price of EA (he knew the cost of Rose). I subtly hinted that EA cost about the same, maybe more, as Rose.

EA came out the clear winner. They nearly died when I told them the price of EA. I bet EA would have had a much harder time had they known the price. It is a funny world we live in...

Regards,
Rich
Ariadne Training

[javelin5]

Yeah, that profile scared the bejeezus outta me too.

I personally have never done EJB, but if you have a simpler method could you breifly explain how you are modeling EJBs.  I am curious what the <<session>> or <<entity>> actually represent.

"I model EJB's extensively using EA; all I require is a <<session>> or <<entity>> stereotype, that works for me. I don't need any more complexity on my models."

[PhilR]

Sorry if the specification scares people.  My intention in posting it was not scare everyone off.  Remember that these specs are defined with enough detail for tool vendors (such as Sparx) to build tools that generate code, deployment descriptors etc.  Check out the official UML spec if you really want scared.  

Also this profile is definately not the way to learn about EJB.  try one of the more gentle introductions at

http://www.cetus-links.org/oo_java_ejb.html#oo_java_ejb_tutorials

The section of real interest in the spec is 3.5 Standard Extensions.  This is where all the stereotypes and tags are defined.  Note that they model EJBs as packages rather than classes.  They also have stereotypes for the component parts of EJBs such as the home and remote interfaces.

At the end of the day the stereotypes <<session>> and <<entity>> will do just fine if you don't need the greater prescision offerd by the stereotypes in the spec.  Be careful though, if you share your models with other modellers.  The <<entity>> stereotype is used in another context for Robustness Diagrams.  Stereotypes <<EJBSession>> and <<EJBEntity>> might be better to avoid confusion.

Rich, I love your comments on the evaluation of EA.  Maybe Sparx need a "Corporate Fools Edition" of EA.  Crippled functionality, additional bugs, fancy 'T' shirts and a ten times price tag  

Phil.