Month Year Issue

June 1998 Issue: 3

Journal of Conceptual Modeling
www.inconcept.com/jcm

Editor's Notes: UML: Put Out or Shut Up
by Scot A. Becker

Live, From Harrisburg

My road trip has now come to a temporary halt in Harrisburg, Pennsylvania. I got me an apartment and a laptop, so I should be able to devote some more time to this Journal soon.

Visio Solutions Conference

A big thanks goes out to those of you who stopped by the InConcept booth at the 1998 Visio Solutions Conference. I hope we were able to inform and gain a few more readers. The conference went well, but the database related content was limited this year. Hopefully, next year will bring more data modeling content. However, I was impressed with Visio's flexibility, I had no idea it was that much more than a drawing tool. Interesting.

Frequency

I've decided to publish this journal a bit less frequently. From now on, the JCM will go out every other month. That means the next issue will hit the streets around August 1st, 1998. Hopefully, that will give more people more time to contribute and to give our regular writers some time off between issues.

News

Visio held an NDA session at this year's Solutions Conference and many of InfoModeler's biggest supporters and practitioners were there.

About halfway through Anthony Bloesch's presentation (sorry 'bout that Anthony!), the flood gates opened and the Visio staffers got an earful. While the comments were sometimes passionate, I thought that overall they were constructive and were echoing the sentiments that I have ranted about (and many others have felt). I think the Visio folks were a bit surprised by what we had to say, but you have to admit, they asked for it! Later on, in the hallway, the discussions continued on a more passive level and I think we got some things sorted out. As a result, we have some news:

The first of which is the fact the Visio Preferred Customer edition and it's support will be available through the Visio sales 800 number (800-248-4746) until the next full version upgrade of Visio. Don't look for Visio to actively promote this, though. They just want to "market" it to folks who are already interested in the product. This leads to some funny sales pitches, by the way. InConcept has gotten quite a few calls from sales folks (not knowing the background of the folks they are talking to) trying to tell us about the tool.

I had to call Visio tech support about an HTML problem I was having with 5.0b and I was pleased to hear the automated drone say something to the effect of "… for InfoModeler Support, press…." I have spoken with a couple of people at Visio who have been really trying to get the InfoModeler support issues resolved; you know who you are…Thanks!

Visio has also recently created an online tech support forum. You can find it at http://www.visio.com/imsupportforum/. While you are there, check out the new format of the Data Modeling portion of the Visio website.

Correction

David Cuyler at Sandia National Laboratories was bright enough and kind enough to find yet another mistake in the last issue of Common Model Fragments. He writes, "The Journal is Excellent! You've done a tremendous job of editing. The format is sensible and attractive. Articles are interesting and helpful. Keep it up!"

Oops, wrong part of the e-mail. I had to take your attention off of the fact that I screwed up again.

In reference to the mistake, he writes, "A minor correction in your May model fragment correction: The last model fragment depicting the dependency of City and State on Postal Code shows a 1-to-1 relationship between Postal Code and City. This would mean that Los Angeles (or Minneapolis) can have only one postal code. I think what works better is to say:

Each Postal Code is for exactly one City. Each City has zero to many Postal Codes. (One to many if you are ignoring cities w/o post offices, like Ramsey)"

You're absolutely right, Dave, and thanks for both the kind comments and the correction.

Stop Me, Before I Rant Again

It slices, it dices, it removes ring-around-the-collar, it's…it's…it's…UML! Am I the only one who is tired of hearing about UML? Don't get me wrong, it's not that I object to the method, it's the fact that everyone is talking about it. And by "everyone", I mean the people who don't know squat about it.

Now, I don't want to get off on a rant here (sorry, Dennis), but it seems that every conference, seminar, and marketing propaganda session has some sort of UML "content". And I, eager to learn more about UML, faithfully attend; me and everyone else.

Just mentioning "UML" causes a lemming effect: hoards of computer geeks storm into the room and do that "EF Hutton" mimic (remember those commercials? "When EF Hutton talks, everybody listens.") I haven't seen anything so effectively over-hyped since Java (and don't get me started on that!).

So, you dutifully attend the "class" and find out that the person speaking doesn't have much to say. Inevitably, you end up hearing some high level talk about how great UML is without providing any proof as to why. Then, you also tend to hear the marketing pitch.

If you ask me, these guys deserve to be locked into a room with That Guy On The Airplane who sees your proficiency with a laptop (or sees you reading a computer related publication) and asks you infinite questions about Windows while chiming in comments like, "Gosh, they sure are doing a lot with computers these days."

Save that explicit marketing for the vendor room, that's why it is there.

I recently attended a session that promised to teach me about UML and the Microsoft Repository. I think "Microsoft Repository" was mentioned twice… in passing. Of course, because it was UML "content", there were so many people in the room that made the later attendees sit in the aisles. Thus, when the marketing pitch began, I had no escape. I really wanted to commandeer the "lecture" and talk to the attendees about what I knew about UML. While that is pitifully little, it would have been more content filling than what was being offered. Ugh.

So what's the solution? Here are Uncle Scot's surefire tips on good UML presentations:

Say something about UML. Anything. As long as it is about UML and at a level other than "It sure is going to revolutionize software design." (Incidentally, see Dr. Terry Halpin's continuing series on UML and ORM for an example of what TO do. He actually has something to say and is providing useful information, thus satisfying this criteria.)
If you notice most of your attendees are looking at the mass of people sitting and blocking the aisles, presumably thinking about a means of escape by walking on their heads, you are probably not meeting the criteria above.
If half of the people are looking at the masses clogging the aisles, and half of the people are looking at you and smiling, then you are failing to meet criteria one but also have something hanging from your left nostril (or, your fly could be open). You have now doubly embarrassed yourself.
If your nose is clean and your fly is up and still you have people paying attention to you even though you are violating criteria one, then you have a population of buzzword speaking drones. Use your apparent hypnotic power to suggest another career path for these folks, such as real estate or insurance sales. The best way to do this is to say something like, "The best way to advance your career by dropping UML terms even though you don't understand them is to sell cars." You may as well do something productive while you are wasting my time and I may even forgive you for your previous digressions.
If you still insist on giving a worthless presentation, at least size the room accordingly. I recommend a room big enough to house 75% of the total conference attendees. That will at least offer me a means of escape (so I can ridicule you in the hall) without me having to plant my foot onto anyone's head on the way out the door.

Are you noticing a pattern?

Preface

We have another great set of content for you this month. Dr. Anthony Bloesch gives us another installment of InfoModeler tips. Dr. Terry Halpin continues his series on the differences between ORM and UML. Dr John Sharp shows us a bit more about NLM and offers a discussion about denormalization. Pat Hallock talks a bit about embedded objects. Dr. Gordon Everest hands out the midterm exam and Dr. John Sharp introduces another analysis problem and gives us last issue's solution.

Scot A. Becker is a software consultant and the founder of Orthogonal Software Corporation. He is also a certified ORM consultant and trainer, a certified Visio trainer, and former Editor of the Journal of Conceptual Modeling.

Contact Information:

Scot A. Becker
Orthogonal Software Corporation
scot@orthogonalsoftware.com
www.orthogonalsoftware.com