Open source UML editors lag proprietary leader

What may have been valid critiques of open source UML modelers was just spoiled by the author's uninformed and biased opinions about the "seriousness" of programming languages and the usefulness of UML itself...

Maybe open source modelers are indeed lagging behind the one proprietary app the author refers to. I don't know, I've never used one. But the author's opinion that only C++, Java and Python are valid choices for "serious" applications (I suppose he includes .NET in that list too) destroyed all the credibility he could have had with me.

As for UML, the early comments give it all: its usefulness is restricted to a very specific environment, and for everything else it's just useless overhead. And in my (admittedly secondhand, I've never used UML, but colleagues of mine on a different project do) experience, even in an environment of business apps built by large teams, too much dependency on UML can lead to a lot of useless overhead too.

Use tools when they are useful to you, but don't tout something as the Holy Grail of programming (there's no such thing), and don't make arbitrary judgements on the "seriousness" of programming languages.

Tsela: Your message brings an important distinction to light. I appreciate what you wrote. I also feel it extends beyond this one case.

The web has opened the flood gates for writers who previously did not have an outlet for their materials. When print media held a monopoly and the only way one could publish, it restricted the contributions of many excellent writers.

Unfortunately, the web also provided a forum for informal language, sparse research and a mix of subject and objective materials. Many writers don't have the benefit of a second pair of eyes when they press "submit". Many publishers are just happy to have content.

No doubt, subjective and uniformed bias takes away from the credibility of an article and its author.

tadelste: it is indeed something that I've seen quite often, especially in the IT press. It was just so glaringly obvious here that I just had to comment (by the way, I have tried to "vote on this", but it seems my vote never gets registered...).

Luckily, there are indeed a *lot* of good writers that the Internet allowed to reach out to people like me. I read quite a few blogs everyday and although I don't always agree with the authors, at least they document and explain the reasoning behind their opinion. This guy, on the other hand, just gave his personal and unjustified opinions as intangible truth. How can you trust someone like that indeed...

Oh well, as they say in Dutch "elk voordeel zijn nadeel": "to each advantage its drawback", and vice-versa ;) .

The real reason that so few Free Software projects have UML modeling of their code is that UML is user driven. Use case modeling is why most places I have worked use UML and it is all part of the requirements side of things. You get requirements, model requirements and show the users something easier to explain than a code walk through of the features.

In an FS project you aren't being driven by user requirements in the same way if at all.

And why model up front, to save reworks of tested code later when you discover it doesn't meet requirements. By doing so you make a major (so theory and in my experience usually in reality) savings. In an FS project you don't have the same worry about getting it right first write and release. You can give the release to people to use it and they can provide feedback on the real thing. An eternal Beta.