Talk:Glossary of Unified Modeling Language terms

UML version 1.x
Hm. At the moment I'm asking myself whether it would make sense to just concentrate on UML 2.x and just leave away the 1.x section. What do you think? --Adrian Buehlmann 18:04, 28 January 2006 (UTC)

I partly agree but I thought this way there would be less scope for confusion between the versions - it is clear were each term belongs and it will be clear even if there is a UML 3.x. I don't want to do 1.x terms myself but this way if someone wants to describe 1.x terms there is a place for them to do it without diluting the 2.x section. Mark.murphy 23:28, 28 January 2006 (UTC)


 * Yeah I see your argument. Let's leave it for now. Maybe somone has a good idea in the future. We have a somewhat similar problem on the Unified Modeling Language article too. I thought about creating separate articles for each version of UML but that would be a bit overkill, especially for Unified Modeling Language. Think it's best to leave everything as it is. --Adrian Buehlmann 00:15, 29 January 2006 (UTC)


 * I say we trash it. There are no entries in that section now. We can note the version differences (and their peculiarities) specific to the article in question.  I have not found a distinction in the wiki between the two versions.  // Brick Thrower 11:10, 22 September 2006 (UTC)

Validity
I'm very concerned with the quality and validity of these definition. They don't seem precise enough or accurate enough to be useful and are not defined within the context of UML.

some examples


 * Inheritance - where a new more specific Class derives part of its definition from an existing more general Class

While this sounds good, UML supports inheritance of other than clasess (e.g., Use Cases)


 * Generalisation - in Software Engineering, a solution to the problem of programming complexity which involves reusing code; it is often implemented using Inheritance: the more specific code resues the more general code

Normally, Generalization and Specialization are the UML terms, inheritance is the programming term. but in any case, it's mostly irrelavant what Generalization means in Software Engineering, if this is a glossary of UML terms, it should be what does it mean within UML. BTW, Generaliation is a powerful organizing technique for reducing complexity, but need not involve reusing code.


 * Type - the options are: an elementary Value type such as integer, string, date or boolean or a Reference type defined in a Class

Seems circular to define Type with Type


 * Association - a relationship between 2 or more Classes. Each end may have a Role, Multiplicity and be Navigable

This is really a common error, while the association is usually drawn between two classes, it can be drawn from a class to itself. But more fundamentally, the association relationship is between 2 or more instances. Generalization is between classes, associations are between instances. To be precise, a link is between 2 or more instances, an association stands in the a relationship families of links in the same way that classes stand in the relationship to instances.


 * Multiplicity - a specification of the number of possible occurrences of a Property; Optional: 0..n; Single Mandatory: 1; A single value: 1; Undefined multivalued: *; Multivalued with a minimum n: n..m

Optional is usually given as 0..1

I would go on, but the majority of the definitions need signficant rewrite to be correct. There is already an official definition in the UML spec (I wouldn't trust Fowler for formal definitions), and definitions should probably be placed in Wikidictionary? Mjchonoles 04:25, 1 February 2006 (UTC)


 * Hi Mjchonoles. Good to see your sig in blue color :-). To the matter: I think as these definitions here are technical ones, I'm rather in doubt whether they should go to Wikidictionary. I assume you mean Wiktionary? --Adrian Buehlmann 09:50, 1 February 2006 (UTC)


 * Hi Adrian. Unfortunately, my sig is an empty page. But back to the matter. If we want the definitions to be technical ones, they need to be technically correct. My guess is that more than 50% of the definitions are currently wrong in some way -- using UML 1.x terminology, misunderstanding UML fundamentals, using a programming point of view, or just generally vague. Should we scrap the whole page, scrap any definitions that seem questionable, or attempt to fix them. Fixing them would be a considerable amount of work, unless we take the definitions from the spec (we'd have to look at the copyright) and even then, the definitio are not meant for end-users. While I believe it is possible to make accurate and useful definitions. Mjchonoles 06:09, 2 February 2006 (UTC)


 * I agree that it would be good if everything were correct on Wikipedia. The copyright might be a problem as we are probably not allowed to take out parts of the UML spec and slap them verbatim into Wikipeda. Smaller quotes should be ok. But I'm not a copyright expert. It is also questionable whether it makes sense to try to duplicate the UML spec here. On the other hand it might be wise not to prematurely destroy Mark.murphy's work. Oh well, I fear I'm not too much of a help to you on this matter :-). Of course if you see any chance to put this here into better shape, please do so. --Adrian Buehlmann 08:27, 2 February 2006 (UTC)


 * I also agree that there are issues here with the definitions. Combining the definitions policy with the policy about what Wikipedia is not has lead me to create my own wiki just for UML that does not have this specific restriction.  I am referring to the portion that says that the wiki is not a place for original material.  I envision a Wiki where folks who elaborate upon definitions for precision create their own articles for whatever class or concept they are trying to discover...an public domain model mart for UML models would be nice, but current wiki policies here seem to exclude that dream.  // Brick Thrower 09:39, 31 March 2006 (UTC)

2007-02-6 Automated pywikipediabot message
--CopyToWiktionaryBot 23:41, 6 February 2007 (UTC)

Request to Disambiguate
There are many terms listed under S/Stereotype that link to disambiguation pages or worse, incorrect contexts. Could someone knowledgeable in UML go through and disambiguate these, or remove wikilinks? Hoof Hearted 15:27, 26 February 2007 (UTC)
 * While disambiguating Responsibility, I found the reference here. I believe it should point to Single responsibility principle, and have changed it to reflect this. If this is wrong, feel free to change. The rest of the list still points to many incorrect pages. Jujutacular talkcontribs 17:16, 29 August 2009 (UTC)