Definitions Of Oo Considered Harmful

Any of the definitions of "object oriented programming" tossed about on this wiki and in the literature necessarily (and often quite purposefully) exclude one or more languages that is widely considered to be an OO language. We contend that "object oriented" has long been a meaningless term, and attempts at a precise definition at this late date will satisfy only those with a partisan agenda, while a broad definition provides no value.

Wait a second, doc. What do you think was the partisan agenda behind NygaardClassification. KristenNygaard's reputation and integrity as a scientist and as a co-inventor of ObjectOrientation have never been questioned. Don't try to paint UncleBob's definition and Nygaard's definition with the same brush.

"Object oriented" has so long been an alleged GoldenHammer synonymous with good that different camps in the development world and on Wiki fight bitterly over ownership of the term, and marketers will stamp it all over any old warmed over piece of software in order to create buzz.

Let's just give "object oriented" away. It's not doing us any good anymore.

The set of languages which are considered object oriented includes such odd bedfellows as EiffelLanguage and JavaScript, which have far less in common with one another than they have with non-OO languages AdaLanguage and ToolCommandLanguage, respectively. It can be seen as a fairly arbitrary grouping. If, however, we look at these languages in terms of features, more useful distinctions are available.

Ada'95 is for all intents and purposes an ObjectOriented languages, until proven otherwise. This is according to its authors and with highly reputed and accomplished scientists and engineers from the Ada community whop wrote books and papers on Ada and its features. To question the object orientedness of Ada is IMHO, kind of on the flamewar side. -- CostinCozianu

Therefore, we suggest that these discussions are better phrased in terms of particular language features rather than using the sweeping and vague term "object oriented".

Discussions of whether or not a given language is object oriented are right out. It is much more fruitful, for instance, to discuss polymorphism in C++ vs polymorphism in Smalltalk than it is to discuss which of these is or is not an object oriented language. Polymorphism, to continue the example, can be reasonably defined, and is specific enough that it can be reasoned about.

Past [based on?] on past debates, I don't think it will be that smooth.

I think this is similar to the DefinitionOfLife battles. Rather than try to find a perfect definition, maybe treat OO as a list of features or tendencies. Something is considered OO (or OO-ish) if it has enough of the features.