Are they static or in decline? I've not seen any data to suggest that.

Anyhow, while it's a proven technique, it hasn't seen wide-spread use since the 80's when the research I'm referring to was conducted. I can only speculate as to why.

Isn't that exactly what we have now?

Where would they begin? They'd almost need to be experts before they could make any sort of meaningful choice. As beginners tend to have little to no preexisting knowledge (that's what makes them beginners, after all) they'd essentially be making their selections at random. I don't see how that would benefit beginners in any way. It would like learning to swim by getting dropped from a helicopter in to shark-filled ocean waters.

Hey, it's your definition. Enjoy. Just don't expect many people to agree with you. (That's my point!)

Hell, even I'll disagree with it on the grounds that it's insufficient. On it's face, it's so broad that it encompasses just about everything, which is clearly not what you intend. Mix in a bit of context and you just get a simple definition of "object".

You'll need to tweak your definition a bit before you can derive anything. What are these principles that naturally derive from your definition (or hopefully soon-to-be modified definition)?

Don't be silly. Meyer doesn't offer anything of the sort in OOSC, nor does anyone anywhere else. (If they could, you'd necessarily have a solid mathematical foundation and no further ambiguity as to what constitutes OOP!) What you do find in his overweight book are a lot of bold claims which are NOT supported by the literature. (For example: Right up front he claims that people who "master" OO techniques can produce software faster and cheaper.) He also makes an irritating habit of misrepresenting functional and structured approaches to bolster his never-ending salesmanship of Eiffel. The unnecessary build-up to the big "reveal" comes off more like a side-show performance rather than a scholarly work (which, by the way, it is clearly not!).

Moving back for a second...

Do you believe that the "principles" Meyer offers in his advertisement -- oops, I mean book -- can be derived from your definition?

Apparently not. I misinterpreted "the unique thing". My bad.

Still, it's clear that you're confused. OOP did not introduce objects.

Ignoring for the moment that that has absolutely nothing to do with my comment: that statement is completely incoherent.

And not at all like OOP as it is generally understood by Java, C#, and C++ programmers. Again, no one can even agree on what constitutes OOP. It's an incoherent concept at best. Dig up a few definitions. You'll find most are vague, many are inconsistent, and a few so broad that they're meaningless.

Indeed it does. Modularity is good and many beginners seem to naturally understand the need. OOP in the *ahem* classical sense, however, is anti-modular and is thus ill-suited.

Classes aren't essential to objects, you know. I'd even go as far as to say that they're a harmful concept. Skim through whatever books you have lying around and you'll soon discover that a good bit about what is written about "OOP" is ultimately about dealing with the problems caused by the flawed concept of classes and class-based inheritance in many popular so-called object oriented languages. An easy example would be design patterns. You'll find that a good number of the "patterns" found in GoF are intended for problems that you simply don't encounter, or which have solutions so trivial that they don't warrant special attention, in other languages, including other so-called OO languages.

It could be that you believe classes to be essential to OOP, but I've not seen many people make that claim.

It wasn't arbitrary. It's a proven approach that has been shown to be successful; even with very young children. I honestly never considered the possibility that it couldn't be near universal.

What do you have in mind as alternatives?

You're assuming that they don't. Which has the more potential for harm? Remember: Until recently, it was thought that babies don't feel pain.

If I'm to err, I'd rather my mistake not result in torture.

So what you like are objects, not necessarily OOP (by whatever definition). Objects, used appropriately, can be useful. They're no panacea, of course, and their should be little doubt that they're misused more often than not.

I can't possibly disagree more here. OOP (in the Liskov sense) is anti-modular by its very nature. Languages like Java and C# make it much more difficult for beginners to learn about modularity as a result.

Please, enlighten me then! Show me the error of my ways.

If you can, stay away from platitudes, they won't actually help you make your case. I completely understand if that's all you can offer; that's my point after all.

People say that, but they never list the alleged benefits.

Give it a try. You'll fail, of course, but you'll ultimately find the exercise enlightening.

I get the attraction to Lisp. It's the same attraction I have to forth. They're both fun and interesting to implement. Well, at least when you first encounter therm.

So, yeah, I guess "Because it makes writing an interpreter really easy" is a good call.

I've heard lots of good things about it. I just haven't heard anything about it that's unique to OOP. Hell, no one can even agree about what constitutes OOP!

I'd ask you to defend it, but you'll just mumble something ridiculous about "modularity" and I'll be forced to laugh at you.

I thought this was supposed to be uncontroversial?

Classes? You can't be serious! Logic? If slashdot is any indication, most programmers couldn't pass an undergrad course in logic. Sure, you'll need an incredibly tiny subset of logic, but it's not a topic that needs special consideration. Most beginners work out the basics themselves without assistance. They certainly don't need a through understanding of the subject like a philosopher. I could go on ...

A better (much shorter) list of fundamentals:

- Data - variables, types, and structures (Just the basics, of course.)
- Flow control (iteration, branching)

That's it, really. (I may have even gone too far!) I can't really add anything else without stepping in to controversial territory.

Useful is a different matter, naturally, and we could add quite a bit more to the list. Though I doubt that it would be nearly as long as yours without becoming controversial.

Design patterns?

Jesus, it's like you're a best troll ever.

Spot on.

Sarcasm aside, it's a useful algorithm that beginners can easily understand. With the right prompts, they can even "invent" it themselves.

I was with you until the last paragraph.

How about we point beginners to a simple imperative language with an interactive mode? Pack it with simple but interesting games they can understand and modify.

You know, the same stuff we had when we were kids. It sure beats the shit we're giving them now.