Pseudo-OOP Code?

[Brenden Vickery]

Brenden, I think you make a good point but base it around the wrong argument: that methods that return "something" should be evaluated for "smell". That, in and of itself, is terribly misleading.

What you're really talking about is changing those methods when composition can be applied to your class structure.

So, to that point, I would agree with the promotion of composition where applicable.

I guess I havent communicated those points very well. What I actually have been doing is looking for return statements in code I just wrote and evaluating it to see if its a bad spot to return. Maybe thats not the best method but its something Ive been doing. What I would like to do, is to be able to make the right decision before hand and not have to refactor things like this, but Im working on it.
Obviously none of this is new, but can be found in Refactoring by Martin Fowler. I think the biggest difference between procedural, Pseudo-OO and OO code is the amount of correct refactoring that has gone into it.

[stereofrog]

- methods that return anything (Shouldnt I tell this method to do something instead of asking it for something?)
- if statements (Cant I use polymorphism here?)
- switch statements (Cant I use polymorphism here?)
- instance of / is_a (Cant I use polymorphism here?)
- nested if statements (Make a new private method, this is easier to read.

- loops (replace with iterators)

Didn't we just invent smalltalk?

[tobyhede]

I use what I call 'Object Procedural' programming with PHP. I predominately use classes as namespaces, rather than as true objects.

I think the PHP model is profoundly broken when it comes to OO, even in PHP 5. The actual page->request model used by PHP is the fundamental issue, it makes all OO programming essentially moot.

In a truly Object Oriented langauge, such as Java, OO for web apps is possible, but I am yet to see a pure OO program written in PHP that doesn't actually make life more complex for developers. The number of times I have seen the Singleton pattern used, for example, is a joke. Given that you can actually have global variables in PHP, there is no reason to use this particualr pattern. In Java, it's required because of the way the VM works.

[lastcraft]

Hi...

And what exactly smells bad about a collection of utilities? A string utility? or a number utility? or a math utility? or a validation utility?

The sense of smell comes from experience of chucking up after eating rotten food.

Util classes by definition lack cohesion, they do lot's of odds and ends. Better is to place them into a lot of small classes and then have them grow naturally. At least that way you don't have to include a lot of junk you don't need when you include the utils.

That PEAR validation class sounds pretty low quality if it has no interface design.

Maybe your nose is the problem. Ever think of that.

This is philosophically feeble. We would throw away all human judgement wouldn't we? If there are lot's of ways then there would be lot's of opinions from the older/experienced practitioners and nobody would ever be able to convince anyone else of anything. In fact expert knowledge tends to converge. I see the "utils class is bad" rule independently discovered all over the place. People don't go back to util classes once they have abandoned them.

Maybe it does to you, but then just plug your nose and walk away.

I don't like smelly stuff piling up around me. I'd rather call the council and have the dustman remove it. It's unhealthy otherwise.

yours, Marcus

[arborint]

In fact expert knowledge tends to converge.

That line stood out to me, because it is the one thing that unexperienced programmers have the hardest time believing. Unless you believe that idea, then things like TDD, patterns, OOP, anything Agile, etc. just don't make sense because they are often counter-intuitive.

[Dr Livingston]

Isn't ignorance got something to play in that though? To be a better programmer you need to have an open mind, in that you want to learn more, learn something better, something smarter.

Then again maybe it's not got anything to do with ignorance, nor the lack of experience either - but it's got everything to do with the lack of support from those above you. The attitude may be that they're not interested in individual personal development within the company, that they're only interested in getting the job done, quickest possible route taken.

It takes a shockwave for instilled attitudes to change.

[a.dotty.dot]

The attitude may be that they're not interested in individual personal development within the company, that they're only interested in getting the job done, quickest possible route taken.

Or, in the case of freelancing, they may not want to pay what it's worth to write more than a crude hack.