What are OOP's Jargons and Complexities?
Dale King
DaleWKing at insightbb.nospam.com
Sun Jun 5 00:47:08 EDT 2005
Anno Siegel wrote:
> Tassilo v. Parseval <tassilo.von.parseval at rwth-aachen.de> wrote in comp.lang.perl.misc:
>
>>Also sprach Dale King:
>>
>>
>>>David Formosa (aka ? the Platypus) wrote:
>>>
>>>>On Tue, 24 May 2005 09:16:02 +0200, Tassilo v. Parseval
>>>><tassilo.von.parseval at rwth-aachen.de> wrote:
>>>>
>>>>
>>>>>[...] I haven't yet come across a language that is both statically and
>>>>>strongly typed, in the strictest sense of the words. I wonder whether
>>>>>such a language would be usable at all.
>>>>
>>>>
>>>>Modula2 claims to be both statically typed and strongly typed. And
>>>>your wonder at its usablity is justified.
>>>
>>>I used a variant of Modula-2 and it was one of the best languages I have
>>>ever used. That strong, static type checking was a very good thing. It
>>>often took a lot of work to get the code to compile without error.
>>>Usually those errors were the programmers fault for trying to play fast
>>>and loose with data. But once you got it to compile it nearly always worked.
>>
>>I am only familiar with its successor Modula-3 which, as far as I
>>understand, is Modula-2 with uppercased keywords and some OO-notion
>>bolted onto it (I still recall 'BRANDED' references).
>>
>>I have to say that doing anything with this language was not exactly a
>>delight.
>
>
> I've been through Pascal, Modula2 and Oberon, and I agree.
>
> These languages had an axe to grind. They were designed (by Niklas
> Wirth) at a time of a raging discussion whether structured programming
> (goto-less programming, mostly) is practical. Their goal was to prove
> that it is, and in doing so the restrictive aspects of the language
> were probably a bit overdone.
I fail to see how they were that different in terms of structured
programming than C. The main benefit I was talking had more to do with
types. It had types that were not compatible just because they had the
same base type. For example you could have a type inches that was an
integer and a type ounces that was also integral. Just because they were
both integral did not make them type compatible. You couldn't just
assign one to the other without you as the programmer explicitly saying
that it was OK (by casting).
In the environment I was programming in (engine controls for cars) where
safety was a critical thing and a programming bug could kill people that
safety was a very good thing. I think that also has a lot to do with why
the government uses Ada.
> In the short run they succeeded. For a number of years, languages of
> that family were widely used, primarily in educational programming
> but also in implementing large real-life systems.
>
> In the long run, the languages have mostly disappeared from the scene.
I've posted before that hardly any language that has ever been somewhat
popular has actually died (depending on your definition of that word).
When asked for someone to name one once I got Simula for example (the
forerunner of OO languages). Turns out that it continues to actually
grow in popularity.
> It has been discovered that "structured programming" is possible in
> about any language. It turns out that programmers prefer the
> self-discipline it takes to do that in a liberal language over the
> enforced discipline exerted by Papa Pascal and his successors.
There are lots of reasons they have not taken over, although Ada is
still in wide use. It seems to me that too many people like playing with
dangerous power tools without the guards in place.
--
Dale King
More information about the Python-list
mailing list