[PD] "Scripting languages" ?

Mathieu Bouchard matju at artengine.ca
Sun May 23 21:52:22 CEST 2010


On Sun, 23 May 2010, Pedro Lopes wrote:

> The truth is with the newest developments, scripting is of course gaining a new meaning. By old "meanings" a scripting lang would have usually
> characteristics like:
> - being interpreted from source code (traditional langs are compiled as you know)

BASIC is a traditional language that is interpreted from source code (or 
from a kind of byte-jcode that is a simple search-and-replace on the 
source code). But as early as the eighties, there was a compiler for it.

LISP is a traditional language that is interpreted, yet became quite 
compiled as an option. It's quite notable as there is a full interpreter 
of itself in the standard library that can be used by compiled programs, 
at runtime (or compile time too).

> - being separate components from main application and so forth (but this 
> has gradually changed nowadays!)

>From what main application ?

> One does not usually call "shell scripting" "shell programming", simply 
> because its interpreted scritps (like so many other langs),

What is usually called "shell" is a series of programming languages that 
were born with the initial goal of providing a main interface to the OS 
and thus have a special focus on managing other programs.

> but OF COURSE when you're scripting your programming... :)

No idea what that means...

> An interesting description could be borrowed Ousterhout from where he 
> states: "they are intended primarily for plugging together components",

Most any programming language is about plugging together components. a 
function is a component, a variable is a component, an object is a 
component. (nothwithstanding the specific use of the word made for 
"component-based programming"). The size and complexity of components 
alluded to in the sentence "they are intended primarily for plugging 
together components" is only there to say that real tough code is made in 
real tough languages and that for the wimpy stuff you can use wimpy 
languages.

There is a very relevant discussion to be made about how Pd fits in all of 
this, if it is to be considered as a "scripting language".

> but then again... they (scripting langs) have evolved so largely that 
> they have surpassed that in many ways. But he was saying this in 
> 98[1]...

So what ?... I've seen a lot of Tcl apps in 1997 and most of them didn't 
have to do with being the house pet of a Real Application. It's ok to have 
people like Ousterhout talk about why they first invented a language, but 
in many cases it doesn't have anything to do with the nature of the thing, 
and you can see that by how quickly its users can ignore the original 
intent. Ousterhout was writing with his own vocabulary about a reality 
that it wasn't fitting in.

He even confused strong typing with an API design that happens to not care 
for concise application code, lack of default args, lack of keyword args, 
etc.

I used to think that this article is really mind-blowing, but that's 
because I was a teen undergrad who was in the process of switching from 
DOS to UNIX. Nowadays I have trouble reading it because it makes me want 
to rewrite the article.

Then maybe the whole thing was just a plot to popularise alternate 
languages with a crowd of programmers that had pretty narrow views about 
what constitutes a Real Programming Language. That's how one needs to 
start to distinguish between Programming and Scripting.

> it is funny to see how it all evolved. There's another usual name for 
> this "glue code" that some of my oldest teachers used to say in my first 
> years of college,

The concept of "glue code" is as ill-defined as the concept of "pseudo 
code"... vague ideas based on impressions and on language politics.
"pseudo code" usually meant anything that looked like C, PASCAL and all 
the competing languages of the time, usually made with a syntax slightly 
different from anything in order to avoid favorising any language over the 
other. It was a kind of neutral ground for sharing research in a 
fragmented world of dozens of languages and dozens of CPU models. As such, 
"pseudo code" was a political game. Now, I suppose you can imagine too a 
politically-defined concept of "glue code"...?

> True true, there's a lot of ecma compliant stuff out there, but still 
> too esoteric for me because apart from AS and js haven't really learned 
> any of them...

ISO-9660 (CD-ROM filesystem) is a well-known ECMA standard but it is not 
well-known that it's an ECMA standard. It was first defined by ECMA then 
adopted by ISO.

  _ _ __ ___ _____ ________ _____________ _____________________ ...
| Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801


More information about the Pd-list mailing list