REST Questions

Dave Winer
Sun, 26 Aug 2001 07:06:35 -0700

Hi Jeff!

I want to tell you a little story.

At WWW9 in Amsterdam after one of the sessions I was in a group of people
talking about where XML could go, and we were having fun, I said something
like this "You know we could do a taxonomy for all information about a
person, it's not impossible, let's work on it!"

Two or three people groaned in a nice way and got up and left. "Been there
done that," they said or something like that. I gathered that one could have
that discussion and never reach resolution.

Scripting vs programming is like that. Of course there's no difference.
Today's machines can run interpreted code at rates that surpass compiled
code of just a few years ago. But you can debate the differences till the
cows come home and like my friends at WWW9 never reach resolution. Having
been a vendor of scripting systems for 13.5 years now, I've been around this
block. It's like the difference between a weed and a flower.


PS: Yes, Perl is a scripting language. And Python and Javascript.  ;->

----- Original Message -----
From: "Jeff Bone" <>
To: "Dave Winer" <>
Cc: <>
Sent: Saturday, August 25, 2001 11:18 PM
Subject: Re: REST Questions

> Dave Winer wrote:
> > For me it's all
> > about cross-network scripting, nothing more or less. Dave
> I'm with ya, Dave:  it's hugely important.
> So Dave, here's a question for you:  what constitutes "scripting" vs.
> "programming"?  How does a scripting language differ from a programming
> language, from a usage perspective and from an (internal) implementation
> perspective?
> Is Perl a scripting language?  Python?  Javascript?  Is scripting just
> synonymous with "interpreted?"  Is any language that can be used
interactively /
> has a REPL a "scripting" language?
> I'm leaning towards a POV these days which says that true scripting
> focus on higher-order primitives for composition / integration /
> of coarse-grained components.  By that definition, I'm not sure that the
> languages, for instance, are truly "scripting" languages;  they focus more
> providing primitives for defining and operating on domain abstractions at
> finer-grained level than, say, the Unix shell.  Not that a scripting
> couldn't *have* those things, just that they might need higher-order
things like
> | and friends.  A scripting language by this definition might in a sense
> regarded as a coordination language.
> Sigh...  definitional craziness. :-)
> Just some random thoughs.  I'm scouring through a bunch of old reuse
> right now, trying to piece together a train of thought...  I've got this
> hunch that there's something really fundamental and quantifiable in all
> REST etc. discussion that speaks directly to qualities and characteristics
> system architecture, modeling, programming paradigms, etc. which
> influence software reuse.
> jb