[REBOL] Re: Periods as parts of rebol words?
From: greggirwin:mindspring at: 11-May-2002 11:15
Hi Joel,
<<
> > Technically, you can do it. I personally don't think it's a
> > good idea, for the following reason: REBOL is designed to be
> > a good language for humans.
>
I just don't buy this one, for a whole list of reasons, including
the following samples:
>>
Let me see if I can clarify my point a bit.
<<
* There's nothing new about idea of having a programming language
that is supposed to resemble natural so closely that humans of
the non-programmer species can write code.
Nonsense.
>>
I didn't say REBOL was intended for non-programmers, or that any resemblance
to natural language in any way endowed people not inclined to study the
craft with the ability to write code. That's what dialects are for. :)
<<
The same kinds of fatuous claims have been made for BASIC,
SQL, spreadsheets and spreadsheet macro languages, and on and
on and on... The claim was also made about FORTH that, as an
extensible language, one could bootstrap up to a set of defined
words that end users could employ as an interactive or scripted
command language.
This "impossible dream" has never worked, and IMHO never will.
Humans can't write real, non-trivial programs without actually
learning programming skills. Sugar-coating that fact is at
best a waste of time and at worst deceptive.
>>
Hmmm. I guess I like sugar coating. OTTOMH, I might say a programming
language could be considered good for humans based on the number of
non-programmers that are able to use it successfully.
My justification for this is that, even though the holy grail of 'not
requiring a computer programmer to program a computer' is an ideal we may
never achieve completely, you have to admit that a lot of non-programmers
are writing programs that do useful stuff these days, generally in
sugar-coated languages.
<<
* There are so many factors in the design of REBOL that already
violate good human factors practice that the issue of dashes
versus periods as intra-word punctuation just doesn't seem to
be that significant. Such issues as ensuring referential
transparency, consistency, the principle of least surprise,
proportionality (things that make a big difference should look
really different, things that don't make much difference
shouldn't) and on and on and on...
>>
I think it *is* significant (and I disagree with the argument that "it's got
other problems so this won't make any difference"). The syntax is very
important and is a separate issue from the semantics and behavior of a
language. That said, I'm not saying that my view is the "right" way to see
things for the example under discussion, just that it's my view. As you
said, pick a style that you're comfortable with and that won't confuse the
target audience.
If there are standards in place, it's generally a good idea to follow them
(assuming they make a modicum of sense). If you take a different path, have
a reason.
<<
Of course, if we're going to take this resemblance to natural
language seriously...
>>
I obviously created some confusion with my post in this regard. REBOL is,
indeed, an artificial language. I'm not suggesting that it should strive to
emulate natural language so much as it should strive *not* to emulate other
artifical languages. ;)
--Gregg
