Re: How Lisp's Nested Notation Limits The Language's Utility
- From: Kent M Pitman <pitman@xxxxxxxxxxx>
- Date: 09 May 2007 11:41:16 -0400
[ comp.lang.lisp only.
http://www.nhplace.com/kent/PFAQ/cross-posting.html ]
David Hansen <david.hansen@xxxxxxx> writes:
On Wed, 9 May 2007 16:24:41 +0200 Kjetil S. Matheussen wrote:
For larger mathematical expressions, I think s-expressions are much
better because you can use more lines and indentations to make
things clearer.
Show me one person that uses sexps to scribble large equations by hand.
There's a fallacy in your question because it presupposes that large
expressions are constructed by scribbling. They often aren't.
Constructing large expressions by scribbling is error-prone and not
best practice.
It also supposes that the people who do this construction by hand do
the full set of operations on their expressions that we do on ours.
They may do some of that symbolic rearranging in their head, but they
may also just deny themselves options that are not easy to administer.
In this regard, it reminds me of writers who stubbornly resisted word
processors for so long, claiming they preferred fixed pieces of paper
for editing. It seems to me objectively clear that the computer /
word processor is a better device because of its ability to save old
copies, to compare copies, to rearrange large blocks, and in every way
to enable "trying alternatives" to make sure what you have is good.
Those who manipulate symbolic algebra without benefit of a mechanical
tool to assure the integrity and validity of the transforms they do are
really just hampering themselves.
Infix notation is, objectively, a device for compact depiction of a
result, and for people who don't want to type a lot. The compact
display has its purpose, and compact input may have its purpose, but
ultimately, trying to argue that there's no purpose in having other
display forms is like trying to argue that mathematical transforms
have no purpose.
Or, if you prefer a more obscure analogy, infix algebra is like
clinging to non-realtime TECO after WYSIWYG editors came along.
[I'd use the the correct names of the tools here, but it would obscure
my analogy, which is obscure already.]
Another problem problem with infix is that its irregular rules of
presentation obscure structural regularities that would probably
produce great insights, but that are hidden by the non-parallel nature
of the notation [in the English grammar sense of parallel
construction; no allusion to computational parallism here] .
Also, there are lots of people that use parens around algebraic
notation because they don't understand and trust "precedence".
(((a*b)+c)/d)
is not an uncommon style for people to adopt.
And, there are people like me who don't like doing math at all. I
prefer to hand it over to a computer to do it for me. Algebraic
notation is fine when the task I want to farm out is just "compute
this". But it's not when the task is more complicated.
And, also, even when editing things myself, I frequently find that
it's hard to safely and reliably rearrange an expression to understand
it. If you see an algebraic expression that goes on for pages, it's
nearly unintelligible. (I deal with symbolic algebra systems all day
for work and I have this problem every day. I have to stop and indent
the expression Lisp-style in order to make any sense of it, and that's
a complicated and tedious process.)
.
- Prev by Date: Re: Does ANSI Common Lisp have pattern matching?
- Next by Date: Re: Does ANSI Common Lisp have pattern matching?
- Previous by thread: Re: How Lisp's Nested Notation Limits The Language's Utility
- Next by thread: Re: How Lisp's Nested Notation Limits The Language's Utility
- Index(es):
Relevant Pages
|
