What computers are for (was: Perl is to stupid ...)
[1/2] from: chris:starforge:demon at: 18-Dec-2001 10:11
Carl Read wrote:
> Perhaps, instead of trying to make software understand documents
> written any old which way by humans, we should create strictly formal
> versions of current human languages that can be tested for
> correctness by computer? We'd then be able to have documents that
> could be examined by computer without the need to worry about an
> infinate number of special cases.
I hope this is a tongue-in-cheek comment! If not, you are making the fatal
error that practically every computer scientist has made at one point or
another: expecting that human beings will modify their behaviour to suit
the computer rather than the opposite. This isn't impossible
(voice-recoginition software was, and to a large extent still is, a good
example of this), but it is far from ideal. It appears to me that the
computing industry seems to have lost sight of the real purpose of
computers: they are (or at least were) supposed to make life easier for
the user. It's more than a bit of a kludge when you tell a user "this
program can understand your documents, but only provided that you write
them in this very constrained, artificial form which allows little of the
form and structure you typical documents contain".
It's a bit like expecting a user to learn C just so that they can type in a
letter! (not that I'm complaining about C you understand - I use it every
day - but it's hardly something you can expect the average user to learn)
No, either we solve the problems caused by rule set size without forcing the
user into overly artifical situations or automatic document parsing will be
constrained to simplistic and case-specific situations.
Chris
--
.------{ http://www.starforge.co.uk }-----. .--------------------------.
=[ Explorer2260, Designer and Coder \=\ P: TexMaker, ROACH, site \
=[___You_will_obey_your_corporate_masters___]==[ Stack: EETmTmTRRSS------ ]
[2/2] from: rpgwriter:y:ahoo at: 18-Dec-2001 10:16
--- Chris <[chris--starforge--demon--co--uk]> wrote:
>
> I hope this is a tongue-in-cheek comment! If not,
> you are making the fatal error that practically
> every computer scientist has made at one point or
> another: expecting that human beings will modify
> their behaviour to suit the computer rather than
> the opposite.
That always happens with technology, to a degree,
so I'm not sure why it shouldn't happen with
computers. Consumers always expect to minimize
their changes, and often those making the tool
expect them to make more than is reasonable;
reality ends up being somewhere in between.
> This isn't impossible (voice-recoginition
> software was, and to a large extent still is,
> a good example of this),
I'm not sure it is; the requirement to change
behavior is one of the (many) reasons no one
I know personally uses voice recognition software.
> but it is far from ideal. It appears to me
> that the computing industry seems to have
> lost sight of the real purpose of computers:
> they are (or at least were) supposed to make
> life easier for the user.
Different tools *usually* require changes
of behavior and still make life easier.
> It's more than a bit of a kludge when you
> tell a user "this program can understand
<<quoted lines omitted: 3>>
> of the form and structure you typical
> documents contain".
Perhaps it is. Its better than nothing at all,
and may still have use in limited markets,
though. Arguably, unless your program is
smart enough to understand meaning in
language -- a strong AI -- its not going
to be as good as a human at recognizing
the content in documents. That's just a
given.
> It's a bit like expecting a user to learn C just so
> that they can type in a letter! (not that I'm
<<quoted lines omitted: 6>>
> constrained to simplistic and case-specific
> situations.
Well, sure, I'd suspect the latter will remain the
case for quite some time.
Chris
Notes
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted