Genuine computer Scrience >Re: Autocoding project proposal.

phil hunt philh at comuno.freeserve.co.uk
Tue Jan 29 12:49:15 CET 2002


On Mon, 28 Jan 2002 18:47:45 -0500, Christopher Browne <cbbrowne at acm.org> wrote:
>Stefaan A Eeckels <Stefaan.Eeckels at ecc.lu> writes:
>> On Mon, 28 Jan 2002 23:08:04 +0000
>> philh at comuno.freeserve.co.uk (phil hunt) wrote:
>>> Doesn't computer science deal with HCI and making computers easier
>>> to use, then?
>
>> Maybe it's in the curriculum. If the current crop of programs is
>> anything to go by, they could as well have left it out.  In any
>> case, it's about people, and how they react to devices (soft or
>> hard). It's psychology, not comp.sci.
>
>Certainly HCI isn't part of _pure_ computer science,

Well it ought to be. The main problem with computers is getting them
so that humans can interact with them more easily.

> and connecting it
>requires a conscious connection to psychology.
>
>It surely doesn't fall into the traditional areas of CS:
> - Algorithms
> - Numerical Analysis
> - Databases
> - Languages

AFAICT languages are *purely* an HCI problem. Machine code is Turing-
complete; therefore the *only* reason to have other languages is to make
it easier for humans to use.

Databases are special cases of data structures; again the only reason why
RDBMSs exist is because it's easier for people to do it that way than to
code it trhermselves in assembler.

Algorithms and data structures are what languages do, so are strongly
connected with the subject of languages.

>The Association for Computing Machinery, "the" CS organization, does
>have an HCI group (called SIGCHI).  It's only one of many special
>interest groups.
>
>I'd argue that HCI is _properly_ a secondary concern in computer
>science.  
>
>Computer science is about understanding what computers are,

They are physicval realisations of turing machines. End of discussion;
because that's all the discussion needed, and everything else they are
flows from that (or is a special case of I/O).

> what they can be made to do, 

anything, constrained by hardware limitations and the imagination/skill
of the programmer.

>and how to accomplish those things.

by paying people like me a lot of money, hopefully!

-- 
===== Philip Hunt ===== philh at comuno.freeserve.co.uk =====
* strong public-key email encryption  * automatically decrypt incoming mail
* automatically encrypt outgoing mail, including attachments and headers
* automatically propagate your public key and manage others' public keys
What encryption package does all this? Herbrip, part of the Herbivore
project. <http://www.vision25.demon.co.uk/oss/herbivore/intro.html>






More information about the Python-list mailing list