Re: Traditional computer science vs. Human-oriented computer science
- From: yaoziyuan@xxxxxxxxx
- Date: 16 Sep 2006 07:12:24 -0700
A good case to study the complexity of a human-involving process is the
Chinese PinYin input method, especially Microsoft's sentence-oriented
Pin Yin input method (MSPY).
In MSPY, the most annoying part is when your pinyin expression for a
Chinese character pronunciation can have multiple corresponding Chinese
characters (usually 30 - 400 of them) and therefore a manual resolution
of such ambiguity is time-consuming. There are two main solutions: (1)
the user temporarily inputs a multi-character word (a collocation of
two or more Chinese characters that is regarded as a single word) which
includes the actually desired Chinese character; this dramatically
reduces the ambiguity to just one or two choices of eligible words; and
then the user deletes the other Chinese characters in the word and
keeps the desired character; this approach has a drawback: When the
user is quickly expressing his ongoing thought, deliberately making a
phrase for extracting the desired character can brutallly interrupt his
thought. This should be deemed a severe penalty of the approach's
usability; (2) The many candidate characters (all of which correspond
to the inputted pinyin expression) are displayed in multiple rows, each
row corrresponding to a common characteristic, such as a common tone
that all the characters in that row belong to. This reduces the total
scan time a user has to spend for finding his desired character, but
still entails more character reads than approach #1.
yaoziy...@xxxxxxxxx wrote:
At the kernel of this human-oriented computer science is a mechanism to
measure how efficient a human-involving process' performance is
(analogous to the big O notion) and how much new value a person's brain
gains after undergoing an execution of a process.
There are two attitudes toward a person's brain's role in a
human-involving process: (1) the human brain as a tool (or machine,
component) of the whole process; (2) the human brain as a speculation
(high risk high return) platform where various idea-creating methods
can be experimented.
Since this human-oriented computer science involves only the computer
world and the human psychological world, we can consider it a "pure
science" as a next wave of computer science. I predict its prime time
will last 10-20 years, just as the most classical discrete algorithms
(routing, sorting, scheduling, matching, flows, etc.) were all invented
during 1950's - 1970's.
Example system problems with a general application:
(1) If a person wants to find relevant information (i.e. prior art)
about a particular idea with a keyword-based search engine, how should
he plan his search strategies, various combinations of concepts, and
various keyword embodiments of a concept?
(2) How can a computer assist a person in writing/reading documents in
a non-native language so well that he would seem to be a native
speaker?
Maybe you don't like system problems since they're always big, complex
projects. Maybe you like simple, discrete data objects. Then, consider
the above "kernel" issues, such as "how to define the big O notion's
counterpart in human-oriented process performance analysis?"
yaoziyuan@xxxxxxxxx wrote:
Traditional computer science is mainly concerned with computer
programs' performance; I think a trend of current computer science will
shift more attention to human brain performance -- how to best maintain
and utilize a user's brain power.
The industrial era's main theme is the performance of machines; the
post-industrial era's new productivity won't come from machines doing
repetitive tasks, but from creative human brains perceiving new needs
and conceiving new ideas.
Yao Ziyuan
.
- References:
- Traditional computer science vs. Human-oriented computer science
- From: yaoziyuan
- Re: Traditional computer science vs. Human-oriented computer science
- From: yaoziyuan
- Traditional computer science vs. Human-oriented computer science
- Prev by Date: Re: Traditional computer science vs. Human-oriented computer science
- Next by Date: Re: Traditional computer science vs. Human-oriented computer science
- Previous by thread: Re: Traditional computer science vs. Human-oriented computer science
- Next by thread: Re: Traditional computer science vs. Human-oriented computer science
- Index(es):
Relevant Pages
|
Loading