ietf
[Top] [All Lists]

Re: [Re: Standartization of User Input for find\search engines]

2000-10-16 12:30:02
Alongwith standardizing the way the search engines take the search
query, I
feel that there should also be some standard with regard to specifying
the
context in which a search is being performed.
...

Let's also require that all natural language query systems use English.

In other words, there are good reasons why successful IETF standards
have generally avoid user interfaces....No, that's wrong.  It's not only
the IETF that has avoided successful user interface standards.  Anyone
who's coded with the zillions of GUI toolkits over the years knows that
"successful user interface standard" is an oxyomoron

Until the world is willing to standardize on a single dialect of
a single human-to-human interface, why should there be a single
flavor of human-to-computer interface?


I think you're taking my proposal to the extreme.
Nowhere did I define the search language, search interface or search results
(although results forms are a rather nice idea). and nowhere do I say that
the standard should be the only way things are implemented. research is
always a blessing, but,
when it comes to user interface and user input, It seems we tend to forget
that there are non-technically-oriented users, which a standard will be very
helpful (since they don't have to learn a new one).

These users will benefit extremely from a standard, since they'll have the
power to search more specifically or more generally, according to what they
want, and get more accurate results.

The common users will benefit from the things like the obvious-ness of
Boolean operations (always being OR\AND\NOT) and simplicity.

Sophisticated users can benefit from a further step like regular expressions
:
Say you want to search a family name in a DB. how would you do it?
6 different searches?
"pickard", "picard", "pikard", "peecard", "peekard", "peeckard" ?
or one search of a regular expression?
"p(ee|i)(c|k|ck)ard" - (perl style) ? (I doubt common users know where the
pipe sign is... (-:   )


or evaluated expressions:
"score:(\integer>5)"
(will bring results like "score:6" or "score:12" but not "score:4" etc')

(*the "search expressions" are just for illustrations)

These are, off course, more suited to special cases and special DB\indices
but clearly needed at some level, and there's lots to work with, so why not
have a standard? it'll clearly benefit the internet community. as in, us.
But it seems that IETF doesn't handle these kinds of things. maybe it
should.

    Cheers.