Re: Ken's G spot (was Character for e, 2.71828...)

From: Kenneth Whistler (kenw@sybase.com)
Date: Mon Apr 08 2002 - 16:19:06 EDT


Tex,

> I was with you up to this last comment.
>
> I can't think of any reason why you *would* want the constant G to match
> with values for G in word searches.

My point was that if I was looking for "G"'s, hoping to find them on
the assumption that they were encoded as "G"'s, i.e. U+0047, along with
"G"ravity and "G"oogle, I don't want my searches to fail, because,
oops!, the gravitational constant "G" was also encoded as something
else that looks the same, but has a different number. (Encoding number,
not constant value. *hehe*)

>
> If I wanted to find the use of gravitational formulae in my database or
> on the web, narrowing the search for G makes a lot of sense.

Of course. And this is part of what is driving this undercurrent
of people wanting special symbols to represent these constants --
on the assumption that if such a symbol was added, everyone would
use it, and thus searching for it would be easier. But in fact,
the net result would be everybody would use some unspecified
mixture of the regular character and the new symbol, because they
can't really tell them apart, and you'd end up having to search
for both, anyway, in any open context like the Internet.

If you have a closed context, like a well-defined database, then of
course, you can use whatever convention you like, and symbols can
be uniquely defined -- even using the current characters, without
having to add constants to the Unicode Standard.

>
> Try searching for "GMm" in google, since this is a common element in
> gravitational formulae and see how many non-gravitational responses you
> get.

Sure, lots of trash, since "GMM" is a common abbreviation for lots
of things. But "GMm gravity" does a whole lot better. Although even
there it turns out that "GMM-1" stands for "Goddard Mars Model 1" ;-)

> I think it makes sense for Unicode to not have a unique character for
> every possible constant or semantic nuance that can be invented, but
> lets not confuse that with the functionality people would want to have.
> A reasonable solution is to capture the semantics in markup or other
> ways.

Correct. Just don't assume that that will make Internet searching
any cleaner. :-)

>
> People should recognize that plain text has some limitations. One of
> them is the lack of stylistic differentiation. Another is the inability
> to distinguish specialized constants from other alphabetics.

No argument there.

--Ken

>
> tex
>
> Kenneth Whistler wrote:
> > In fact, there
> > is every reason *not* to have separate characters for them. I
> > don't want a "G" for the gravitational constant not to match
> > on a text search with the "G" in "Gravity" or "Google".



This archive was generated by hypermail 2.1.2 : Mon Apr 08 2002 - 17:06:18 EDT