Re: The case for lower case


"Robert Redelmeier" <redelm@xxxxxxxxxxxxxxx> wrote in message news:6x_Ce.355


> > The question of GetProcAddress vs GETPROCADDRESS seems
> > clear cut; but what about source code that uses lowercase
> > (getprocaddress) vs camelHump (getProcAddress) vs the MS
> > CamelHump (GetProcAddress)? Alternatives include get_proc_address,
> > get.proc.address (where permitted, but common in object compound words).
>
> I prefer Caps on all words as a form of separator, but only
> for variable names. fns are all lower, and macros/constants
> are all UPPER.

That's C style, not good writing style.
It is amazing how a backwards typographical convention from the early days
of computing still persists today. All uppercase is hard to read. Period.
Indeed, that's the point -- to slow people down and make them pay more
attention (just like boldface type, for example).

Lower case letters contain more visual "information" than upper case
characters. As a simple experiment, try the following: take a *** of paper
and cover up the top (or bottom) half of a row of all upper case characters.
Look at how hard it is to read. Now repeat the experiement with all
lowercase characters. You'll actually be able to make out the words in most
of the sentence. Actually, you don't even have to go that far. Just reread
this paragraph in all upper case and see how easy it is to read:

LOWER CASE LETTERS CONTAIN MORE VISUAL "INFORMATION" THAN UPPER CASE
CHARACTERS. AS A SIMPLE EXPERIMENT, TRY THE FOLLOWING: TAKE A *** OF PAPER
AND COVER UP THE TOP (OR BOTTOM) HALF OF A ROW OF ALL UPPER CASE CHARACTERS.
LOOK AT HOW HARD IT IS TO READ. NOW REPEAT THE EXPERIEMENT WITH ALL
LOWERCASE CHARACTERS. YOU'LL ACTUALLY BE ABLE TO MAKE OUT THE WORDS IN MOST
OF THE SENTENCE. ACTUALLY, YOU DON'T EVEN HAVE TO GO THAT FAR. JUST REREAD
THIS PARAGRAPH IN ALL LOWER CASE AND SEE HOW MUCH EASIER IT IS TO READ.


This isn't brand new knowledge we're talking about here, this is information
that typographics have known about for a couple hundred years. It is
amusing listening to C programmers talk about how uppercase is easy to read,
based only on a lame style convention that Ritchie invented back in the days
when people didn't really know what readable code looked like.

As for monitor foreground and background colors, the easiest to read page is
one that comes closest to matching the 18% gray scale that are eyes are
accustomed to seeing in nature. That is, 72% white and 18% black. Easily
achievable by putting black letters on a white background. Difficult to
achieve (in a readable form) by putting white letters on a black background
(too much contrast). This, in fact, was the reason for green screen monitors
in the first place -- it reduced the contrast. Believe me, if reading white
letters on a black background were truly more readable, that's what people
would be looking at. Xerox, Apple, IBM, and Microsoft wouldn't have
universally adopted the paper paradigm if they could achieve better results
by changing the colors around. There's too much at stake in the desktop
competition. If white on black were truly more readable, at least one of
them would be pushing that as an alternative; as it is, about the only ones
you see doing black backgrounds are Linux hackers who've never done a
readability study in their lives ("Hey, I grew up on a greenscreen terminal
and all UPPERCASE CHARACTERS, it *must* be more readable!").

Vantana Press had some real good books on typography and such. This was
about 15 years ago, so they might not be around anymore. But look for books
like "Type From the Desktop" and "Newsletters From the Desktop." They've got
a lot of real solid advice for producing readable text. And all those rules
apply fairly well to source code as well as English text.
Cheers,
Randy Hyde


.

Loading