In a message dated 2001-04-27 4:22:51 Pacific Daylight Time,
karl-pentzlin@acssoft.de writes:
> 1bc> Why don't you make the next print edition of the Unicode
> 1bc> standard (not to mention online) with Unicodepoints
> 1bc> in decimal as well as hex?
>
> In fact, I do not see any reason to use hex numbers in documents
> released for the general public. In my opinion, future print editions
> should use decimal numbers *primarily*.
>
> Of course, hex numbers are a concept which all computer and coding
> experts know as good as ordinary (i.e. decimal) numbers. They have
> advantages if you have to discuss technical details of communications
> on the bit or octet stream level - but I cannot see any other
> advantage. They are legacy. They are a habit of the experts inherited
> from the time when 8-bit or 16-bit entities were strong constraints
> for code tables - now, the possible number of Unicode points is not
> even a power of 2.
>
> On the other hand, "numbers" (i.e. decimal numbers) are a concept
> everybody is familiar with. Thus, why not say to the public simply,
> "Unicode gives every character a number", instead of geek speak like
> "Unicode gives every character a code point, and as we are very cute,
> we use a special numbering system with 16 digits designed for
> computer experts, and to use Unicode you have to become an expert too
> and have to learn this system"?
Because the "experts" are the only ones who have to know this stuff.
When you talk about people who "use Unicode," or any other character set for
that matter, you are really talking about two completely different groups:
1. "geeky" programmers like me and many others on this list, who must deal
with the actual numbers (expressed in decimal, hex, or whatever), and
2. the general public, who does "use" the character set in a sense, but has
no need or desire to know the numbers.
For many in group 1, hex is handier than decimal for Unicode -- not because
we have stayed up nights learning to do hex multiplication and division in
our heads, or some other "get a life" endeavor, but because hex represents
the underlying structure of Unicode better.
For example, the combining diacriticals are in the block beginning with
U+0300. Calling it 768 is less memorable and sheds less light on what is
going on.
The BOM is U+FEFF, and the reversed BOM (which is not a character, but can be
important to identify) is U+FFFE. Referring to these as 65279 and 65534 is
no help whatsoever. (How long after reading this message will anyone even
remember the decimal number 65279?) Almost 20 years ago, I used to program
8-bit computers in their built-in BASIC languages, where decimal PEEKs and
POKEs were the order of the day, and believe me, memorizing numbers like
65279 is *much* geekier than using hex.
> If "character numbers" are at last used commonly, e.g. every Chinese
> businessman can spell his name to every secretary in the world, simply
> telling them the numbers. And every secretary can enter the name
> correctly without any knowledge to Chinese, having no more to learn
> than the single function key to be depressed during the digit input
> (which may be standardized for future keyboards). And (maybe the
> strongest argument) they have no chance to confuse decimal numbers with
> hex numbers which accidentally do not have digits in the A..F range.
But what Chinese businessman, or anyone else, is going to memorize the
numbers that spell his name, regardless of the base? I doubt seriously that,
for example, Karl thinks of himself as 76, 97, 114, 108 and can recite that
to his secretary (talk about geeky!). The Chinese won't have to do any such
thing either, as keyboards and input method editors and other user interfaces
improve to provide better support for Unicode and non-Latin scripts.
> The fact that these "numbers" have a relevance for the various UTF
> encodings is convenient for computer and coding experts. But for
> the ordinary user, it has no more relevance as knowing the cylinder
> head diameter for driving a car.
Again, it is not the drivers who have to know the numerical code points, only
the mechanics.
-Doug Ewell
Fullerton, California
This archive was generated by hypermail 2.1.2 : Fri Jul 06 2001 - 00:17:16 EDT