Re: unicode Digest V2 #167

From: Otto Stolz (
Date: Fri Jul 05 2002 - 10:25:07 EDT

I had written:

> You don't need web fonts; it is enough to have installed fonts for the
> languages you can read -- which you normally would have, anyway.

Michael Jansson wrote:
> Your average Internet user may not have a suitable font, may not
> know how to install a font, may not be allowed to installed a font (e.g.
> would have to ask their helpdesk at work), nor would not know where to find
> a font.

Your average user may not have a suitable browser, or no suitable browser
settings, or not a suitable operating system, or not a suitable display,
or not a suitable windows size or resolution, or not some other
to make use of a proprietary system as WWW fonts are, currently. I have
tried to convey this thought with the following words Michael chose not
to quote:

> Web fonts do not work with all browsers alike, as there is no widely
> accepted standard. Rather, two market-leaders have their proprietary,
> mutually incompatible, mechanisms. We have seen that indic example which
> tried clever tricks to select a web-font mechanism appropriate for the
> respective browser -- and still failed to provide my (fairly standard)
> browser with a font to display that page. Even if you do better (server-
> side mechanisms rather than Javascript), you will never be able to cover
> all browsers in the field (i. e. all browser brands, all versions; all
> operating systems, all configuration settings; all display types, re-
> solutions, and windows sizes; ...).

Michael Jansson wrote:

> Also, I would need several different fonts for a variety of platforms

> and you web page have to be encoded differently for different fonts.

No, nay, never. The extend-repertoire-by-fonts-switching hack is
bound to fail, in many ways. Basically, you are lying to the reader
about the content of your HTML page. Imagine an "acoustic" browser
that reads the page aloud to the user; how will it cope with that
font hack? For more arguments and examples, cf.

Rather, the WWW page will have to be coded in Unicode, as the standard
sayeth, cf., e. g., <>.

> Also, telling people to make changes to system files causes problems.

Most people will have fonts for their native language installed, anyway.
A user interested in other languages, and capable of reading them, will
certainly be interested in having his computer configured, accordingly.
So, they will either buy their system cofigured suitably, or they will
learn how to do it by themselves.

> You don't base a serious commercial web on such a solution, because people
> won't use it. A solution for the public have to just work. Browsers should
> support Unicode *and* web fonts, intelligently.

And while they don't, you would not base a serious commercial web on such a
solution, either.

Clearly, Unicode is the way to go. Modern browsers support it quite well.

A couple of weeks ago, I needed a Unicode example when I was not in my
office; I asked a local secretary whether I could use her computer, fetched
me <> and printed the page. The
Latin, Georgian, Hebrew, Arabic, Greek and Cyrillic examples were displayed
and printed properly. Classic Greek hat two characters replaced with boxes;
Far-East, and Nunakut examples were missing. If I remember correctly,
and Ethiopian were also missing, whilst Armenian was there. Granted,
that this
was a standard Windows + IE/Netscape installation, I deem this result fairly
impressive. If that secretary would be able to write, e. g. Japanese, she
would certainly have the Japanese IME installed, and that would include a
Japanese font, so the browser would display the Japanese examples, as well.

Best wishes,
   Otto Stolz

This archive was generated by hypermail 2.1.2 : Fri Jul 05 2002 - 08:45:01 EDT