Re: HTML - i18n / NCR & charsets

Drazen Kacar (
Wed, 27 Nov 1996 09:39:04 -0800

Message-Id: <v03007803aec22d21c55e@[]>
Date: Wed, 27 Nov 1996 09:39:04 -0800
From: (Drazen Kacar) (by way of Walter Ian
Subject: Re: HTML - i18n / NCR & charsets

Walter Ian Kaye wrote:
> At 6:03a +0100 11/27/96, Drazen Kacar wrote:
> >
> >But you didn't consider cheesy little editors (I don't mean Emacs).
> >They will do it for you.
> Or "to" you? <G>

Hmm... To somebody else, actually. Somebody who doesn't use the same

> >On Windows, in particular. When you're not using
> >Latin 1, they will assume you do, because, what else could you be using?
> >
> >And they will insert entities, tons of them. Some of them will be in the
> >128-159 range. These will be numeric, others will be simbolic.
> Well, that's not an "existing practice" problem, that's software bugs which
> should be reported to the developers of those..uh.."cheesy little" editors.

But it is. I want Latin 2 page. I can use my cheesy little editor to
generate it. I'm on Windows and I use windows-1250 code page. That's not
ISO 8859-2. There are letters on 128-159 positions. I write some text
that uses them. The editor puts entities in place of every character which
has code greater than 127. It assumes ISO 8859-1. Symbolic Latin 1
entities are put for 160-255 positions and numeric entities (again Latin 1)
are put for 128-159 positions.

Now I use my crummy little web browser to preview the page. I don't have
the faintest idea about charset parameter in content-type header nor
about META tag. So I tell my crummy little web browser to use Latin 2 page.
3.0 version is a bit confusing. It can deal with ISO 8859-2 and windows-1250.
But I know what everybody use. I tell the browser to use windows-1250.

My crummy little browser has some bugs that I'm unaware of. It will map
numeric entities to current code page. More than that, for symbolic
entities it will do a table lookup, always find the same number and show
the character in the current code page.

Everything works! As long as I'm on Windows. Bug in crummy little web browser
negates the bug in cheesy little HTML editor. The web is full of these things.

> How about a "certification" service of some sort, to award a "badge" to
> those editors which output valid HTML? (There won't be too many awards!)

I wonder if there'd be any...

Life is a sexually transmitted disease.