[l2h] Re: Guillemets and OE

[Ross Moore]

> On Thu, 9 Sep 1999, Ross Moore wrote:
> 
> > > How are they coded in LaTeX? It should be fairly easy to provide an
> > > appropriate conversion as they are part of the ISO-Latin-1 character set.
> > > Ross?
> > 
> > No, they are not in Latin-1; that's the whole problem.
> 
> RU sure? See http://www.w3.org/TR/REC-html32.html#latin1 lists (HTML 3.2)
> 
> <!ENTITY laquo  CDATA "&#171;" -- angle quotation mark, left -->
> <!ENTITY raquo  CDATA "&#187;" -- angle quotation mark, right -->
> 
> These look like << and >>.

Yes, the guillemets are in Latin-1 and other charsets,
including Latin-9.

It is the \oe and \OE that are not, but are now in Latin-9.

If you have a single keyboard which allows << and >> as a single
keystroke, then these now stay as this character (provided it
is at the correct code-point) with LaTeX2HTML.

More generally, if you use  \inputenc{latin2} because your key-board
makes it easy to type Eastern European characters (e.g. Polish)
then the HTML pages will use the raw 8-bit characters wherever
possible, and mark the charset as  iso-8859-2 .
The parameter entities  &#<num>; are only used for characters
that do *not* exist in Latin-2.

Formerly parameter entities were used for all non-ascii characters,
but Mariusz and Alan Flavell convinced me that this was not best,
especially for Europeans where Latin-2 is much more commonly used
than Latin-1.
(unless the unicode.pl extension was also used, there would be too
many images; but older browsers don't support enough of utf-8,
so even this has significant problems)


Can anyone comment on whether Netscape 4.6 is any better than earlier
versions for support of character sets ?
(and stylesheets, for that matter ?)


Cheers,

	Ross Moore