Re: several messages about Netscape and charsets


From: guest <GUEST@DAHLIA.BITNET>
Subject: Re: several messages about Netscape and charsets
Date: Mon, 4 Dec 1995 20:38:14 +0100

Next Article (by Date): Re: several messages about Netscape and charsets Hynek Med
Previous Article (by Date): netscape character set Thomas R Rocek
Next in Thread: Re: several messages about Netscape and charsets Hynek Med
Articles sorted by: [Date] [Subject]


Please ignore the From: header on this message.  The LISTSERV processor
loses the Reply-To: field which I have, which should be <barryb@tuke.sk>
for messages not intended to be posted to the whole list.  Sorry...
 
On Mon, 4 Dec 1995, Hynek Med wrote:
 
> On 3 Dec 1995, Barry Bouwsma wrote (discussing META MIME charset
  tagging and Netscape 2.0 releases under X Windows):
 
> >     When these new versions of Netscape encounter such a document, they
> > will attempt to display it with a Latin-2 font, provided one is availab=
le.
>=20
> How does it select the font from the available Latin-2 fonts?
 
    You can see this in the Preferences:  Under the pull-down menu
``Options'', the top item -- ``General Preferences'' (the language
encoding is something different which I shall mention later).
    Within the General Preferences pop-up window, the second folder is
called ``Fonts''.  Look to see what is there.
    You will see an encoding.  This will probably be the western Latin-1
encoding.  If you click on this box, a menu will appear with all the
charset encodings for the fonts which are available on your system.
You want the Central European (Latin-2) encoding.  Select that.
    You will then see what fonts Netscape has chosen to use as the
default when displaying documents with this type.  You can select from
the pop-up menus provided by these boxes, if you don't like the default.
Save your preferences if you change them.
    These fonts are found by searching the list of available fonts for
those which use the ISO 8859-2 encoding, then Netscape picks two.
 
    If the Central European encoding does not appear, that means that
Netscape was not able to find any fonts with the ISO 8859-2 encoding on
your system.  You can check this by listing the fonts...
 xlsfonts | grep 8859-2
    If no fonts are listed, that means your Xserver has no Latin-2 fonts
available -- they were not installed, or not properly installed.  If this
is the case, you can add fonts to your font path from a font server -- we
have the fonts found on our FTP server available, which can be added
with the command...
 xset fp+ tcp/fs.vszbr.cz:7000
    If you got no error, the xlsfonts command should list about half a
dozen fonts which match.  (This will not work with X11R4.)
 
    Provided you have ISO Latin-2 fonts, Netscape2.0 will then be able to
support automatic switching from the default charset encoding to Latin-2,
*provided* that the document describes itself with the META tag for
MIME charset that I mentioned earlier.
 
    Note that the item under the Options pull-down menu entitled
``Language Encoding'' has nothing to do with this.  What this does is to
define the default charset assumed to be used by Web pages.
    The HTML spec specifies ISO-8859-1 to be the default encoding.  That
means, in the absence of any charset specification in the document, it is
assumed to be Latin-1.  Of course, this is useless in countries where
Latin-1 encoding is not used, such as in SK/CZ, so pages have been created
with a number of different byte->character mappings to match the displays
available to the user.  Until now, there has been no client support for the
META charset tagging, so the best one could do was to use a different
default font/charset-encoding.  This is what is defined here.
    When a document without tagging is encountered, the encoding
specified here will be that which is assumed to be in use.  For browsing
documents in these countries, some Latin-2 encoding can be assumed, so if
all you do is surf in the area, you can define your encoding to be
Central European and see everything correctly when viewing a document
delivered to your browser with ISO-8859-2 encoding.
    If you spend time outside, such as in Germany, France, Scandinavia,
and so on, those pages will use the Latin-1 encoding.  Unfortunately,
most of those pages will be without the META tagging, and will therefore
display using the default character set you've specified, so you may have
to change this encoding on-the-fly as you view documents to match the
encoding in use.
 
    As I mentioned, it would be nice if all documents were to include a
tagging to guarantee that the displayed text matches.  However, the
burden falls on non-ASCII or non-Latin-1 documents, which would include
virtually all Czech/Slovak pages with diacritics.
 
 
> We have
> these free (monospaced) ISO-Latin2 fonts on our Linux machine, and when I
> tried it, it didn't do anything. Well, that's probably because I don't
> have the <META> tags in documents..
 
    Exactly.  In the absence of a META tag describing the charset, it is
assumed to be Latin-1, or more correctly, it is displayed with the
default Language Encoding which you have selected.  In order for this
automatic switching to work, the META tag mentioned in my earlier
message must be included in the document.
 
    There are two documents to which I have added this META tagging, one
shamelessly stolen from another server and to which I merely added the
tag, which can be viewed (it's short) at...
http://www.vszbr.cz/~guest/czdocs.html
    And the one I first mentioned, at...
http://www.vszbr.cz/~guest/
 
    These two documents can be used to verify that this Netscape2.0
feature works.
 
    If you want the ISO Latin-2 documents which you serve to always
display correctly when viewed by this latest Netscape, then you must add
this tagging.  Without it, it will not work automagically.
 
 
>  (But wouldn't these tags confuse
> browsers, which use non-ISO-Latin2 fonts, like Netscape for MS-Windows?
> S =E8e=B9tinou je to v=BEdycky hr=F9za..)
 
    No more than delivering the document without the tagging.  Ideally,
the client (Netscape) can perform the mapping from the over-the-wire
encoding defined in the META MIME charset tag, to the display font which
is in use locally.  But I don't know if the ISO-8859-2 -> CP1250 mapping
is in place in MS-Win Netscape, likewise with Mac Netscape.
 
    All I know about this META tagging was obtained from the Netscape
Release Notes, or News, or help, or wherever.  I do not know which
charsets are understood for mapping by Netscape, nor have I looked at
the spec to see if it is permissible to tag the alternative Latin-2
encodings with the tagging from the Assigned Numbers RFC.  It could well
be that in order for a given tagging to be displayed properly, you must
have a font with that encoding available, but I really have no idea.  I
guess it's time for me to play with this...
 
    In any case, I can't see that it would hurt, if you are serving pages
with MS-Win Latin-2 encoding, to tag them with the charset value of
CP1250.  After all, that is correct, because this tag describes what is
used, which is better than guessing, and a smart client can use mapping
tables, or GNU recode, or whatever, to convert the encoding to the local
display.
 
 
 
On 4 Dec 1995, Thomas R Rocek wrote:
 
> Does anyone know how to get the character set coding to work in the 1.x
> netscape versions?  If I select czech fonts in the preferences sections,
 
    Netscape 1.* does not support charset tagging, and is only aware of
the western Latin-1 encoding, ASCII, and Japanese.
    Also, in the two versions of Netscape 1.1 that I tried to run, I
saw no possibility under Preferences to select the font.  (I may have
overlooked something, and my window is too big to fit the screen.)  In
one of them, I saw a variety of encodings, the default from which was
Latin-1, but this was not in the other; both Unix Netscape.
 
 
> netscape allows me to do this, but keeps Latin1 coding (it offers only
> Japanese coding as an alternative).  So, when I view czech documents,
> accented characters are replaced by blank boxes.
 
    Netscape probably searches for a specified font, but only if the
description of the font matches the selected coding (that is, if the
string ends in 8859-1 for Latin-1).  So, it's going to find a Latin-1
font -- except not all fonts tagged with 8859-1 are really Latin-1.
Netscape may have had to use an ASCII font.  These are just guesses.
 
    I would have to ask what documents you are trying to view (to
determine the encoding used), and how you are running Netscape.  But I
think it would be easier to obtain a beta copy of Netscape2.0 for viewing
non-western-European language documents, because a minimum of
configuration would be necessary, and everything should be automatic with
properly-tagged source documents.
 
    I'd be interested to hear the results of non-Unix Netscape when
viewing tagged ISO-8859-2 documents...
 
 
Barry Bouwsma
<barryb@tuke.sk>
visiting MZLU Brno, CZ

Next Article (by Date): Re: several messages about Netscape and charsets Hynek Med
Previous Article (by Date): netscape character set Thomas R Rocek
Next in Thread: Re: several messages about Netscape and charsets Hynek Med
Articles sorted by: [Date] [Subject]


Go to listserv.cesnet.cz LWGate Home Page.