Re: Personal web pages: your suggestions...

Daniel Déchelotte wrote:

Hi,

I have recently reworked my little place on the Net and would like to hear
your suggestions. The site is here:
http://yo.dan.free.fr/

Some issues I tried to address, but maybe incorrectly: is the content
accessible, is the presentation coherent, is it "obvious" to get the
translated version (if available), or to find links (not underlined...),
and important details such as ALT texts, "prettiness", and so on.

A bit of warning: presentation relies on CSS and, as I am not an expert,
might break on older (and not so old) browsers; IE is served the same page
but uses a simplified version (thanks to conditional comments).

Looking forward to reading you...
Cheers,

Looks great to me! I like your icons.

mark
www.sixstringtheory.com

On Sun, 23 Jan 2005 03:27:46 GMT, "mark (AT) sixstringtheory (DOT) com"
<mark (AT) sixstringtheory (DOT) com> wrote:

Daniel Déchelotte wrote:
Hi,
I have recently reworked my little place on the Net and would like to hear
your suggestions. The site is here:
http://yo.dan.free.fr/

Some issues I tried to address, but maybe incorrectly: is the content
accessible, is the presentation coherent, is it "obvious" to get the
translated version (if available), or to find links (not underlined...),
and important details such as ALT texts, "prettiness", and so on.
A bit of warning: presentation relies on CSS and, as I am not an expert,
might break on older (and not so old) browsers; IE is served the same page
but uses a simplified version (thanks to conditional comments).
Looking forward to reading you...

Looks great to me! I like your icons.

Looks very good in my old Netscape 4.08, and loads very fast on a
dialup connection.

www.sixstringtheory.com
A bit slow for me, but interesting tabs.

Geo

Daniel Déchelotte wrote:
Hi,

I have recently reworked my little place on the Net and would like
to hear your suggestions. The site is here:
http://yo.dan.free.fr/

Some issues I tried to address, but maybe incorrectly: is the content
accessible, is the presentation coherent, is it "obvious" to get the
translated version (if available),

This is what I got in IE/Win XP:

Not Acceptable
An appropriate representation of the requested resource /index.phtml could
not be found on this server.
Available variants:

a.. index.phtml.en , type application/x-httpd-php, language en
b.. index.phtml.fr , type application/x-httpd-php, language fr

--------------------------------------------------------------------------------

Apache/ProXad [Dec 20 2004 12:37:37] Server at yo.dan.free.fr Port 80

--
Markus

On Tue, 25 Jan 2005, Daniel [ISO-8859-15] Déchelotte wrote:

| Not Acceptable
| An appropriate representation of the requested resource /index.phtml
| could not be found on this server.
| Available variants:
|
| a.. index.phtml.en , type application/x-httpd-php, language en
| b.. index.phtml.fr , type application/x-httpd-php, language fr

Ouch, bad. Could it be because your "accept-language" list contains some
(or no) languages that confuse the server?

I think you'll find that it's the content-type, not the language,
that's spooking it out. The client should never get to see the
"magic" content-type of application/x-httpd-php.

There seems to be something wrong with the server configuration, which
means that the content negotiation hasn't properly understood what the
final served-out content-type is going to be.

On Tue, 25 Jan 2005, Daniel [ISO-8859-15] Déchelotte wrote:

If that does not fix the problem, I could try and tinker my
.htaccess file

Sorry, I've changed my mind. My guess now is that MSIE is requesting
language en-US, and you haven't got an en-US version, only a generic
English version.

However, there -is- something wrong with the content-type,
nevertheless. The negotiation machinery apparently thinks this
resource is of type application/x-httpd-php and is negotiating it as
such. If a browser were to "Accept: text/html" (which is what you're
really going to send to the client, after all) and the client didn't
include */*, then the negotiation would conclude that
application/x-httpd-php was not an acceptable content-type, which
isn't what you want.

My hunch (this -is- Apache, and I guess you're using MultiViews,
right?) is that you should stop using the magic content-type
definition via AddType, and instead use AddHandler to tell Apache that
you want PHP called (and possibly also use AddType to define the
ultimate content-type of the PHP handling for MultiViews - or use a
type-map for the negotiation instead).

Why none of my test browsers has such a problem?

That depends on what they send in their Accept* headers. If it was
me, I'd put a custom logging directory into the server to make sure I
knew what the various browsers were sending in this regard. But there
are other ways, if you can't do that.

MSIE, if not explicitly configured to the user's preferences, has this
stupid tendency to reject all languages except en-US, instead of just
omitting the Accept-language header.

Daniel Déchelotte wrote:

Thanks to Markus for pointing out the problem and to you for the
explanation and a fix.

Yes indeed it works now for me, too.

Alan: It is actually de-ch, which of course makes no difference regarding
your technical explanations.

It's always nice to learn from other's problems :-)

--
Markus

On Wed, 26 Jan 2005, Markus Ernst wrote:

Alan: It is actually de-ch, which of course makes no difference
regarding your technical explanations.

Indeed. According to the RFC, your browser should have made it
evident to you, when you chose de-ch, that by doing so you were, in
principle, excluding all other kinds of de, and suggesting that you
maybe wanted to add the generic "de" at a somewhat lower preference.

.... Do we know *any* browsers which do that?

But since this situation is so common (especially with MSIE), it's
evident that information providers have to make arrangements to deal
with this situation anyway on the server side.

all the best