Form spam? What to do?

[Toby Inkster]

Your script will serve application/xhtml+xml to a browser with an Accept
header of "text/html,application/xhtml+xml;q=0.1', which is completely
broken behavior.

This function implements somewhat more thorough HTTP Accept checking:
http://message-id.net/[email protected]

 

[cwdjrxyz]

This function implements somewhat more thorough HTTP Accept checking:
http://message-id.net/[email protected]

You will find dozens of discussions of this subject on the web. Some
are very elaborate. Some are simple. Some work only for xhtml 1.0. I
tried many approaches. If you get elaborate with using q, many of these
approaches will connect you to serve html rather than the desired true
xhtml 1.1. My desire was to force using true xhtml with mime type
application/xhtml+xml for any browser that reports it can use it at
all. In such a case you have to be careful to check major browsers that
can use true xhtml to see if they have any bugs when using it. I found
only a few minor bugs that are easily corrected for pages of interest
to me. This I decided the most simple method met my needs. I started
out by writing two pages of the type something.xhtml and
something.html. My server is set to deliver .xhtml with the mime type
application/xhtml+xml. It is set in the usual way for the .html
extension to deliver text/html. You can connect to a short redirect
page that will allow you to select either the something.xhtml or
something.html page. Most of the modern browsers can use either, but
IE6 is the exception that can not use the something.xhtml page.
Although I have plenty of storage on the server for near-duplicate
pages, I find that I do not have to use this method for any browser of
interest to me. However, if I ever have a problem, there are several
things you can fall back on such as duplicate pages, selection of a
range of q allowed, etc.

The IE7 browser is in early beta test by some selected software people.
Of course there is plenty of discussion about this on the web. You
never know what to believe. However one report I read suggested that
IE7 still will not support application/xhtml+xml. If so, this is very
disturbing. However someone could just be attempting to stir up
trouble. However there are reports that several of the IE6 CSS bugs
will be fixed in the IE7. Time will tell.

 

[Jukka K. Korpela]

CSS is not "part" of HTML 4. One can happily use style sheets in
validating HTML 3.2 documents. (Just not the style attribute or <style>
element.)

HTML 3.2 even had <style> elements (though as incompatible with HTML 4:
no attributes were allowed, whereas HTML 4 requires the type attribute,
rather pointlessly, since text/css could have been defaulted even
formally).

HTML 2, on the other hand, had no <style> element - but you could still
have used style sheets with HTML 2, if browsers had implemented style
sheets in a manner that does not require their binding to HTML in HTML
source. In fact, that's the path things _should_ have taken.

 

[Leif K-Brooks]

I just like to keep up with the latest versions of code, when possible.

So you want to comply to an experimental and pointless standard like
XHTML just for the sake of compliance, but you don't want to comply to
the HTTP standard even when there is a good reason to do so?

 

[cwdjrxyz]

I started
out by writing two pages of the type something.xhtml and
something.html. My server is set to deliver .xhtml with the mime type
application/xhtml+xml. It is set in the usual way for the .html
extension to deliver text/html. You can connect to a short redirect
page that will allow you to select either the something.xhtml or
something.html page. Most of the modern browsers can use either, but
IE6 is the exception that can not use the something.xhtml page.
Although I have plenty of storage on the server for near-duplicate
pages, I find that I do not have to use this method for any browser of
interest to me. However, if I ever have a problem, there are several
things you can fall back on such as duplicate pages, selection of a
range of q allowed, etc.

Another page I have with the php include is
http://www.cwdjr.info/media/playersRoot.php . I still have an earlier
version with a short redirect page that I mentioned at
http://www.cwdjr.info/media/playersConnect.html . This page uses the
Windows conditional comment that only IE and close relatives see to
automatically direct IE browsers to a .html page served as text/html.
If you do not have a IE browser then you can select the link for newer
browsers that is a .xhtml page served as application/xhtml+xml . Or you
can select the link for older browsers that might not support true
xhtml, which takes you to the same .html page as does the automatic
redirect for IE browsers. Of course if you have problems with the
..xhtml page, you can also go back and select the .html page route. So
you see, even if I found a situation where I could not or did not want
to use the php include, I always could use this redirect page method
for a special case.

 

[Toby Inkster]

HTML 2, on the other hand, had no <style> element - but you could still
have used style sheets with HTML 2, if browsers had implemented style
sheets in a manner that does not require their binding to HTML in HTML
source. In fact, that's the path things _should_ have taken.

In Gecko based browsers you *can* use style sheets with valid HTML 2.0
documents! Gecko supports the "Link" HTTP header, so style sheets can be
bound to an HTML document through HTTP headers, or even through <meta
http-equiv="Link">.

 

[Alan J. Flavell]

In Gecko based browsers you *can* use style sheets with valid HTML 2.0
documents!

This is nothing new. I was playing with CSS applied to HTML2.0 (valid, of
course) quite some years ago, and I didn't find any CSS-aware browsers
back then which were upset by the lack of the type="text/css" on the
<link...>

It was a nice way to remedy the horrible default presentation which
browsers seemed to have copied from the original NCSA Mosaic, down the
years.

To do anything impressive, though, one has to be fairly creative to get
around the lack of a "class" attribute in HTML2.0. Ho hum. Pity that I
lost the test documents that I produced back then.