commments on techniques doc.

Some comments on the latest techniques document dated 16 of March: 

2.7 content negotiation.

I think one should use both content negotiation and 
include links to alternative versions.

Content negotiation determines the default format served,
1. a user may prefer French to English, but may still prefer 
   to have access to both versions.
2. User agents may not support content negotiation well and the
    format served to the user may be not the optimal one.
2.8 Automatic page refresh.

First the sentence:
"Developers cannot predict how must time a user will 
 require to read a page;"

The word "must" should read "much".

Second, the term "server side redirect" is incorrect and misleading. Both
in "META refresh" and in HTTP status code 301 the server provides to the
user agent information about the new location and 
the user agent is making a new request.

The differences are:
1. Using HTTP headers vs. markup in the "entity body".
  Using HTTP headers is prefrreable because:
  it reduced internet traffic, and download time, it may by 
  applied to non-HTML documents, it may be used by agents who 
  requested only a HEAD request (e.g. link checkers) 

2. giving information on the nature of the redirect.
 Returning status codes of the 30x type provide 
 information such as moved permanently or temporarily etc. 
 that cannot be given with META refresh.

Thus, I propose to recommend to use the appropriate HTTP status code
rather than recommending the usage of "server side redirects".

3.7 Image maps.

The following text is confusing:
"Image maps are created with the MAP element (available in 
HTML 4.0 and 3.2). HTML allows two types of image maps: client-side 
(the user's browser processes a URI) and server-side 
(the server processes click coordinates)."

MAP is used only in client side image maps. 
Indeed later in the text the example which intends 
to be one of a server side image map uses a client 
side one, which ofcourse adds to the confusion.

3.8 applets and object.

I can't see why the recommendation using EMBED / NOEMBED for 
an applet. authors concerned with backward compatibility with browsers that
do not support OBJECT should use the deprecated 
APPLET element, which is a part of the specs for exactly this purpose.

One may consider using EMBED for some other object, say an MPEG video.
However in this case one should also supply a DTD with which people who
follow the recommended technique may use to validate their documents.

In addition, the EMBED example has an IMG element with no alt text.

That's all for now.



Nir Dagan

"There is nothing quite so practical as a good theory."
-- A. Einstein

Received on Wednesday, 17 March 1999 12:18:35 UTC