W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > April to June 2005

Re: frames and no frames content

From: Matthew Smith <matt@kbc.net.au>
Date: Thu, 19 May 2005 12:03:42 +0930
Message-ID: <428BFB06.1010000@kbc.net.au>
To: Lynn Alford <lynn.alford@jcu.edu.au>
CC: w3c-wai-ig@w3.org

Antony Tennant wrote:
>> Also as a side query, does anyone have any stats on browsers used that 
>> do not understand frames.

Lynn Alford replied:
> I'm not sure which browsers don't understand frames, but search engines 
> frequently don't do frames.  So if your content doesn't appear in a 
> noframes area, people searching for information will not find your site.
> Search engines work better on sites with an accessible design.

Matthew comments:

It seems that most people are thinking of accessible content as that 
which can be rendered appropriately by one of a set of existing software 
agents; this is little more than extending the scope of the old "best 
viewed in XYZ user agent".

Maybe we should think beyond the "ordinary browser" that does little 
more than make a literal representation of content provided and consider 
the existence - present or future - of software agents that actually try 
to make sense of the content and perform further processing on it before 
passing it on to the user (we're getting into the realms of Semantic Web 
here).  Lynn has provided us with a perfect example: search engines.  If 
a page cannot be understood by a search engine and therefore be found, 
is it accessible?  I would say that hidden does not equal accessible.

My own experience of writing agents that try to retrieve certain data 
from pages (real-time search agents that attempt to summarise and 
aggregate interesting pages), is that pages that are reasonably 
accessible and have decent semantic markup (proper use of headings, 
etc), are much, much easier to handle.  Whilst I haven't, as yet, tried 
to handle frame-based pages, I think I would be inclined to programme my 
agent to take the term "frameset" to mean "look somewhere else".

To summarise, we cannot use a study of a few available user agents to 
test accessibility - we also have to consider agents unknowable to us 
(search engines) and agents as yet unwritten.



Matthew Smith
Kadina Business Consultancy
South Australia
Received on Thursday, 19 May 2005 02:33:57 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:36:25 UTC