E-mail List Archives

Re: URI's in print media


From: Terence de Giere
Date: Jul 25, 2004 5:21PM

Some staff are wanting to standardize company policies for print
URI's, excluding the &quot;http://&;quot; string.


Recent technology has routines for auto URI completion. Screen readers
like JAWS will work if the browser correctly auto completes the URI.
Some older technology may not do this however, so it depends on the
user's knowledge to supply the missing information. I remember
incomplete URIs not working. Users of older technology might have a
problem. I was about to test an older assistive browser I had not used
in a while and discovered that it no longer runs under Windows after all
the recent patches, so I couldn't even test it. Old technology with a
novice user is the most likely scenario where not having a complete URI
will fail. Netscape 1.22, which I have on my machine, completes the URIs
just fine. So do the more recent Windows ports of the Lynx text browser.

The full form is http://www.domain.whatever/

If you truncate the full form of the address, then the browser has to
try things to fill in the missing pieces. Usually that is the &quot;http://&;quot;
or the &quot;www&quot; or the final slash &quot;/&quot; or up to all three. I think it is
probably safe to leave off the protocol, the www, and the final slash as
it is quite common to see this in print, and the browsers fill in the
blanks. If there are other protocols used, such as HTTPS or FTP, that
should be written out in full because browsers don't look for these and
they are less familiar to users. If you are referring to a specific page
other than the default home page, it might be better to have the full
address. I don't know of any statistics on this question.

If the printed material has a URI as a reference, for example, in a
footnote or endnote, you should use the full form always, as this is a
formal citation. If it's advertising, a brochure or something, from a
visual designer's point of view, a full URI looks cluttered, and most
users just type incomplete URIs except people like me. HTTP is the most
common protocol, so think of it as a de facto standard to leave it off,
like not dialing an area code for a local telephone call. Browsers do
handle other protocols, like GOPHER, which are hardly used anymore. As a
precaution, in Web pages themselves, the full form of the address should
always be used in the code. For email addresses,
<EMAIL REMOVED> er should be always used in the link text as the
email client may not be associated with the browser, and the user will
have to type it in because the client will not start the email
application under these conditions, and if it is not in the link text on
the page, the only recourse the user has is to look at the underlying
code. Email URIs in print are of course always complete, it is just
there is a tendency to use different text on a Web page like 'Bob
Smith', instead of ' <EMAIL REMOVED> er'.

Auto completion actually helps some users, such as those with motor
disabilities, because they have to type in fewer characters. Auto
completion is also good for small portable devices with tiny keypads or
keyboards. Unsighted users have to be smarter than the rest of us to
read Web pages anyway. I am for leaving it off for the default page in
the domain for public documents, but less sure about subsidiary pages.
If the documents are internal to the company, you could probably leave
them off even for subsidiary pages, as long as you know the browsers you
use function with incomplete URIs. Anyone else want to vote on this?
Does anyone know of a major piece of recent assistive technology that
bombs on incomplete HTTP URIs?

Terence de Giere