lynx-dev
[Top][All Lists]
Advanced
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: LYNX-DEV lynx on www.microsoft.com

From: Laura Eaves
Subject: Re: LYNX-DEV lynx on www.microsoft.com
Date: Sat, 3 May 1997 01:19:17 -0400 (EDT) 
> Date: Fri, 02 May 1997 22:04:46 -0500 (EST)
> From: Foteos Macrides <address@hidden>
>...
>       If an Anchor with an HREF is opened, it can only have empty
> content if the Anchor was closed, and there are two situations when
> that happens.
>
>       One is that the content has an IMG with ALT="", because the author
> does not consider that link suitable for browsers with image loading turned
> off.  If the IMG has a client-side image map indicated, Lynx will force a
> pseudo-ALT so you can use it.  Otherwise, if you toggle on image links (*),
> Lynx will create *two* links, e.g.:  [LINK]-[IMAGE], so that an intelligent
> user can realize that the link for the image goes with an anchor, which now
> is most likely appropriate to follow as well.  That's the behavior which
> presumeably was behind Hiram's suggestion about the -ismap fote mods, though
> inappropriate in the latter situation, as I explained.
>
>       The other case is when a site is creating links for mindless
> robots to follow, and they should not be displayed for human begans
> who use neurons instead of just transistors.  That's what's being done
> at the bottom of the www.microsoft.com page:
>...
> <A HREF="/products/default.htm" TARGET="_top"></A>
>...

First case: If the content of an anchor is an img with alt="", why not
treat that as no alt at all?  (Print the [LINK] string.)  Since the img
is there, the link is obviously intended to be visible (even if the author
doesn't care about text-only browsers).

Second case: Apparently these links are intended to be invisible.  (But it's
nice to see them anyway...:)  Is there a reason to leave them invisible?
How about a default string of [xxx] where xxx is the href string?

Before the mods to parsing error recovery, bad html would also create invisible
links -- but I can'at seem to reproduce that now.  Here's a test that runs
differently in 2.7.1 and ac-0.19:
(Note; keypad_mode=LINKS_ARE_NUMBERED)

With the devel code (new parsing), </b> is never processed, so the rest of the
page is underlined.  When I hit ^V to use fote's parsing, only the first is
highlighted.  In either case, all 4 links are there, but only 2 are "visible".
In 2.7.1, the first link disappears, but underlining is also terminatead.

<html>
<head><title>testing null links</title>
</head>
<body>
<pre>
<b><a href="file1.html"></b>a link</a>
<a href="file2.html"><img src="file.gif" alt=""></a>
<a href="file3.html"></a>
<a href="file4.html">another link</a>
</pre>
</body>
</html>

--le

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;
reply via email to
[Prev in Thread] Current Thread [Next in Thread]