[WEB4LIB] Re: Is Netscape dead?

[Thomas Dowling]

Josh Kuperman writes:

> ...The test of a browser
> should be foremost in its ability to correctly display HTML. The
> authorship is irrelevant. The web site,
> http://tycho.usno.navy.mil/howclock.html, which shows how to get an
> accurate clock on a web page is one of many, that points out something
> that NetScape got right that MSIE didn't. (I haven't checked this
against
> the latest version of MSIE.)

The page in question has nothing to do with displaying HTML correctly, but
rather with inline support for a particular MIME type.

Do you have a citation where the MIME type in question
(multipart/x-mixed-replace) is defined as a standard?  The "x-" indicates
that it's a privately defined type, and various search engine hits suggest
that it was a Netscape invention.  If so, not supporting it is no more of
a shortcoming than not supporting BLINK or LAYER (or MARQUEE).


Valerie Lawrence writes:
> One reason I'm not ditching NS is that IE5 doesn't behave properly when
> I'm trying to download a search in Ovid.  Instead of a popup window
asking
> where I want to save the file, when I try to select a search to
download,
> I get a screen with all the text running together.  My option is to copy
> and paste from this??? No thank you.  Ovid says Micro$oft "knows it's a
> problem" but the Ovid rep said they didn't sound like they planned to
fix
> it.

Agreed, this is one of the most serious failures of MSIE. MS does know
about it, and they don't plan to fix it to my knowledge, because they
don't really consider it a bug.

You could make a case that Ovid is somewhat abusing MIME types to coerce a
browser into downloading instead of displaying text.  As I recall, Ovid
uses a common hack to do this; they send out text with a MIME content type
of "application/octet-stream", which is a server's way of saying, "this is
some sort of binary data, I have no idea what you're supposed to do with
it, good luck."  Other browsers will automatically open a Save As dialog
on the assumption that they won't be able to display it.  IE actually
examines the incoming data stream, and if it decides it's text/plain or
text/html, it goes ahead and displays it.  While this is arguably not
wrong, IE goes further, and A) uses file extensions in preference to
content type and B) uses its own examination of the data stream in
preference to content type.  This is clearly wrong, and prevents you, for
example, from creating an HTML tutorial and displaying the HTML markup by
copying foo.html to foo.txt.  IE will display it as rendered HTML either
way.  This is a We Know Best attitude that is pretty galling.

[Ovid isn't without recourse here, by the way.  If you send a
content-disposition header, IE will not only open a Save As dialog like a
good little browser, it will also default to the file name you send along
with it, a standards compliance checkpoint that other browsers haven't
worked out yet.  I mentioned this to our Ovid contact once or twice, and
didn't get a response.  But they're apparently still in "Hey, it works in
Netscape..." mode, as are way too many library vendors.]


I should point out that, while I do believe the Netscape browser is
moribund, I'm also well aware that the great majority of our member
libraries prefer it for public workstations, and that our percentage of
Netscape hits is seldom below 75%.  The installed base is not going away
any time soon.


Thomas ("Beautiful Plumage") Dowling
OhioLINK - Ohio Library and Information Network
tdowling at ohiolink.edu