Warp <war### [at] tagpovrayorg> wrote:
> It's actually not by accident. Internet Explorer deliberately ignores
> the HTTP header protocol and tries instead to guess the right MIME type
> by examining the received data itself. For this reason it will display
> eg. HTML pages as HTML, image files as images and so on, regardless of
> what the HTTP headers claim the MIME type is.
>
> I'm guessing they made it like that because back in the day a good
> majority of web servers were misconfigured and were sending files with
> the wrong MIME type in the HTTP headers. (For example, while gifs and
> jpegs might have been configured on the server side to be sent with the
> proper HTTP headers, newer image formats like PNG were typically not,
> so the server would by default claim that they were plain text, thus
> looking like text garbage when looked with a compliant browser.)
>
> The problem with this is that it breaks the HTTP standard and by being
> back then by far the most popular web browser, the majority of web servers
> never got configured properly. In fact, many people thought that it was
> the other browsers that were broken, and IE the only one that worked
> properly. (When in fact the problem was a misconfigured HTTP server, and
> IE deliberately hiding the fact.)

Similarly (though in the opposite direction), Netscape 4, one of the two
buggiest Web browsers ever written (the other was IE3), lasted years longer than
it ever deserved to, because Web authors kept programming around its bugs.  It's
hard to blame the Web authors, because no one wants their Web pages to be
screwed up by a popular user agent.  But because they inadvertently covered for
Netscape, the end users never learned what a piece of crap they had on their
desktops.

I have to side with the standards.  If I'm doing it wrong, I want to know.
Micro$haft does me no favors by giving me a false sense of security.

Perhaps it's because of my programming background, where I'm used to the
computer doing exactly as I tell it, not necessarily what I intend.  More
likely, it's because of my experiences with Micro$haft since 1997, and more
recently with The Google and with KDE's KWrite: when computer programs
second-guess me, they USUALLY GUESS WRONG.

I appreciate a /warning/ if the program senses that I'm about to do something
regrettable or irrational.  The most obvious example of this is "Document is not
saved. [Save] [Discard] [Cancel]"  Similarly, I appreciate a search engine's
/suggestion/ if I misspell a word.  But don't /guess/ what I want and then
follow through summarily.

Post a reply to this message