> Let[']s not totally blame IE.
I wasn't thinking of blaming IE. It's just a product; a series of
frozen snapshots at moments in time of an evolving codebase that
represents a state of mind.
The state of mind of the authors of IE.
I was thinking of blaming the authors of IE, who, incidentally, munge
other agreed standards to their own ends. For example, to generate
those non-standard non-ISO apostrophes you keep using to encourage us
all to use Microsoft products with Microsoft fonts that will render
those 'apostrophes' correctly, because if we're not using your
products we just don't matter in the scheme of things.
> The fault here is that there is no mime
> types file to map htm to text/html on the server. (as it was said
> earlier). All IE does is notice the htm and cope with the fact that
> many unix server administrators dont setup mime-types to reflect the
> types of content that their content authors use.
...and this coping effectively bypasses mimetyping, making use of
mimetypes completely redundant in the process, of course, and screws
up the results for more standards-compliant browsers - like Netscape,
[IMO, Open File... should be entirely separate.]
> IMHO what should happen here is that the server admin should either
> communicate to the content authors what the right extension is (
> .html or .htm) OR he should update the mime-types file to reflect
> what they are using (htm).
That's a localised fix for a global problem. By introducing the
original workaround for your own legacy problems, you then force the
planet to work around the legacy problems that result. Do this enough
times, and a standard is meaningless; everyone is too busy working
with cascaded legacy problems to move forwards.
[We can see this happening to http. How much 1.1 traffic is _your
> Would it be smart to change IE so that it doesnt do this? It might
> cause a few authors to notice that their server's mime types arent
> set right, but mainly it will just break tons of files that are just
...and encourage administrators to upgrade their webservers.
Prevention is always better than cure.
If IE had followed the standards originally, this would have been a
non-problem, and you'd even have sold more copies of your products
with semi-decent filing systems on the basis that the upgrade was
essential for webserving.
Arbitrarily tweaking implementations of standards to suit your own
legacy problems will always result in cascading legacy problems. This
a good, simple, demonstration of that. Unlike, say, http. or tcp.
would someone please remind me what the point of having standards is again?