Re: default return is "text/html"?

Paul (
Thu, 14 Jul 1994 16:27:18 +0100 (BST)


@ > I mean, come on. We are connecting to HTTP, "HyperText" Transfer
@ > Protocol, shouldn't the default be text/html?
@ Good point.

Very, indeed.

@ What this suggests, then, is that HyperTextTPDs should
@ *always* include a content-type header line. In other words,
@ the server should pin a default type (text/html) on all questionable objects.

Or the default Content-type should be defined as text/html as mentioned

@ I mean, come on. What's so hard about tacking on ".html"
@ to any file it serves-out? That's just one way to eliminate the
@ ambiguity. There are others.

This would be bad. Then a file such as say "my_text" would be served out
as "my_text.html" when it isnt, if Im reading the above correct.

@ Or (again) the clients could be "smarter", and upon seeing
@ "<html>" or even just "<head>" or any other SGML-ish strings,
@ decide, oh yeah, this *is* HTML, otherwise make it plain text.

Isnt this where the <!DOCTYPE...> comes in? But the problem is what
level of tag do you look for? Its the old story of what can we do
currently without breaking things. But we do need to go ahead and do
them before any more "damage" is done.

I would think that key things to look for would be:

<!....> Any of <!--, <!DOCTYPE etc...
<HTML> obvious
<HEAD> Some people miss the <HTML>
<BODY> ditto
<H1> A lot of people here seem to start docs this way

Hrm, but that is cumbersome, and you cant rely on "<" as the 1st meaning
part of the message being right.

I cant offer any reasonable suggestions that would break things one way
or the other.


.--------Paul Wain ( X.500 Project Engineer and WWW Person at Brunel)---------.
| Brunel WWW Support: MPhil Email: |
| Work Email (default): (Brunel internal extn: 2391) |
| or |
`-------------------So much to fit in, and so little space!-------------------'