Re: cache-busting document

W.Sylwestrzak@icm.edu.pl wrote:
> Drazen Kacar:
> 
> > >   Don't use content-negotiation until HTTP 1.1 is more widely
> > >     deployed, since in HTTP/1.0 it interacts badly with proxy caches.
> > 
> > What am I supposed to use until then?
> 
> Hm, what are you specifically asking about ?

PNG vs. GIF, for example. Or language negotiation. Or charset negotiation.

> > >   Don't use server modules or scripts to convert document's character
> > >     set on the server side.  Leave it to the client.
> > 
> > What if the client can't do it? The text will be unreadable. That's not
> > acceptable.
> 
> It's the responsibility of the client to parse and display the document
> to the end user. The server should not make any assumptions 
> about the client's capabilities. The same priciple as with frames,
> images etc. Especially that charsets are platform dependent.

I agree that it's the client's responsibility. But many clients can't do
it. Even worse, some clients have annoying bugs or almost-but-not-quite
implementations. Take NSN 4.0 betas, for example. The browser sends
Accept-Charset which lists `*' and `utf-8' without q factors. So, I'll
send ISO 8859-2 document and it will be rendered with adequate charset
if it's available on the client machine. If it isn't, unknown characters
will be rendered as `?'. Now, suppose I write some page in English
where my name is the only non-English text. Do you suppose I'd get
mail starting with "Dear Dra?en Ka?ar..."? From my point of view, that
would be amusing, but I don't think anyone will send something like that.
Certainly not in a business letter.

There are many clients which can't do charset stuff right. Am I supposed
to put "Usable only with the latest Lynx beta" on my pages? If not,
I have to make content readable in some way. Cache busting is a reasonable
price to pay, if you ask me.

BTW, charsets are not supposed to be platform dependant. That's what IANA
registry is all about, along with recommended charsets in HTTP RFCs.
There are brain damaged implementations, unfortunately. If I use
Content-Type: text/html; charset=iso-8859-2
and put <FONT FACE=Arial> in my document, certain clients on Windoze 3.1
will use ISO 8859-1 font, because Latin 2 glyphs are in "Arial CE" font.
But that is a *bug* in client (or OS, GUI or window protocol) and not the
proof that charsets are platform dependant.

-- 
 .-.   .-.    Life is a sexually transmitted disease.
(_  \ /  _)
     |        dave@srce.hr
     |        dave@fly.cc.fer.hr

Received on Sunday, 8 June 1997 16:15:59 UTC