Re: Netscape 2.0 beta MS windows 95 virus

Mike Meyer (
Tue, 7 Nov 95 08:55:05 PST

> >Of course, until people either 1) stop doing content negotiation based
> >on user-agent, or 2) start paying attention to browsers other than
> >Moscape, this is immaterial - authors (and users) of browsers that are
> >going to lie about the user-agent in order to get the best document
> >they can.
> We have dozens of browsers in our "browser-negotiation" database.

Having dozens puts you well ahead of most people who play this game.
Of course, since there are hundreds of user agents, you're probably
feeding a lot of user-agents substandard documents (exact numbers
depend on how you map from the user-agent value to a browser).

> Including items listing the features that we believe Microsoft's browser
> supports. However until someone proposes a way to tell me via the headers
> that this browser supports '<p align=3Dcenter>' and that this browser
> supports tables but not percentage widths, and this one supports tables
> within tables but not with forms in them; I'm going to have to keep doing
> browser-based presentation.

You want a way to distinguish between experimental implementations of
features proposed in various drafts? Not likely - having to live with
broken and partial implementations is the price you pay for working
outside the area for which there are accepted standards. That the
major players profit from such FUD probably has something to do with
why content-negotiation has had so little attention.

> What really gets me
> though is trying to figure out whether I need to send a RealAudio file, a
> WAV file, an AU file or an AIFF file. Why the !@#$% aren't the major
> browser manufacturors sending that helper-application information?

I believe they are. They just aren't sending any quality information
so you can determine that they can play something, as opposed to being
able to save it to disk. At least, it looks that way for the image/*
types. I haven't looked closely at the audio/* types.