Re: Revisiting Authoritative Metadata

Karl Dubost wrote:
> The issue is then for the thousands, sometimes millions of users who
> land on a Web site which does the wrong thing. The tool is not the
> server, the tool is the browser and so the tool is wrong.

Sorry, but I've never understood browser vendors blaming everyone under
the sun for their own astonishingly bad UI choices which I've never
read a spec which called for; deluding themselves so far as to actually
assign an employee the Sisyphean task of getting the likes of Microsoft
to change terminally-borked code.

> In the mean time **for the users**, Opera had/has to do horrible
> hacks for coping with wrong server configurations and abusive user
> agent sniffing.

Opera's my preferred UA, but honestly all they ever had to do for me
was gracefully degrade to HTML parsing and note the _website_ error
unobtrusively in the status bar.  Silent error recovery is just as bad,
name me one Web developer who hasn't learned from making mistakes and
tell me how anyone will learn when no mechanism remains to even make
them *aware* they've made a mistake.

Subjecting users to the worst usability decision I've encountered since
my CP/M days was never strictly required by anyone, even XML developers,
and with that sort of error presentation I have no sympathy for the
user wrath it engendered, attempts to lay off that blame or call it
misguided, or the horrible hacks such misbegotten decisions led to.


Received on Wednesday, 27 February 2013 23:17:55 UTC