[5960] in www-talk@info.cern.ch
Re: Forms support in clients
daemon@ATHENA.MIT.EDU (Mike Meyer)
Fri Sep 30 22:29:15 1994
Date: Sat, 1 Oct 1994 03:26:50 +0100
Errors-To: listmaster@www0.cern.ch
Errors-To: listmaster@www0.cern.ch
Reply-To: mwm@contessa.phone.net
From: mwm@contessa.phone.net (Mike Meyer)
To: Multiple recipients of list <www-talk@www0.cern.ch>
> What we should be spending air time on is the abstract API between
> the browser and the script interpreter. Trusted scripts can be recognised
> as such as they will carry a digital signature plus certificates in the
> HTTP header. I will therefore concentrate on untrusted scripts:
>
> o scripts can [...]
> I don't have time to refine this, and would love to see other people
> work together on refining this into a detailed proposal, rather than
> arguing on which language to choose.
Fair enough. You missed a few I think are important. However, I think
the script language should double as a macro facility for the users,
which adds a number of things:
o scripts can manipulate the GUI directly (open/close/resize/move
windows & dialog boxes)
o scripts are associated can be associated with a specific
window in a multiwindowed browser.
o scripts can feed HTML to the Browser for display.
o scripts can change the browsers preferences settings.
And you've got one odd one here:
> o scripts can't send messages
Why not? That's the most useful thing about the scripting extensions
I use in Mosaic. Turning this off for "unsafe" scripts is probably a
good idea, but that's no reason to leave it off for everyone.
<mike
