begin  quoting Gregory K. Ruiz-Ade as of Thu, Apr 12, 2007 at 10:01:42PM -0700:
> On Apr 12, 2007, at 5:37 PM, Christian Seberino wrote:
> 
> >The only thing I can see stopping Javascript from taking over the  
> >world is M$'s slowness to fix compatibility problems in IE.  That may
> >be  enough to sink your favorite technology I'm afraid.  Perhaps AJAX
> >will work  around the problems until IE improves?  One can hope.
> 
> AJAX is Asynchronous Javascript And XML.  Which makes the above  
> statement amusing to me, because it's already happening.  Most of the  
> "AJAX libraries" that are available do extensive testing to see what  
> browser they're running in to select the best (i.e., non-broken)  
> method for getting something done.  So, to an extent, it's already  
> happening.
 
So, it's a sort of built-in autoconf, eh?

> The problem with Javascript, as Andrew and Stewart will happily point  
> out, is that there is absolutely _zero_ security model with  
> Javascript.  I'm pretty sure you could have an AJAX app start  
> uploading random files from your hard drive to a server of your  
> choosing and the user would be none the wiser unless you completely  
> tanked his/her bandwidth.  Even then, they'd be puzzling over why  
> everything is so slow.

I believe that security in Javascript is pretty ad-hoc -- I do not
believe that you can (anymore, that is) have a Javascript program
upload random files from your hard drive, at least not trivially.

Security is getting better -- shoot, even my evil javascript page
doesn't cause heartburn anymore, and it used to be capable of crashing
a Linux box -- but the assurances haven't changed: "It works fine for
me, what's your problem? Are you a paranoid luddite or something?"

> I need to break out my Javascript book and see just what you can do  
> without needing to ask permission via the browser...

I should probably pick up a Javascript book and set about devising
a new and improved evil javascript page.

-- 
I should probably start with an enumeration of all objects in the system.
Stewart Stremler


-- 
[EMAIL PROTECTED]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to