- From: David Dailey <david.dailey@sru.edu>
- Date: Wed, 11 Apr 2007 10:13:48 -0400
- To: public-html@w3.org
At 02:38 AM 4/11/2007, Cameron McCormack wrote:
>David Dailey:
>
> > >execCommand('SaveAs',false, '.txt') that was (is?) available in
> > >IE. About seven years ago I put together a little web app using VML
> > >(IE only) http://srufaculty.sru.edu/david.dailey/grapher/
>
>[...] For this particular case you could construct a data: URI and
>have the browser navigate to it, so that the user could then save the
>file you want.
data: URI is not yet available in IE (according to Wikipedia). If I
follow you here, it means that if data: URI becomes a part of our
standard web arsenal, then we can use it to allow users to store
files client-side?
It is apparently covered by IETF RFC 2397. The only references in
http://www.whatwg.org/specs/web-apps/current-work/ to either RFC 2397
or data: URI seems to be in the <canvas> section of the document. If
this is how users are to save their work, then I think we oughta make
sure it's in the spec.
Ian Hixie wrote:
>Why wouldn't you just provide a link to the file and let the user use the
browser's Save Page As feature?
If this is a different question than Cameron's, then the answer would
be that it is because the content of the page is constructed by the
user/visitor -- it's a graph drawing program. Save As saves the
application, not the work done in the application.
I think the use case for allowing users to save content would pretty
much apply to all situations where the user uses a web application to
make stuff.
It would be fine with me to allow that to happen through data: URI --
that apparently circumvents some of the security risks, but the
ability to read and write files client side seems like a fundamental
requirement to me of "web application" development.
David
Received on Wednesday, 11 April 2007 14:16:41 UTC