[whatwg] "Content-Disposition" property for <a> tags

Markus Ernst derernst at gmx.ch
Mon Aug 2 10:07:01 PDT 2010

Am 02.08.2010 18:21 schrieb Michael Kozakewich:
>>> Dennis wrote:
>> Yes, but that wouldn't help since I want to force downloads regardless
>> of the browser settings. Maybe it would do if the type was set to
>> application/octet-stream, since those, by default, always get downloaded.
> People don't often like it when they're forced to do something. If they 
> want
> to download it, they can select "Save Link As..." from their browser. If
> they actually want to view it full-screen, they don't want to be forced to
> download it.
> I see where you're coming from, but we try not to force users to do things.
> For the same reason, @target was removed from links.

But it has made it back into HTML5:

I assume the reason is that some de-facto standards have established 
that users have got used to.

One of them seems to be that clicking on an external link in a corporate 
web site usually opens a new window or tab, whatever they set in their 
browser prefs. Some users (like me) might forget to hold the ctrl key 
because most corporate websites use target="_blank" anyway on all 
external links, and get annoyed if the original page disappears and they 
have to go back their session history.

Another de-facto expectation could be, that on a "download" page with 
several links, they behave consistently, starting a download:

<h1>Press downloads</h1>
   <li><a href="portrait.jpg">Portrait image (JPG)</a></li>
   <li><a href="band-photo.tif">My band (TIFF)</a></li>
   <li><a href="cv.pdf">Curriculum vitae (PDF)</a></li>
   <li><a href="discography.xls">Discography (Excel File)</a></li>

Every link is likely to invoke an other default action, which does not 
seem consistent to me in this context - even if the downloading 
journalist as an experienced user might be able to anticipate the action 
based on the file type information in the brackets. Dennis' and Roger's 
proposal would definitely be helpful here.

More information about the whatwg mailing list