[whatwg] <INCLUDE> and links with @rel=embed

Tab Atkins Jr. jackalmage at gmail.com
Fri Aug 6 12:43:37 PDT 2010

On Thu, Aug 5, 2010 at 7:43 PM, Bjartur Thorlacius <svartman95 at gmail.com> wrote:
> On 8/5/10, Tab Atkins Jr. <jackalmage at gmail.com> wrote:
>> I'll note, though, that that isn't a very useful pattern for users in
>> the first place.  Most users won't have any idea what to do with a
>> video file, especially if it doesn't come with an identifying
>> extension.
> I don't understand the comment about identifying extension. What's wrong with
> IANA/MIME media types? Format metadata is often represented as a name/icon
> combo anyway.

Servers often don't serve their videos with the correct mimetype.

Do browsers supply a file extension for un-extensioned files based on
the mimetype?  I didn't think they did.  A file without an extension
confuses most people.

> And having the video prevents them from opening it in their media player how?
> Video playback consist mostly of a mouse click and some watching.
> Codecs are completely irrelevant.

Huh?  Codecs are certainly relevant *if you don't have the correct
codec installed to decode the video*.  For example, I don't think
Windows Media Player by default includes the necessary codec to decode
Ogg Theora, and I know it doesn't to decode WebM.

>> Doing a Flash or similar fallback that still plays the video is
>> virtually always the best choice for the user.
> "Virtually always" seems like an overstatement based on how often I
> see people extract sound tracks from Youtube videos (and how I often I
> see people don't in part due to them only having a dumb one-off
> Flash-based video player). Or try to sequence playing of arbitrary
> songs from Youtube with multiple browser windows + manual play/pausing
> hackery. Or other things that would be trivial with their usual media
> player.

You're running with a very non-representative crowd if those are the
sorts of things your friends do.

Ordinary people watch videos.  That's that.

> What's wrong with solving the problem once for all?

Every solution costs something.  If solving it "once and for all" is
more expensive than the benefit it'll bring, it's not worth doing.
That also presumes that you have a good, correct solution in the first

> And as SVG is not universally supported, what if I want to offer some
> logos/icons/images as both (gzipped) SVG and PNG?

SVG is in the process of being universally supported right now.  Once
IE9 is out, every modern browser will support it.

It's quite certain that you'll be able to use SVG before you'll be
able to use a generic media container holding both SVG and PNG for
fallback.  It's not usually worthwhile to try and solve things that
will solve themselves, given a little more time.


More information about the whatwg mailing list