[whatwg] <video> application/octet-stream

[Henri Sivonen]

Chris Double wrote:
> As I mentioned in a previous email, the sniffing could result in a
> reasonable amount of data being consumed. I'm sure people who run
> sites that share HTML 5 video would appreciate browsers not consuming
> data bandwidth to sniff files that they've already specified as being
> something the browser doesn't support.

I think the solution to this concern is to allow authors of bandwidth-sensitive to specify the type attribute on <source> or the Content-Type header on the HTTP response to say something other than application/octet-stream or text/plain. For best performance, authors should use the type attribute in multi-<source> cases anyway.

I don't think it follows that it's necessary to fail hard in the application/octet-stream or text/plain cases for all sites--even ones whose authors value ease over bandwidth conservation or ones whose authors have carefully specified the type attribute on source but are using a hard-to-configure CDN to serve the files.

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/