- From: Harald Alvestrand <harald@alvestrand.no>
- Date: Wed, 17 Jul 2013 13:15:49 +0200
- To: public-media-capture@w3.org
- Message-ID: <51E67CE5.9050004@alvestrand.no>
On 07/17/2013 11:35 AM, Robert O'Callahan wrote: > On Wed, Jul 17, 2013 at 8:55 PM, Harald Alvestrand > <harald@alvestrand.no <mailto:harald@alvestrand.no>> wrote: > > We may have a problem here if we expect a MIME type for each blob. > MIME types are usually defined in terms of the whole file (or the > whole stream for RTP MIME types), including any headers that occur > only at the beginning. > > My expectation is certainly that the MIME type of the file that > results from concatenating all the blobs needs to be clearly > defined and useful, but I don't see a reason to have a MIME type > for each blob. > > If Blobs have MIME types in general, that may be an argument > against using Blob. > > The MIME type of a Blob is optional. See for example > http://www.w3.org/TR/FileAPI/#constructorParams. See also the APIs for > concatenating and slicing Blobs --- the idea that a Blob can represent > part of a resource is not new. Thanks! I think it would make sense to me to return the MIME type chosen for the whole file as the MIME type of the first blob, and then return no MIME type at all (or, almost equivalently, the "fallback" application/octet-string) for subsequent blobs.
Received on Wednesday, 17 July 2013 11:16:19 UTC