Re: [dispatch] Stephen Farrell's Yes on charter-ietf-cellar-00-00: (with COMMENT)

[Michael Niedermayer <[e-mail]>]

On Mon, Oct 26, 2015 at 02:23:54PM -0500, Ben Campbell wrote:
> On 22 Oct 2015, at 16:19, Michael Niedermayer wrote:
> 
> 
> [...]
> 
> >>>>
> >>>>I wonder is there any interest in the security of the archived
> >>>>stuff here, and whether the WG would want to possibly take
> >>>>that on, most likely later. I'm not saying they should, as that
> >>>>could be a lot of possibly contentious work, but if they did
> >>>>want to do it, it'd be good to say in the charter.
> >>
> >>Does he mean security as in only allowing certain people to
> >>access the
> >>data ? IMO it's out of the scope and better handled out of the file
> >>format/streaming. It's more a transport/storage issue than format.
> >>
> >>There is encryption possible in WebM though that's working with a EME
> >>DRM module.
> >>http://www.webmproject.org/docs/webm-encryption/
> >>
> >
> >>If he means data safety, that's something we can work on. We already
> >>have the CRC for checking. In the past we also discussed redundancy
> >>but it was ruled out (again, better handled outside of the file
> >>format).
> >
> >forward error correction, that is to add redundancy to recover from
> >damaged or lost parts of an archived file or stream/track, is
> >something that sounds interresting. CRCs also exist at the video
> >stream
> >(FFv1) layer. Theres currently no means to restore lost frames, slices
> >or disk sectors except through concealing them based on previous
> >frames. The existing CRCs are sufficient though to correct rare bit
> >errors, if that occurs in any actual use case.
> >Adding some layer of protection to protect from damaged or unreadable
> >disk sectors or other larger damges losslessly is possible and
> >interresting. Iam not sure at which layer that would be best to do
> >though.
> 
> For the record, I think Stephen was asking more about
> confidentiality than about error correction.
> 

> Do you think anything relative to error correction needs to be added
> to the charter? It seem to me that is something that could be added
> in a future recharter if people think something becomes needed
> beyond the currently planned work.

the charter is fine, i just wanted to comment on "data safety", which
as it turns out the original question wasnt about ...

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Rewriting code that is poorly written but fully understood is good.
Rewriting code that one doesnt understand is a sign that one is less smart
then the original author, trying to rewrite it will not make it better.