Re: [Cellar] Adding private elements to Matroska

[Steve Lhomme <[e-mail]>]

But would we define an element for each level where one can appear
(that would be a lot) ? Or do we define a global element similar to
Void/CRC32 that is open bar for whatever you put it in, but a bit too
loose for validation.

The latter may be a simpler possibility and we use an element ID with
2 or 3 bytes to avoid false positives. Like Michael suggested an ID +
value may be useful too.

The issue would be for readers, they will have to map and signal this
element every time it's found.

At least this approach is a lot simpler that defining a whole system
to put in the EBML header.

2018-04-15 15:49 GMT+02:00 Moritz Bunkus <[e-mail]>:
> Hey,
>
> I'm not particularly against private _data_ inside Matroska, but more
> against private _elements_ as we already have ways of attaching arbitrary
> data (tags & attachments). We could simply define a schema for private tag
> names and be done with it.
>
> Kind regards,
> mosu
>
> _______________________________________________
> Cellar mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/cellar



-- 
Steve Lhomme
Matroska association Chairman