Hi Bjoern, Many thanks for the quick review. See inline for comments and answers. I will work on an update and hopefully be able to send it out before we put it into the next version of the TS 26.346. Bjoern Hoehrmann wrote:
* Magnus Westerlund wrote:
Please review the proposal in the attached word document in the previous mail as it intended to fix the total lack of MIME type specifications in the referenced 26.346.
Well, I did, but it seems that this information was added later to the document; not all viewers support such editing information, using more portable formats might be more appropriate here.
Sorry about that, word format with change marks are the by 3GPP required format for contributions. Converting into text would have required me to reformat the sections which wasn't time I like to spend on it.
The +xml types all lack a charset parameter. This is needed to allow general purpose XML applications like Validators to decode the format without hard coded knowledge of the media type.
Okay, adding an optional "charset" parameter is not a problem. If I understand this correctly this equals the encoding="UTF-8" attribute in a <?xml version="1"?> line?
Some of the XML types have encoding considerations like "The content is well formed XML document without any binary parts" which sort of misses the point of the field. These should simply refer to the considerations in RFC 3023.
Okay, I will use the "Same as [charset parameter / encoding considerations] of text/xml as specified in RFC 3023." reference sentence. Although I don't think the RFC is easy to use when it comes to referencing it and separating what is for the types registered and what is the general parts.
For application/mbms envelope+xml it is unclear what the syntax of the optional parameters is, from the brief description it seems these are meant to be boolean parameters but it's not clear how to specify true or false values.
They are intended to be boolean thou their presence or absence.
This media format contains XML and may contain binary embedded objects using CDATA sections within the XML. Thus for transports not supporting binary content BASE64 [82] encoding is suitable.
This is incorrect, XML documents cannot include binary data, only text. CDATA sections are not exception, they are just concerned with characters that introduce markup like & and <.
Okay, this is something I will have to discuss with the group. I think there is some general lack in the scheme on how to embed files within the envelope.
application/mbms-user-service-description+xml is also a bit unclear about the serviceID parameter, in particular, it is not clear whether a single service identifier must stille be quoted, or if there are constraints about what is a legal identifier for the purposes of the paramter; it should be possible to derive a clear grammar for legal parameter values from the registration.
The identifier is defined as "xs:anyURI" which I think is most suitable to always quote even single instances. I will clarify this.
The security considerations are
This media format is used to configure the receiver on how to participate in a service. This format is highly suspectible to manipulation or spoofing for attacks desring to misslead a receiver about a session. Both integrity protection and source authentication is recommend to prevent missleading of the receiver.
It is unclear how to protect the integrity of such data objects, a brief look at the corresponding schema suggest that e.g. use of XML digital signatures with such content is prohibited.
This protection is intended to be done through the security mechanisms in MBMS. I don't think using the XML digital signatures has been considered at all. Thanks Magnus Westerlund Multimedia Technologies, Ericsson Research EAB/TVA/A ---------------------------------------------------------------------- Ericsson AB | Phone +46 8 4048287 Torshamsgatan 23 | Fax +46 8 7575550 S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@ericsson.com