- From: Jean-Claude Dufourd <jean-claude.dufourd@telecom-paristech.fr>
- Date: Tue, 21 Jun 2011 16:01:16 +0200
- To: "public-web-and-tv@w3.org" <public-web-and-tv@w3.org>
Dear all, Thinking about a discussion on the level of APIs, which is full of various misunderstandings, I am summarizing here my take on this issue. Option 1: lowest possible API = something like WebSockets at the UDP level (WebSockets being TCP) The minimum to be able to implement UPnP is to have a UDP socket API, with broadcast and multicast, in the browser. I believe the same would work for other protocols, but that needs to be confirmed. WebSockets seems frozen because of security concerns, so obviously such an lower-level interface would be even more dangerous. On top of this, we would need code to implement UPnP and equivalent (in whatever language, C or JS). On top of UPnP, we would need code to implement generic discovery and messaging (in whatever language, C or JS). If the UPnP layer is to be loaded per page, there are issues with multiple instances, resets, etc... Option 2: specific discovery and service protocol interface, e.g. UPnP-specific interface An example of that would be adding platinum, an open source UPnP package, to a browser and exposing a UPnP-specific API to JS. On top of that, we would need code to implement generic discovery and messaging (independent of UPnP). Option 3: generic discovery and service protocol interface An example would be to add platinum to the browser, then implement (in whatever language, C or JS) a generic API on top of the option 2 UPnP-specific API (which might not be exposed) Security considerations: Option 1 seems simply too dangerous, unless the API is accessible only with severe restrictions. That and the other issues disqualify this option for me. Option 2 has the security issues of UPnP. Option 3 probably has the same security issues as option 2 (at least, I have no proof that it does not). Preference: My stated preference is for option 3, including a messaging interface. Other stated opinions cover the range of all above options. There is actually many possibilities in between the above options, but I believe this provides useful references. Best regards JC -- JC Dufourd Directeur d'Etudes/Professor Groupe Multimedia/Multimedia Group Traitement du Signal et Images/Signal and Image Processing Telecom ParisTech, 37-39 rue Dareau, 75014 Paris, France Tel: +33145817733 - Mob: +33677843843 - Fax: +33145817144
Received on Tuesday, 21 June 2011 14:01:51 UTC