- From: Hamish Cunningham <hamish@gate.ac.uk>
- Date: Tue, 23 Jun 2015 09:05:35 +0100
- To: Vlad Trifa <vladounet@gmail.com>
- Cc: public-web-of-things@w3.org
- Message-ID: <CAAcCExOxFj3m_oNTSoXQmV4NokBFCqzEjDWnM+i=VY_==KGOSg@mail.gmail.com>
hi Vlad Looks good! A few random questions: I've been playing with the ESP8266 developing wifi-connected sensor/actuator boards. The MCU on these chips is pretty powerful, and we plan to host several sensors on each board. Would the WoT server running on the MCU expose these as multiple things, or expose the board itself as a single thing? Is there anything in the model about joining? (E.g. the ESP can act as a wifi access point, so we can have quite a convenient "configure me" routine that offers an access point and serves a config UI, e.g. to enter the credentials of a more permanent local access point to join. I guess this is outside of the WoT space, or...?) I can see the advantage of a WT and a WoT server sharing a core API, but presumably the latter will expose additional functionality, e.g. authentication? I guess that WTs will often be too constrained to do anything clever with authentication and will rely on local firewalls and the like, whereas part of the point of a WoT server might be fine-grained access control. (Perhaps this is the type of thing that an "Interoperable Web Thing" does?) Would a WoT server host a Linked Data repository by default? And therefore a SPARQL endpoint? (I'm not familiar with JSON-LD; perhaps there's some equivalent query API there?) I wonder what the "semantic extensions" might look like? If we're already in Linked Data then we can presumably apply additional logic up to and including OWL over the top of the data exposed by WoT servers? Perhaps this logic will often be application specific (and/or non-existent), which might suggest they may be an expensive thing to include in the basic model? (I like this diagram -- https://camo.githubusercontent.com/8dfbcb18d4bd494f4247b80ce573072ad1d23615/687474703a2f2f7765626f667468696e67732e6f72672f77702d636f6e74656e742f75706c6f6164732f323031352f30342f776f742d6c6576656c732e706e67 -- but also think that is potentially quite a lot of layers to understand.) In https://github.com/w3c/wot/tree/master/TF-AP/wot-compose-recommendation/web-things-model the table entry for "name" is wrongly copied from another entry I think? Re. the reqs doc -- https://github.com/w3c/wot/tree/master/TF-AP/wot-compose-recommendation/web-things-requirements -- I guess that some things will be too resource-constrained to fulfil all the desiderata? In which case they have to operate via a proxy, or? (Perhaps that is what a "gateway" is?) I'll try and make the interest group telecon tomorrow morning; see you there? Have a good one, Hamish Cunningham Professor of Computer Science, University of Sheffield, UK 07920 765 455 http://twitter.com/@HCunningham hamish@gate.ac.uk http://pi.gate.ac.uk http://hamish.gate.ac.uk http://gate.ac.uk On 22 June 2015 at 10:10, Vlad Trifa <vladounet@gmail.com> wrote: > Hi Hamish, > > we have just published last week a first iteration of this definition, > which you can find here: > > https://github.com/w3c/wot/tree/master/TF-AP/wot-compose-recommendation > > This will show you the different capabilities (API/data models) of a Web > Thing (an entity of the Web of Thing), what is the data & services it > offers, etc. > > > https://github.com/w3c/wot/tree/master/TF-AP/wot-compose-recommendation/web-things-integration-patterns > > It's important that the interface of a Web Thing and a WoT server are > common. Obviously, a WoT server has different and additional capabilities > (can expose many Web Things, much more storage and processing power, user > interface to manage & create Web Things, support for many protocols, etc.). > > Let me know what you think! > > Vlad > > > > On 19 Jun 2015, at 21:37, Hamish Cunningham <hamish@gate.ac.uk> wrote: > > hi > > Is there a definition somewhere of what a WoT server does? > > (I've had a brief poke around in > https://github.com/w3c/web-of-things-framework -- is that > the best place to look at present?) > > Thanks, best > > Hamish > > -- > > *Hamish Cunningham* > Professor of Computer Science, University of Sheffield, UK > +44 7920 765 455 https://twitter.com/@HCunningham hamish@gate.ac.uk > https://pi.gate.ac.uk https://hamish.gate.ac.uk https://gate.ac.uk > > > -- *Hamish Cunningham* Professor of Computer Science, University of Sheffield, UK +44 7920 765 455 https://twitter.com/@HCunningham hamish@gate.ac.uk https://pi.gate.ac.uk https://hamish.gate.ac.uk https://gate.ac.uk
Received on Tuesday, 23 June 2015 08:06:08 UTC