- From: Erik Wilde <dret@berkeley.edu>
- Date: Fri, 10 Apr 2015 10:02:49 -0700
- To: Dave Raggett <dsr@w3.org>, "Lynn, James (Fortify on Demand Security Services)" <james.lynn@hp.com>, Public Web of Things IG <public-wot-ig@w3.org>
hello. On 2015-04-10 05:51, Dave Raggett wrote: > XML or JSON aren’t sufficient by themselves. We need semantics, data > formats, and bindings to scripting languages and protocols, e.g. HTTP, > Web Sockets, CoAP, MQTT, STOMP and XMPP. We also need a consistent > approach to security, for instance, see: Bruce Schneier’s post: "The > Internet of Things is Wildly Insecure — And Often Unpatchable" well, if the goal is to have a "unified world model of things", which seems to be what dave is suggesting, then step 1 would be to define such a model, in some shape or form. step 2 then would be to represent that model in some shape or form. that could be based any kind of metamodel that's around, including XML and JSON (and of course RDF as well). the value, though, would be in the shared model, and not in the chosen metamodel. to me, the bigger question is whether it's a sensible mission to give ourselves. nobody doubts that if everybody spoke the same language (i.e., used the same conceptual model when talking about the world), then things would be a lot easier. that's true for machines as well as, very generally speaking, for humans. however, so far we have many lessons of where this approach failed, but admittedly also some where it succeeded. often, the biggest problem is that conceptual models are always contextual and thus cannot be created independent from usage, and that the more users you have, the more contexts you have. another big problem is semantic drift, where concepts change over time in subtle and often unforeseeable ways. so, my opinion is that trying to find/invent the "HTML for IoT" might end up being as successful as inventing the "HTML for Services" and the many other places where this has been tried: not very much. only because it would be nice if something existed does not necessarily mean that it is very likely that it will be easy or even possible to create it. cheers, dret. -- erik wilde | mailto:dret@berkeley.edu - tel:+1-510-2061079 | | UC Berkeley - School of Information (ISchool) | | http://dret.net/netdret http://twitter.com/dret |
Received on Friday, 10 April 2015 17:03:15 UTC