Deborah, (01)
> Very interested in this - have you seen the ProjectHaystack tags? (02)
Yes, I'm been watching that project for while. As described by Brian Frank
back in May 2011: (03)
> Haystack is focused on defining a basic taxonomy or ontology for building
>automation energy and environmental systems. The goal is to lay the groundwork
>for how data can be modeled to enable data mining and analytics. We achieve
>modeling through the use of applying one or more tags to the "entities". Tags
>are the unifying mechanism for many different purposes:
>
> • model semantics such as this sensor is for sensing discharge
>temperature
> • relationships between entities (this VAV is supplied by this AHU)
> • modeling metadata such as units of measurement
> So tags are the model/ontology. Haystack does not specify any database or
>protocol for what those entities are. (04)
To folks in this audience there are glaring and enormous problems with this
approach. That they make little or no distinction between a taxonomy and
ontology, or that there is more to modeling than applying tags, is a indication
of the kinds of awful soup of unorganized bits of cruft they (we!) deal with on
a regular basis. (05)
To their credit they have built a fairly large community of regular
participants and have cooperated on a set of tags. At some point they might
make the transition from folksonomy to ontology, but they have a long way to
go. At Cornell we have naming conventions for our things that go pretty far,
but it is inconsistently enforced because of the velocity of changes on our
campus and that names have to be applied by humans, and they are subject to
individual subjective bias. What is the naming convention for an "air
conditioning unit discharge temperature high limit alarm"? (06)
To solve that problem we a Gellish-like comprehensive list of all the terms
that are in our industry along with the Cornell naming convention for each row.
Of course it would be nice for MIT and Stanford and Yale to add to the list
with their own naming conventions, where we can boil it down to an ontology (so
every instance of a "discharge temperature high limit alarm" is necessarily
associated with an instance of a "discharge temperature high limit" which is in
turn associated with a "discharge temperature"). To have an ontology like that
would be TOTALLY AWESOME because we could build the ontology reference directly
into the control system modeling tools and the humans would not be involved in
naming anything other than this thing is called AC-1 and not AC-2. (07)
This is exactly what I have in mind for a future hackathon - and I probably
can't wait until next year - the Building Automation Federated Lexicon (BAFL).
I have in mind something like the AceWiki with a GitHub backend. Everyone in
this industry should be able to pull the system down, use it with wild abandon,
add their own definitions and terms, share their results back up to the cloud,
etc. Where there are referenced standard definitions for terms (like this
stuff http://wiki.ashrae.org/index.php/ASHRAEwiki) there should be a way to add
citations, but put your own twist on it - "we don't call that 'discharge'
around these parts, that's only the stuff that comes outta pipes!" (08)
Note that this hackathon is not that one, in my mind this one has to come
first. Thank you for your interest! (09)
Joel (010)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2013/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2013
Community Portal: http://ontolog.cim3.net/wiki/ (011)
|