EMAIL INTERVIEW – Therese Sullivan and Ken Sinclair

Therese Sullivan, Principal, BuildingContext Ltd

Sullivan:  Brick Schema is another open source organization launched around metadata tagging for buildings. Automatedbuildings readers have been learning about and, in some cases, deploying Haystack tagging for over five years, so the easiest way to define Brick Schema is in terms of Project Haystack. In fact, that is what the Brick Schema developers do themselves. The Brick academic researchers presented their work in mid-November at a conference on the Stanford campus called BuildSys 2016.  According to the paper introducing the demonstration project:

Sinclair:  This is great stuff I think we all have discovered what open is all about.  You do not need to ask to build on someone else’s great open work. People working with open to create "open-er."

Sullivan:  I don’t think anyone asked permission. But, I hope the two organizations will work together toward the goal. All the brain power behind this Brick Schema launch validates that tagging and taxonomy are the right problems to be working on. Project Haystack has the lead on that. And open means open, so "Haystackers" can now copy any concepts in Brick Schema that they like. It is quite a compliment to Project Haystack that all these PhD-types from top universities in California, Carnegie Mellon University, IBM Ireland, etc. studied the landscape and decided to build on the Project Haystack metadata open source effort. It’s an acknowledgment that Project Haystack Methodology is the best technology out there. The Brick research papers say as much. Their charts also draw comparisons to BIM-related hierarchies, but the Brick graph draws directly from Haystack. BACnet has considered standardizing a metadata schema as well. ASHRAE Addendum 135-2012ba provides a way to add metadata to BACnet devices, but this too includes a provision that allows BACnet to access Haystack Tagging.

Now, the open-source Brick Schema organization will add its heavy data science expertise with the same goal of an open metadata schema for buildings. I think that is how open source works? They’ll want to engage even more commissioning experts, master systems integrators, and other building data wrangler-types. Domain knowledge is vital. University data science untethered to experience is not a formula for success. As you say ‘Our Assets are People, not Technology.’ The greatest accomplishment of Project Haystack is its community. Some real innovators in the BAS industry have joined, but many are holding back. Haystack plus Brick Schema would be the start of an awesome community enlargement, I think.

Sinclair:  What do you see as the main point of comparison between Brick and Haystack?

Sullivan:  Like Haystack, Brick Schema starts with a ‘graph’ of building entities.

The simple meaning of ‘graph’ in this context is having agreed-upon names for things in a domain set, and then creating a visual map of those things by defining their relationship for a specific instance, for example, equipment in a building.

To get on the same page as the data scientists about the word ‘graph’—think about Facebook, LinkedinIn, Twitter and their social graphs. In the article that I co-authored with Alper Uzmezler in September, we talked about how Machine Learning would enter Smart Buildings. Data structuring for machine learning happens according to a graph. If you have a winning graph, you may be able to build a Reciprocal Data Application (RDA) around it. This is an app that is so attractive and useful that it gets knowledgeable people to willingly contribute their efforts to the creation of training data to feed machine-learning.

[an error occurred while processing this directive]Sinclair:  So, they both have a graph. How has the Brick Schema team diverged from Haystack and added value?

Sullivan:  The Brick Schema team has added its ontology or hierarchical way of ordering and defining relationships among the agreed-upon names. They have described this further in their paper. It has been used in 6 buildings, mainly university campus buildings. Looking over the BuildingSys presentation on Brick Schema, they would have done well to ask Project Haystack about their reference implementations. They say Haystack has zero implementations. Of course, Haystack deployments are happening wherever supporting members have been doing projects. There are millions of points out there in real working buildings under Haystack-enabled building management. I’ve curated a collection of tweets that sample just some of the largest and most well-known building energy-efficiency projects that incorporate Haystack tagging here.

Sinclair:  Brick Schema seems to have a heavy California flavor with all the UCs involved? Is there a reason for that?

Sullivan:  California has enacted the SB 350 Golden State Standards around renewable energy and energy-efficiency—among the most ambitious climate action legislation in the world. So, yes, the UC’s are expected to help lead. The launch of Brick Schema at Stanford in Silicon Valley is no coincidence either. All the big IT and OT companies are either based or have representation in the Valley, and many are heavily vested in IoT innovation. They are all coming around to understanding that a unified metadata schema for buildings is in their path. The challenge is bigger than our BAS industry. That’s clear.