Assembly - videochat tomorrow

[Greg Bloom]

Hi folks - 

We have an Assembly video chat scheduled for 11am PST (2pm EST) tomorrow, via Google Hangout. Here is the link.

We can only host 10 people at a time, but this time we'll record it so it will be viewable after the fact. NOTE: this is NOT going to be a formal decision-making meeting -- don't worry if you can't make it. I'll be hosting these (and Orientations) on a regular basis.

I had a number of people indicate their interest and availability, but if you'd like to participate, please email me directly again with: a) your google email address, b) indication whether or not you have had a chance to read at least the documentation on our website, and c) any questions you would like to ask.

As our first Assembly, this is probably going to be pretty similar to last week's Orientation - we'll take less time for introductions and more time to dig in to details. But since we'll still have new people getting up-to-speed, we probably won't have enough time to really have a lot of back and forth; most likely, I'll simply invite everyone to go around one-by-one at least a couple of times to ask questions and respond. Any good questions that come up will get reported back to the list here. 

Thoughts? This is all still new, so I'd welcome your feedback on format.

~greg

[Greg Bloom]

Hi folks - 

We had a good (wonky) Assembly last Friday. My notes about it are pasted below this message. 

I'd like to solicit your feedback on the last two video calls, the Orientation and Assembly. 

If there were enough interest, I would be willing to host them on a regular basis, even weekly -- Orientation one week, Assembly the next -- or we can do each one once a month (i.e. a monthly Orientation call for newbies, and a monthly assembly for deeper discussion of the initiative). 

So, you, reading this: I want to hear from you. Newbies, would you like an opportunity for an "Orientation" this week? Folks who are watching closely, do you feel like there is enough going on right now that you'd like to see another "Assembly" next week? (Bonus question: do you have suggestions for the format of these video calls? What do you want to talk about?) 

I could draw up a survey about this, but let's keep it direct for now. Respond to the list or you can just email me directly.

Okay - notes about the Assembly chat are below. Thanks!

Greg

#

NOTES FROM ASSEMBLY

[meta note: next time, I'm going to ask someone to take notes and report out to the list afterwards.]

There is a video recording here [http://www.youtube.com/watch?v=VUrPDrg3E1U&] — it’s long and not super exciting, but if you're so driven, you can at least skip the first 10 minutes.

I have included a (clunky) transcript of our chat, pasted below — in there, I kept a running tab of most of the key points being made. 

Much of the conversation revolved around the identity of our domain. A point was made that there are actually three interlocking layers of an information-and-referral system -- a resource directory, the taxonomy of that directory, and then the means of referral — and, from some perspectives, these are all entangled. 

In our effort, we're aiming to disentangle the resource directory domain, at least to some degree, by developing a common vocabulary to describe services (resources), so that this data can be exchanged and shared among a variety of systems that may have their own means of classifying and referring people to services. We discussed the logical twist embedded in the name 'Open Referral' -- which, at least in this initial stage, won't actually address the subdomain of data about client cases and referral actions. This was really a continuation of our discussion from this thread.

There seems to be a shared sense that it's okay to put this boundary around this domain in order to work with a degree of modularity -- but that we should be mindful of the implications of this. 

I want to highlight Lisa's point that the prospect of a common model with a 'lowest common denominator' set of data points seems like a viable approach to standardization, but also seems like it may actually yield real-world obstacles and pain for people on the ground. I share this concern. I also note that we're designing a process so that people on the ground will actually test this as it goes -- so we'll see real data about it soon enough.

I also note that Derek raised a provocative and challenging point that merits further discussion -- suggesting that anchoring our process around pilot projects and user stories may actually warp the process, as opposed to "looking at the world DIRECTLY." That kinda sounds to me like an ontological wormhole, but I recommend that anyone reading this far also read Derek's recent blog post: http://humanserviceinformatics.wordpress.com/2014/04/23/the-three-patterns-of-holistic-human-service-information/ ~ and I'd like to think about how to create space within this forum to hear more from his perspective. 

That said, overall, it was a helpful discussion that highlighted some of the challenges but also (for me, at least) affirmed the prerogative to act and learn.

Okay - chat transcript below.

~greg

###

me

11:15 AM

Public Documentation : https://docs.google.com/document/d/17cJxF_1P6fafcsFJQERFQifKKc_kPbAKmAXwe2LWDcI/edit#heading=h.a937466uk7p6

FAQ: https://docs.google.com/document/d/1VfoH8a5GL0ZKCNMzm1Y_oiezvDxsHkhLB9MeWGjuNlI/edit#

How might we test these use cases (shared by Dale)?

What is the role of the taxonomy in this initiative?

Derek Coursen

11:22 AM

Main question is about putting sub-boundaries around sub-projects.

seems like there are various sub-projects here

me

11:23 AM

Eric asks: what role might Linked Data play in Open Referral's plans?

Derek Coursen

11:23 AM

One has to do with cataloging I&R records and using taxonomies. Another has to do with I&R program records, sharing and exchanging them.

me

11:23 AM

Dale in response to Eric: linked data would be important in integrating data about referrals and outcomes among clients.

Derek Coursen

11:25 AM

Yes

Right

Lisa Glass Kornstein left group chat.

me

11:26 AM

Derek's question in Greg's words:

How do we understand the separate challenges of classifying service records (taxonomy?) and service directory record exchange.?

Lisa Glass Kornstein joined group chat.

me

11:31 AM

Sameer's question: How do we feel about the delineation between directory data (Open Referral's domain - service descriptors) and 'referral data' (the elements and processes of connecting people to services)?

Derek Coursen

11:31 AM

They are distinct areas of cataloging

me

11:32 AM

Dale and Eric agree with this distinction.

Derek Coursen

11:35 AM

3 domains:

1: taxonomy

2: program records

3: connection to clients

me

11:36 AM

Lisa says: the question of whether or not this really is about 'referral' or not is a good one. If you want to connect one individual with one service, that's a referral. If you want to connect different data sets, that's different. I wonder what kind of challenges we face by separating them and not trying a more holistic approach. 

Derek Coursen

11:37 AM

not exactly

there's a directory list of programs

there's a taxonomy used to catalog the programs

somewhere there's a client list

and the client record is, in the service provider's case mgmt system, connected to the directory record.

me

11:38 AM

Dale says: you cannot build a directory divorced from the type of scenarios of the people who are looking for the services in the directory...

Sameer says: is the data without the context valuable? is there value in managing that data. I've been hearing that there might not be enough value. It's important to be capturing the way that people interact with the data.

Lisa says: there's a huge difference for example between abortion and abortion services. One could be the service of abortion, and the other could be the counseling not to get abortion.

Eric says: suggests that 'Open Resource' may be an alternative name. And 'Open Referral' should be developed in tandem.

Lisa says: my concern about taking a minimal information approach is that we might be putting more of a burden on people who are trying to help.

Derek Coursen

11:52 AM

Is the goal to build a set of tools that can be picked up by independent players who want to use them in whatever way meets their organizations' needs? If so, then the various domains need to be kept separate and modular. One group of agencies might just want to adopt the open taxonomy and use it with their existing directory system; another group of agencies might want to pick up both an open taxonomy and a standard for directory records; some agencies may want to integrate both (taxonomy and directory records) into their case management systems. This kind of flexibility requires modularity.

me

11:54 AM

Eric asks: what about Open Eligibility? And not just the taxonomy that exists, but a protocol for communicating eligibility requirements, matching that up with a person. One-e-App does this, but it's not open source and not a protocol.

Derek Coursen

11:56 AM

Even if this were a project to build a giant behemoth system to store all data about everything, it would still make sense to design and plan and build it in a modular way, just for project management reasons.

Eric Jahn

11:57 AM

I agree, Derek. This is ambitious, and needs to be split up into related parts.

Derek Coursen

11:58 AM

I think OpenReferral is still fine. All the parts do lead up, one hopes, to a successful referral.

Eric Jahn

11:58 AM

or enrollment

Derek Coursen

11:58 AM

right

Dale Fitch

11:58 AM

I agree - I think OpenReferral ensures we don't draw a boundary limit just around a directory.

Eric Jahn

12:01 PM

I have to leave get to another call. Take care all, and talk soon!

Sameer Siruguri

12:02 PM

i have to leave too ... greg: see you at 1ish.

Lisa Glass Kornstein

12:02 PM

Thanks Greg!

Lisa Glass Kornstein left group chat.

Sameer Siruguri left group chat.

Derek Coursen

12:02 PM

I have a few minutes

Eric Jahn left group chat.

Derek Coursen

12:08 PM

A couple of core concerns are (1) need for modularity; and (2) danger that having this driven by on-the-ground pilots might lead the structure of the artifacts to serve only the pilot sites and not be truly generic and useful for a wide range of settings.

Yay!

How wicked is this problem?

Will there be THINGS that can be picked up, neatly, from the pilots, and carried over, neatly, to other organizational and technical environments?

I have been making an argument toward a domain model in other areas of the human services. I would make the same argument here. So I don't agree with Sophia on avoiding a logical model.

me

12:19 PM

re the plan: Dale, says make sure this stays through the evaluation phase - articulate the evaluation parameters as best you can. because that way when you get lost in technical land. There could be a tendency to reconceptualize based on your tech.

There are semantic and conceptual linkages in those scenarios and how they are related to services. those use cases are tests of the database, and the search interface of the database.

Derek Coursen

12:21 PM

My (heretical) opinion is that use cases are good for designing interfaces and application logic, but not for domain modeling. A use case can only confirm or diconfirm or extend an aspect of a domain model; but the domain model itself comes from looking at the world DIRECTLY (WITHOUT looking THROUGH the requirments).

Perhaps you could read my latest blog post and think about the role I might play.

OK

Thanks!

Dale Fitch

12:31 PM

I hope we get to see you next time Derek.

Derek Coursen

12:31 PM

Lacking a microphone and camera make me identify with the voiceless and the invisible. This has been a Freireian experience. Perhaps liberation will come when the oppressed learn to type really, really fast.

Ciao!