Initiating a minor upgrade for Open Referral's Human Services Data Specification

[Greg Bloom]

Hello, friends! I'm excited to share that we are going to conduct a quick development cycle for the Human Service Data Specification and API protocols in March. 

The summary version of this message is: I'm proposing we have an introductory kickoff call on videochat this Thursday (2/27) at noon EST. Then we will start properly reviewing issues on a working session video call next Thursday (3/5) at noon EST, and then one final call on March 12 at noon EST to field final feedback before developing a proposal, which will go to the whole group with two weeks for comment.

I'm eager to solicit your participation and feedback to make this process as effective as possible.

Now for more details: 

This cycle has been requested by our new collaborators in the UK, led by Porism and the Local Government Association, with support from the UK Ministry of Housing and Local Government, as described here. They have submitted a range of issues to our queue, and offered a small amount of funding with which we will enlist the Open Data Services Cooperative to resume their role as technical stewards of our schema.

Working with Rob Redpath and Ben Webb at Open Data Services Coop, I'll co-facilitate a process of reviewing these issues, fielding feedback from the community, prioritizing and/or tabling depending on complexity and the tenor of the feedback, and submitting a proposal for an upgraded specification to the community to review.

This will be a brief cycle, limited to a relatively small set of issues. Some important topics (like data editing and verification processes) are almost certainly going to be ruled out of scope, to be addressed in the hopefully-near future. That said, some important issues, like how we structure our taxonomy table, might be within scope! Review the issue queue to get a sense of where things are, and/or reach out to me directly to discuss your priority issue(s) and other questions.

Open Data Services Cooperative has suggested a quick series of weekly video checkins between now and mid-March, starting this week, on Thursday 2/27, and I'm proposing noon EST, which is a time I have open and one that reasonably stretches between Pacific West Coast and Greenwich Mean Time. (Australian friends, we'll check in to see how to involve you too as desired.) This first call will be very introductory and high-level; on next week's call (prospectively, March 5th at noon EST?) we will start digging in. We'd follow that through with a final call on March 12th to field any last input before Open Data Services Coop develops a proposal. Once the Coop develops their proposal, it will be shared with the community with two weeks for comment before being finalized; if we don't already have rough consensus by the end of the third call, I'll loop back with a proposed process for the RFC period. But, barring surprises, for now I expect things to go smoothly :)

Please let me know if you have any questions or suggestions!
~greg

--

Leading up the Open Referral Initiative.

Book a time to talk on my calendar.

Reach me directly at 305.962.2859

Web: OpenReferral.org

Community: Open Referral Forum

[Greg Bloom]

Thanks to those who joined us today. 

Here are notes from the call: https://docs.google.com/document/d/1N3Vg_ZOXGkWuF8L2rVpLInXpKtfHgRIcIMA-y_7qPqg/edit ~ Next time I will aim to get a video recording.

Our next two calls will be next week, thursday March 5th at noon EST, and then three weeks later, Thursday March 26th at noon EST.

Then Open Data Services Cooperative will put together a set of proposed changes, and we'll have another two weeks for comment before final revisions and approval. 

Let me know if you have any questions or suggestions!

onward,

~greg

[Greg Bloom]

Hi folks – I hope you all are safe and sound. 

We have scheduled a call to review proposals for an upgrade to HSDS this Thursday at noon ET.

We set this HSDS upgrade cycle in motion one month ago, and things have changed dramatically in that time. If someone out there has strong opinions about the proposals on the table, yet can't participate right now due to pandemic contingencies, I'm going to put things on pause. However, I've been checking in with folks who have participated in the conversations so far, and to my surprise nobody yet has said that they aren't comfortable with the schedule proceeding as planned. I haven't heard from everyone, but as of now I'm expecting that we're going to keep to our plan.

Notes from the calls so far are here. 

The main issues being considered are as follows:

• Changing the way HSDS handles taxonomies. Right now, we have one table for a taxonomy, which links to the 'service' table. But implementers have asked to be able to associate classifying schema (taxonomies, code sets, enumerations, what have you) with other entities aside from 'service' (like 'organization,' 'location,' accessibility, eligibility, and others). So one proposal is to restructure the taxonomy table so that it can link among multiple taxonomies and multiple entities. Another proposal is to have a 'service taxonomy' table, and then an other 'taxonomies' table. There might be other variations on this proposal. The most popular proposal would seem to be a 'breaking change,' so I want to hear from as many people as possible on this issue. 

• Changing the way HSDS handles schedules. Right now, we specify data points for weekly and holiday schedules. Various implementers have come across the need for recurring but non-weekly schedules (like 'first monday of every month') or even irregular schedules and one-offs. There are existing standards for things like this (like iCal and RRULE). We're considering our options. 

• Specifying conventions for custom fields. This issue would provide formal guidance for extensions to HSDS, so that implementers can adapt the spec and document their adaptations in a consistent way. Less tricky than the above challenges, but still something we want to hear your feedback upon.

If you're not able to make the call on Thursday but do want to participate in this dialogue, please reach out and let's chat!

Meanwhile, ICYMI, my message to the Open Referral mailing list from today is pasted below.

Stay safe and sound out there please,

greg

#

Hello . I hope that – all things considered – you are doing okay. 

During these very difficult times, I am so glad to be in touch with the network of people in Open Referral's community – many of whom are already working to help people through this pandemic around the country and world. Thank you all for your many different kinds of service.

I don't have an update in this email. Many people have already asked me: what should we do? Given the nature and scale of the need, I don’t presume to have answers to that question. Instead, I want to ask you: 

Are there any ways that Open Referral – and even I, personally – can help you and your community respond to this crisis? 

Especially (though not necessarily exclusively) with regards to sharing information about health, human, and social services available to people in need. Can the tools we’ve put together (like our free resource database management system in Airtable, or our API specifications, etc) be useful to you? Are there other ways that this network might be of use during this emergency? 

Given the disruption to employment, healthcare, and people's daily lives, the distress in our communities is going to be greater than ever before. At the same time, many health, human, and social service providers have fallen out of their normal operations and capacities. Some may be closing down. Others may be changing or even expanding their services. It may be even more of a challenge to keep information about these resources up-to-date. Which may make it even more important for organizations to be able to work together and share data in real-time. 

Toward that end, I can only commit my own time and attention, though I also know a lot of people out there (and in this community) who have more time and attention than usual on their hands right now.

So please don't hesitate to reach out.

Thanks again for all that you do. Stay safe out there.

~greg

[Greg Bloom]

Hi folks – it's been a very challenging season on multiple fronts, and in deciding whether to rush or pause, we've chosen to delay this HSDS upgrade process. I'm sorry for the momentary lapse in communication about the matter.

Right now, the plan is for Open Data Services Coop to resume working on HSDS next month. We hope to have a proposal to bring back to the community by mid-to-late July, for a final comment period. 

Thank you for your patience. The good news is that there is more time for community members to share feedback on the issues at play (see the earlier messages in this thread, or our Github queue). 

(In particular, I've proposed one new feature for us to consider for this upgrade, which we've heard a lot of requests for: Announcements, to enable the sharing of time-sensitive information about services.)

Please let me know if you have any questions or suggestions. 

Stay safe and stand for justice,

Greg

[Rob Redpath]

Hi all!

Apologies for this taking so long.

As you'll no doubt remember, there's three main areas that we're looking at during this upgrade cycle - classifications and descriptions, dates/times and data beyond the standard. The latter two we've broadly reached agreement around, but our first proposals around classifications and descriptions needed more work.

So, we've done some more work, and we've drafted a new proposal for consideration; please do take a look and let us know if this works.

On the specific question of Announcements, I've discussed this briefly with Greg and it's my view that we should try to avoid premature standardisation - but that we encourage people to align with schema.org's work and share their implementations with the community (and ask for support in their efforts) so that we can look at what practice is emerging ahead of the next upgrade cycle.

Thanks, all.

Rob

[Greg Bloom]

Just a reminder that we have a schedule survey for those who want to participate in this final round of discussion about this proposed upgrade to HSDS. Please indicate your interest and availability here: https://doodle.com/poll/wswpidci5ra6mhwm

FYI, I'm asking everyone who participates to read through the proposed upgrades and prepare your feedback in advance (and ideally share it in advance) so that we can structure the agenda in accordance with your interests. See more details and associated links here: https://openreferral.org/release-announcement-benetech-service-net-upgrade/

thanks!

greg

[Greg Bloom]

the Human Services Data Specification version 2.0 is approved!

We have video of the call here for anyone interested: https://youtu.be/Aa0zb_2RA6w

And notes here (which I am still revising): https://docs.google.com/document/d/1N3Vg_ZOXGkWuF8L2rVpLInXpKtfHgRIcIMA-y_7qPqg/edit#

Only one issue came up on this final review, which I will try to summarize here:

Regarding the new method of handling of taxonomy and classification in HSDS 2.0, Kate Lambacher pointed out that I&R systems using the the 2-1-1 Taxonomy tend to support a pattern in which taxonomy terms can be combined together and referenced as a compound term in a unique link. So if there are terms Red and Yellow they can be linked in an term known as Orange. Kate expressed concern that, in the HSDS 2.0 schema, such combinations would have to be created as a new term (say, Red*Yellow) and it would add complexity to unwind that. In Kate's words: "It’s doable, but it would be a shim-job." Complicated and messy.

We discussed this for a while: Andrew Benson suggested that this method of compound combinations could be workable if the method was clearly documented and followed, so that vendors know how to unwind the compound terms in consistent ways. Meanwhile, Mike Thacker of Porism shared that, in the UK, they’re dealing with similar instances in which a type of service needs to be combined with a specific eligibility criteria to represent a compound type, so they’ve created yet another table to handle these combinations; this additional table for compound terms could be developed into a formal or semi-formal extension, perhaps specifically for use with the 2-1-1 taxonomy when needed, without pushing that additional complexity on everyone using HSDS. 

If I've captured all this correctly, I believe this means HSDS 2.0 can offer two methods of handling this complexity of compound classifications: 1) we can provide clear guidance on how to document the method of the "shim-job" of compounding terms, and 2) we can develop an extension table to the spec that specifically accommodates compound taxonomy terms. In the week ahead, I will loop back with all parties to confirm that this holds water. Would welcome any other feedback on this matter.

Thanks again to Rob Redpath and the Open Data Services Coop for once again shepherding us forward!

~greg