On the name and purpose of Open Referral

[Greg Bloom]

[Note: this may seem obvious to some, but if anyone is wondering about the fundamentals as we see them, i'll expound a bit...]

Dale Fitch said something important in his introduction:

In principle I am very attracted to the idea of Open Referral; however, I believe at this time we are talking more about directory standards as opposed to actually doing referrals. That's fine as long as we remember the difference between the two.

Indeed. I think there is clear logic in the name 'Open Referral,' but it does have a twist, so this bears repeating: 

The traditional name of this field is 'Information and Referral,' which indicates the distinct but conjoined challenges of aggregating service directory information and helping people successfully get and act upon it. This field developed long before not only the internet but even computers -- as I understand it, the practice of 'Information and Referral' began in libraries early in the 20th century. Today the service landscape is far larger and more complex; and on the other hand, we also have the internet with its expanding reach and seamlessness. The challenges of 'information' and 'referral' can and should be disentangled. And I think we can all agree, normatively, that some core set of information about health/human/social services should be accessible to all in as many ways as possible. 

In other words: we call it "Open Referral" because we want to see a world in which there is an open set of ways for people to find good information about services available in their community. Our strategy assumes that establishing standards for this information (and the technology to share it) is a critical step on the path to creating that world. On the flip side of that, Open Referral's scope (at this pilot phase, at least) should not actually prescribe any methods for referral. Our concern is strictly in addressing the information challenge. 

(If we're successful in this, then the tension that Dale and Derek and others point to -- the question of data scraping and theft -- will be made more or less irrelevant: the information will be more valuable because everyone can use it, and so the concept of it being 'stolen' would no longer make much sense at all.) 

As we learn and build, this standard will evolve in ways that we might not anticipate from here. But at this stage, we have a narrowly defined scope. This period of time right now is an open invitation for you all to review the documentation to this effect, and chime in with any thoughts along the lines of a 'requirements analysis' that Dale and others have suggested.

For bonus reading, if you're interested: we discussed this and other topics in a thread last November, when things were just starting to come together, and when we first suggested the name 'Open Referral' - if you're interested, you can read that discussion here. 

~greg

[Derek Coursen]

Greg,

 

I actually have very deep misgivings about this way of framing the issue.  To formalize the argument below:

 

TERMINOLOGY  "Open referral" term COMES FROM "we want to see a world in which there is an open set of ways for people to find good information about services available in their community" AND THEREFORE "if we're successful [AT INSTITUTING OPENNESS] ...then... the question of data scraping and theft... will be more or less irrelevant... [because] "the concept of it being 'stolen' would no longer make much sense".

 

My objection is that here we have a word "Open" which has a great many meanings. It's an awfully attractive word.  (Can I have a show of hands for everyone who would rather be "Closed" than "Open"?) But we don't really know what it means. That's a danger sign.

 

Your argument implicitly suggests that we are at a place where we all somehow know that "open" implies "free sharing of information without any property rights pertaining to the information".  In some contexts, I would be perfectly willing to participate in that argument. I might be for it or against it. But whether I'm for or against is not really germane.  The points I would make for our purposes are:

(1) It represents what I referred to in my introduction as a "narrow ideological perspective".

(2) The presence of hidden ideological positions often torpedoes projects. It's better to surface them, then either resolve them openly or proceed in a way that doesn't depend on them.

 

For that reason, I think it is very important to keep the issue of theft of information on the table where everyone can see it.

 

Derek 

 

--
You received this message because you are subscribed to the Google Groups "OpenReferral" group.
To unsubscribe from this group and stop receiving emails from it, send an email to OpenReferral...@googlegroups.com.
To post to this group, send email to OpenRe...@googlegroups.com.
Visit this group at http://groups.google.com/group/OpenReferral.
For more options, visit https://groups.google.com/d/optout.

[Steve Eastwood]

I couldn't agree more with Derek and I'll go just a bit farther.

I&R databases are the life blood of the agencies that currently own and operate them. It's how we do something that's of some demonstrable worth for our communities and, in exchange, find ways to monetize those databases to keep our efforts moving forward. To say that data scraping and theft would be "irrelevant"... well, you might as well just ask us to open a vein.

Individually, I&R agencies already open our databases for people in need "to find good information about services available in their community," so I'm frankly confused. I'm all for openness and transparency and, as my data partner would say, "Information for the people!"

Seriously, she says that.

That being said, to what end is this openness intended? What isn't I&R already doing that Open Referral intends to do?

[Greg Bloom]

Derek - 

I can see how you read it this way, and I suspected you were signaling this concern in your introduction :) ~ so I welcome this discussion. 

I want to leave space for others to chime in, so I'll keep my response simple: I think that Derek's question -- how do we define 'open'? -- is itself an open question. In this initiative's documentation, this is made explicit -- see below, and in the FAQ page here. Some aspects of 'open' I expect we'll all readily agree on; others that may vary by context. 

As per our proposed process, we'll explore possible answers to the question through experimentation, learning, and the deliberation of stakeholders. 

I encourage y'all to check assumptions here and identify concerns. We want to hear them; there's time and space to recalibrate. That said, we're going to answer tricky questions by doing stuff and evaluating the outcome. 

~greg

##

What is Open Data?

‘Open data’ is a popular term right now. It’s also ambiguous. What does it mean?

Open indicates availability: we have “open access” to things like roads and libraries — these are public goods, and anyone should be able to use them.

Open means ‘free,’ as in ‘free speech’: we are all entitled to it by fundamental right.

Open does not necessarily mean ‘anything goes’: you’ve gotta return books to the library, and in good condition too. Even on open roads, there are speed limits, and eventually there are tolls, plus construction and cleanup crews, etc.

Open does not necessarily mean ‘free’ as in without cost. For something to exist in an open state, a lot of energy and resources must go into keeping it so.

Open data can mean many things, but at its core, open data entails:

Access: open data must be available as a whole and at no more than a reasonable reproduction cost, presumably downloadable over the internet. The data must also be available in a convenient and modifiable form.

Reuse and Redistribution: the data must be provided under terms that permit reuse and redistribution, including the intermixing with other datasets. There should be no discrimination against fields of endeavor or against persons or groups. The data must be machine-readable. The data can be licensed to prevent changes and/or to ensure clear documentation of changes.

This is a modified version of the Open Knowledge Foundation’s Open Data definition. 

Openness entails a state of possibility: by making data available for anyone to use it, we can dramatically broaden the scope of its potential value. We assert as a given that the production of open referral data can, will, and should happen. But we also assume that such value can (even must) be realized in such a way that the cost of its reliable reproduction is sustainably carried into the future, in order for this data to effectively meet our collective needs.

For the purposes of Open Referral, the concept of ‘open data’ is itself open to some degree of interpretation. Essentially, we are asking: how should this data be open?

--

• gjb •

202.643.3648

[Fitch, Dale K.]

Dear Group,

If I might parse it a bit more, I am okay with sharing information. However, there is a difference between sharing information and taking information from those whose labor produced the underlying data when paying for that labor is not acknowledged. From what I understand from the documents previously posted about the San Francisco implementation, people are bringing datasets to the table to share and people are mapping out a schema for an API. That’s nice; that’s sharing. But to go and scrape data originally produced by a non-profit that had to pay someone to keystroke every byte, and not to ask for permission, is illegal.

 

And that brings me back to a question I previously posted regarding the SF data, how often is that data updated? It is relatively easy to produce a one-time directory. It is much more difficult to keep it updated and accurate. APIs can potentially save a lot of labor and cost IF people are willingly sharing their labor. Note I said “labor” and not “information.”

 

Dale

[Devin Balkind]

I'm registering my objection to the idea that copying publicly available information about nonprofit services is "theft". 

Here's the Wikipedia definition of theft.

In common usage, theft is the taking of another person's property without that person's permission or consent with the intent to deprive the rightful owner of it.^[1][2]

Britanica has a similar definition.

People have the ability to publish information under whatever license they choose and I think people should respect each other's licensing choices.  However, there are fair use limitations to copyright that have been established to include search engines and archiving.

Fair use is a limitation and exception to the exclusive right granted by copyright law to the author of a creative work. In United States copyright law, fair use is a doctrine that permits limited use of copyrighted material without acquiring permission from the rights holders. Examples of fair use include commentary, search engines, criticism, parody, news reporting, research, teaching, library archiving and scholarship. It provides for the legal, unlicensed citation or incorporation of copyrighted material in another author's work under a four-factor balancing test.

Even if economic conditions and business models make people feel like these issues are black and white, legally they certainly are not.

"What isn't I&R already doing that Open Referral intends to do?" is a great questions that needs to be addressed clearly for this project to succeed. 

My own answer is this: when Superstorm Sandy hit my community and I wanted to create a database of human services for people who were affected, the only taxonomy I could legally access and use was that of openeligibility.org.

Going deeper, I wonder why I can find in-depth information about restaurants in my neighbourhood, including their location, hours, menus, reviews etc, and get that information from a variety of data providers and a variety of devices, but I can't do the same for nonprofit services (even though many of them are funded by government through taxes).  I can't speak for the Open Referral project, but from my perspective, the answer to "What isn't I&R already doing that Open Referral intends to do?" is to create conditions whereby Yelp-style websites for human services can emerge.

Devin Balkind
Founder & Director

Sarapis
134 Spring St, Suite 302
New York, NY, 10012
m.917.748.1048
de...@sarapis.org

sarapis.org

[Steve Eastwood]

On Friday, April 11, 2014 11:18:30 AM UTC-7, Devin Balkind wrote:

Going deeper, I wonder why I can find in-depth information about restaurants in my neighbourhood, including their location, hours, menus, reviews etc, and get that information from a variety of data providers and a variety of devices, but I can't do the same for nonprofit services (even though many of them are funded by government through taxes).  I can't speak for the Open Referral project, but from my perspective, the answer to "What isn't I&R already doing that Open Referral intends to do?" is to create conditions whereby Yelp-style websites for human services can emerge.

Well, if you lived in Arizona, you'd have something quite similar. My agency's database has free smartphone apps which will allow you connect to our website, dial our call center, or sort your search results by distance from your location using GPS. Our vendor is also working on adding user reviews as with Yelp.

While not every database vendor has such apps yet, some do.

[Clive Jones]

Steve,

You probably know what I am going to say ... (and I also agree with Derek!).

I&R data is available for free for people visiting our websites but it cannot be easily moved between different applications and software programs in order to maximize its potential. If your state's department of social services wanted to integrate your data into their case management software in the same way that Google maps are integrated into your own software/web interface -- it can't happen under its present structure/attributes. It needs to be 'open' to other systems.

So at least each I&R should have a choice. A creative commons licence might impose your own conditions on how that data goes out -- no reengineering, no changes (though please submit info on things that need changing) -- and maybe if X is a nonprofit using it for Y, you are happy for it to be "free" and maybe there is a fee for big health if they want to use it. And this way, we do not have the work currently involved in providing access to the data.

Clive

--
Clive Jones
AIRS
www.airs.org
www.twitter.com/AIRSplace

[Derek Coursen]

Dale -- I appreciate your emphasis on sharing labor.  That's why I thought Greg's CFA article referring to a possible "cooperative" (which I summarized at http://wp.me/p3z00P-aW) might be a powerful basis for solving the I&R conundrum.

 

Devin -- Certainly the balancing test about copyright is how the courts determine these things. (And I think about that frequently when trying to save my students $ on classroom materials!) Numbers 3 and 4 would point to the relevance of I&R economic conditions and business models to determine whether copyright has been infringed.

1. the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
2. the nature of the copyrighted work;
3. the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
4. the effect of the use upon the potential market for or value of the copyrighted work.

Greg -- Thanks, yes, let's do put some careful thought into defining what we mean by open. As that conversation evolves, I think clarity will come into view.

 

 

Derek

On Fri, Apr 11, 2014 at 1:23 PM, Fitch, Dale K. <fit...@missouri.edu> wrote:

[Hailey Pate]

Hi all, 

I am not an I&R professional, but I know well the painstaking work of curating and maintaining meaningful datasets. A few of my thoughts in response to this thread: 

-- I am grateful for the work done by community-minded I&R organizations. Thank you all for that you do! 

-- I would argue that knowing 2 things about a resource record -- its source and the date it was last verified -- comprise the bulk of its value. No wise developer wants to risk his or her reputation by serving up unsourced, potentially unreliable data. I would like to see experienced resource curators leverage this in their business models. I've never known open data to mean uncredited data.

-- As more government agencies adopt open data policies, is it possible that public-funded I&R organizations will be expected to make certain data assets open where appropriate*? I think it's coming. Why not get out in front of it now and experiment with new business models based on user-friendly digital publishing with attribution, before open data becomes a condition of funding?

*Please note emphasis on 'where appropriate'. Disclosure of personally-identifiable information (about 211 callers, for example) would not fall under the "appropriate for open data" category.

[Fitch, Dale K.]

Derek,

Thanks. And I've mentioned one of these potential update models before, but I'll reiterate again. In the Michigan implementation we used the RTM product. A feature of this database was to send automated emails (i.e., no labor) to agencies on a quarterly basis with the email message containing the directory information for the agency. If no changes were needed, i.e, the data was accurate, nothing happened. If the data was inaccurate, then the agency corrected the information and it was returned to the database as an update. A confirmation email or phone call might be done by a 2-1-1 operator, i.e., the labor might be a phone call only.

The beauty of this model is it personifies the "information push" typology, which has been shown to reduce labor costs and increase data accuracy. As you might imagine, "information pull" requires more labor and is less accurate.

What is the agency motivation? Much like your description of community-based outcomes, the 2-1-1 call center ran quarterly reports for the community called the "unmet needs" report. Reasons might include caller ineligible, no funds available, waiting list, etc. I know these are not true client outcome measures and separate provisions were made for those. These reports were reviewed by an interagency (city, county, non-profit, state) funding board and funding determinations were partly based on the "no funds available," and "waiting list" criteria (and other measures were used for triangulation).

Two, not mentioned by you but discussed by others, the 2-1-1 directory database was freely shared with a public entity and a non-profit. Doing so was seen as a way 'by the community' to reduce costs for agencies that already had insufficient service funds. We could have executed MOUs, but the community was very collegial and it was not necessary. We used ODBC for one of the connections and perhaps an API would be better now.

In sum, I firmly believe this community already personified an open data commons. Whether they are still doing it or not, I do not know. It also sounds like other communities are sharing data. I also know that some communities do not exemplify this behavior.

Dale

---

From: OpenRe...@googlegroups.com [OpenRe...@googlegroups.com] on behalf of Derek Coursen [derek....@nyu.edu]
Sent: Friday, April 11, 2014 1:43 PM

[Steve Eastwood]

On Saturday, April 12, 2014 3:03:47 PM UTC-7, Dale Fitch wrote:

Derek,

Thanks. And I've mentioned one of these potential update models before, but I'll reiterate again. In the Michigan implementation we used the RTM product. A feature of this database was to send automated emails (i.e., no labor) to agencies on a quarterly basis with the email message containing the directory information for the agency. If no changes were needed, i.e, the data was accurate, nothing happened. If the data was inaccurate, then the agency corrected the information and it was returned to the database as an update. A confirmation email or phone call might be done by a 2-1-1 operator, i.e., the labor might be a phone call only.

"No labor" or the only labor being a phone call only. Sounds magical!

Here's what happens with the system we use here in Arizona (Bowman's ServicePoint and CommunityPoint) which is similar to RTM's set up:

Set up accounts for providers after verifying who they are so we know they're authorized to send updates about their records.

Select the update requests we're going to send out (we have a large database and split up our update requests), modify the email that goes out to providers to include the update cycle's deadline and other pertinent information, then send them.

Weed through all the undeliverable requests to clean up our list of authorized representatives and resend those requests by mail to someone else at the agency.

Set up new people with online accounts from the requests mailed out.

Run support for the people who can't figure out how to sign in or do updates even though they have perfectly good instructions in front of them and often readily admit they didn't bother to read those instructions before complaining they couldn't figure out the process.

To comply with AIRS standards for having documentation that we've reached out to and connected with each provider at least once a year, providers are asked to submit a "no changes" report on records they're responsible for if there are no changes. Those must be processed by resource staff manually to document that the correct person has told us there are no changes.

If there are changes, we must manually approve/modify/deny/follow-up on the requested changes based on our database standards and community needs then educate providers on what kind of information we're looking for (straightforward descriptions of services, for instance, not PR fluff pieces about how wonderful they are).

Less than 20% of my providers update online despite that being our preferred (and most pushed) method. Many providers simply would rather update by mail or fax. Others want to update online but lack the technical savvy to do much beyond post cat videos on Facebook.

Before the current model for 2-1-1 services here in Arizona, the state government tried to implement a system. They were going to go with 100% online updating. Their attitude was, and I'm quoting here, "everyone has email." As it turns out, no, not everyone has email. And many of those who do don't want to get emails from us telling them to update their information online. And, as pointed out above, some who do have email and are willing to cooperate with online updating should never be allowed near a computer for the sake of all humanity.

In an ideal situation, the no/low labor push model would be lovely to use. In reality, it just doesn't work that way.

[Fitch, Dale K.]

Steve,

Thanks for your post as it serves to show that different implementations deal with different realities. Why the implementation in Arizona had those difficulties I do not know. Clive can speak to this better than me, but from my experience talking with folks who have implemented 2-1-1 applications and, also, HMIS applications, the experiences differ greatly by locale.

Regards,

Dale

 

From: OpenRe...@googlegroups.com [mailto:OpenRe...@googlegroups.com] On Behalf Of Steve Eastwood
Sent: Monday, April 14, 2014 11:29 AM
To: OpenRe...@googlegroups.com
Cc: openre...@googlegroups.com
Subject: Re: [openreferral] On the name and purpose of Open Referral

 

--

[Cat Dwyer]

I think Steve's experience is more the norm than the exception!

--

Cathleen Dwyer, CRS, CIRS

[Greg Bloom]

Hi Cat - I'm nudging everyone on this list to introduce themselves in the introductions thread, and I'm going to be extra insistent with everyone who actually chimes in on other threads: we welcome participation, but first we want to know who you are! Please head over to the introduction thread to share a bit about your experience :)

~greg

• gjb •

202.643.3648

[Steve Eastwood]

Yes, and just to clarify: We didn't have "difficulties" implementing our online updating. That's simply the process. We moderate changes submitted to us and we make it as easy as possible up front for providers to submit changes or verify there are no changes to be made. We're not a wiki.

To post to this group, send email to OpenR...@googlegroups.com.

Visit this group at http://groups.google.com/group/OpenReferral.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "OpenReferral" group.
To unsubscribe from this group and stop receiving emails from it, send an email to OpenReferral...@googlegroups.com.
To post to this group, send email to OpenRe...@googlegroups.com.
Visit this group at http://groups.google.com/group/OpenReferral.
For more options, visit https://groups.google.com/d/optout.