Cooperation with ESS Global and Value Flows

(Adrien Labaeye) #1

This discussion was initially reflecting on the ESS Global process and its interoperability guidelines.
With contributions from @mariana and @bhaugen it evolved into an information exchange about the state of Linked Open Vocabularies and possible collaboration between ESS Global and the work of @mariana and the work of @bhaugen @fosterlynn.

[Further summary needed]

Background information

I have met on Skype with Vincent Calame (he isn’t active since 1.5 year) on 20.03.2015 and Jason Nardi on 05.05.2015. There is considerable overlap between ESS Global and TransforMap visions. There is will to collaborate on both sides. Let’s gather all useful elements on this thread.
Various parallel contacts have been underway, let’s gather them all here: cc @almereyda @josefkreitmayer @Silke @mattw @jnardi @mariana

Ressources on ESSGlobal

Some bits about ESSGlobal (more coming from my discussion with Jason):

From the discussion with Vincent Calame who has been involved in the development of ESS global a couple of years ago. Here is what he says:

The idea behind ESS Global was to have a common query protocol for different databases of ESS (Social and Solidarity Economy) initiatives. Maps from essglobal.info and those of socioeco.org are two trials to use this protocol (one on a google background, the other on OpenStreetMap with Leaflet). On essglobal.info, only Quebec and Brazil’s databases are queried, on socioeco.org, the French one is also queried.

This idea of a common query protocol is similar (much more simple) to the OAI-PMH.

The difficulty is that each database must do the effort to implement the protocol.

But the most complicated was to agree on a common [taxonomy] to filter along economic activities.

2015-09-28 | Mapping the Commons + Commons Institut | Wikimedia
General communities' requirements for TransforMap
Complementary (?) System: Share n Save
Geodata storage evaluation
(Adrien Labaeye) #4

In May I had a conversation with @jnardi. I failed to report on it right away. Basically Jason reported on their work (see summary below).
Focus should be on scheduling working session between TransforMap and ESS Global to better identify technical and community overlaps during Solikon.
@Silke @almereyda @josefkreitmayer @toka

@jnardi: could you provide useful link to latest tools and info for those willing to get an update.

Summary:
ESS Global is a a group made of 5/6 people at the moment.

  • Daniel Kilgelitz who has created the Brazilian portal on the solidarity economy, he is now for 2 years after being the secretary for the Brazilian forum of the solidarity economy, the main network there, he’s now in a small software company (EITA) that is dedicated specifically for the development of tools in open free software.
  • Greg (not sure of the name) who is working on the US mapping. He is at University of Pennsylvania. He’s part of the US solidarity network.
  • In Canada, Chantiers de l’économie sociale et solidaire. They were active until a year ago. Now we have to put them back into the loop
  • In Portugal Mariana Malta (@mariana) , she is a researcher at the university of Porto, she has a research group. Her doctorate was on ESS Global. She is the one working on promoting the semantic side, on how to create an inter-operable for classifying
    • Actors: networks, initiatives that identify themselves with solidarity economy.
    • Products and services
    • Field of action
    • Thematic approach

As the result of work with 8 national ESS networks, they put together 2 years ago guidelines for interoperability and data exchange: https://trello.com/c/9uy5CB7A/15-understand-the-interoperability-standards-and-script-of-ess-global
After two years, we have a first version, a first vocabulary and a description of the approach.

(Jon Richter) #5

I know her from the CAPSSI application. Very solid in knowledge and supportive with critical questions.

(Mariana Curado Malta) #6

The context of ESSGLOBAL is the Semantic Web (or Web of Data) [1] and Linked Open Data [2]. The Semantic Web has to do with structured metadata (RDF [3]), data that describes resources. LOD has to do with technologies (turtle [4], J-SON [5] and others) that allow this metadata to be connected, building this way an inteligent Web since machines can infer over this data a give answers to complex questions that users can make.

In the end, users or institutions can develop applications over this data…and we can use e.g a smartphone to find a certain information about SSE organisations or activities.

But for all this to work we need to structure the data in a common way, and then to describe it in such a way that it can be interoperable among us and hopefully between us and the world outside our community.

The first step is to understand each other. That was not easy. We needed to find a data model that would be an abstract model of the SSE “reality”. We have defined one, we know it is very imperfect. We are open to improve it. We call this data model a Dublin Core Application Profile (DCAP) [6].

ESSGlobal has though developed a DCAP for the SSE [7] that structures the data of SSE (organizations, networks, relations, products/services, cost composition of the products) and defines the vocabularies that describe this data model. All but one vocabulry are standards and used by the rest of the world: this way we can be interoperable and be present in the LOD ecosystem. We needed to create a new RDF vocabulry called essglobal since there were parts of the data model that could not be described by the standard existent RDF vocabularies.

Next setp is to “use” the DCAP-SSE. For that we need to program SSE Web Based Information Systems (the web sites with the information on the SSE organizations and its activities), translate the info that is on the these Databases (relational databases…whatever) in datasets of metadata, and make them available on the Web of Data.

Where can we cooperate? TRansformaps could put its data in LOD. And for that could follow the DCAP-SSE model.

Regarding your taxonomies you could have a look in our taxonomies (we call it Vocabulary encoding schemes). You can see here the list (in SKOS) [8].

Mariana Malta

[1] http://www.w3.org/2013/data/
[2] http://linkeddata.org/
[3] http://www.w3.org/RDF/
[4] http://www.w3.org/2007/02/turtle/primer/
[5] http://www.w3.org/TR/rdf-json/
[6] http://dublincore.org/documents/profile-guidelines/
[7]http://purl.org/essglobal/wiki
[8] http://purl.org/essglobal/standard
[9] http://maltas.org/essglobal/index.php?title=Arteiros_Aroma_Brasil (http://purl.org/essglobal/website for the home page of the website we are working on)

[Proposal] end 2016 having some components running | mid 2017 - mid 2018 have an RDF-based system running
2015-09-28 | Mapping the Commons + Commons Institut | Wikimedia
2015 11 25 | Berlin | Federating Civic Data ... on the fringes of Share-PSI 2.0
(Mariana Curado Malta) #7

I just would like to say that you don’t need to invent the wheel (if we want to be the most interoperable possible).
If you want to be present in the LOD ecosystem just define what you want to show to this ecosystems (what information, what resources, …). This means that you just need to create an Application Profile where you define what Classes/properties you have and what are the vocabs that describe them. Example: a title of a activity will be described with dct:title (dcterms vocab) and an email of an organisationi with the foaf:mbox (FOAF vocab [2]), etc. ESSGlobal vocab was created because there was no way for ESS to describe e.g. the open price of a product/service.

Mariana
[1] http://dublincore.org/documents/dcmi-terms/ (thee standard from where you should always start if you want interoperability!)
[2] http://xmlns.com/foaf/spec/

(Bhaugen) #8

ESSGlobal looks like the closest thing I have seen to what we are trying to do in the Open Vocab value flows project.

I’m clicking around. Is this the best place to start?
http://purl.org/essglobal/wiki

Will study.

(Bhaugen) #9

Here’s a list of our slide decks published so far.

@almereyda ping.

(Mariana Curado Malta) #10

YES indeed! Check there in the wiki page the essglobal vocab and the UML diagram to see how it is used in our DCAP (sort of an ontology).

(Bhaugen) #11

Lynn appreciated the UML diagram. (We like pictures!) We’ll play with the RDF later today.

What has been done with this vocab beyond publishing it? Has it been used in actual social solidarity economic practice? That would be hugely interesting.

P.S. I liked this; very informative:
http://maltas.org/essglobal/index.php?title=Arteiros_Aroma_Brasil

(Bhaugen) #12

See also https://github.com/valueflows/ovn/issues/24#issuecomment-144100873
longer discussion about conversations for action

I seem to remember some mentions of offers and wants in TransforMap. Did that continue? If so, where is it at now? (Or did I imagine it?)

Social marketplace
2015 12 04 - 06 | Lille | Federating civic geodata
(Mariana Curado Malta) #13

This vocab was one of the outputs of my PhD, the work I have developed with the RIPESS group (see the first publication here [1] - more to come).

So it is very green yet. We need to work on dissemination and help SSE organisations (specially the ones that are in the end of the line, joint organisations like RIPESS - a network of networks) to understand the advantages in publishing in RDF and help them to do that. I need to find human resources to work with me on producing better Web Pages with more information/manuals and with better documentation for non-technical people.

I know that in cirandas.net (implemented with Noosfero open source software [2]) they have implemented a plug-in that publishes some data in RDFa related to the products a SSE org comercializes. But this is not the final goal for them since RDFa is something that is not LOD. It is more only for Semantic SEOs.

The page you have found, the mediawiki pages, still did not define a purl for that since it is under construction. It was built by a student supervised by me during an Erasmus + placement. She was very good and comitted, and she is still working on those pages on a volutary basis after her summner placement. But I need to find budget to have more people working with me. Looking for a European Project program …I am happy that you liked it.

I hope I have answered your question.

Regards,
Mariana
[1] http://dcevents.dublincore.org/IntConf/dc-2015/paper/view/372/361
[2] http://noosfero.org/

(Bhaugen) #14

Yes, thanks. Your model is compatible with ours (the model behind the valueflows vocabulary, which is also a work in process, as you can probably tell from some of the inconsistencies in the github repositories).

We have implemented earlier versions of the model as working software:
http://mikorizal.org/software.html

The participants in the valueflows vocab also intend to produce working software for actual social-solidarity economic networks, based on a common vocabulary.

Maybe we could all collaborate: RIPESS, TransforMap, valueflows? I know, easier said than done, but still…

(Mariana Curado Malta) #15

We really need to cooperate if we want to be interoperable. Where to start?

(Bhaugen) #16

What would be most comfortable for you? You’re welcome to participate in the valueflows conversations, where @almereyda has popped in occasionally as well. Raise issues, float ideas. And we appear all to be here in this TransforMap Discourse environment, where we could do the same kinds of things.

Do you have other environments where we could all congregate now and then?

I think the valueflows conversations can always use help from somebody else who knows Linked Open Data very well. I am very much a newbie. @elfpavlik is our teacher, but might appreciate some help herding his unruly students…

But I’m open to other ideas.

Beyond the conversations, though, we want running code. And real social-solidarity organizations using it. All interconnected in networks of networks.

(Bhaugen) #17

@mariana - @elfpavlik posted an issue here which you can take as an invitation to talk about ESSGlobal.

(Mariana Curado Malta) #18

I am sorry for being away for so long. I can’t cope with so much work. I am trying hard to be active here but it has been difficult!

I think all your “vocab” are conceptualisations. Did you deefine “new” terms (properties and classes?). For instance the Agent Vocan could be described with the FOAF RDF VOcab terms?

Can they be used in LOD? Have you specified them in RDF/XML or turtle…or do you have and I don’t see it? I am sorry if I am being not clever here.

I need to understand better your VOCABs and in what context/paradigm you are in order to answer the “how to start question” question.

(Bhaugen) #19

We’re using existing vocab terms where they fit, including FOAF:Agent. And adding new ones as needed.

For now, json-LD. We start with models and examples, and then derive json-LD. And Lynn is also working in RDF.


We’re generating json-LD from code for Agents and their Relationships and our current priority is to do a round trip where one app outputs json-LD and another app uses it as input and does something with it.
Example of output: http://dhen.webfactional.com/accounting/agent-jsonld/

Still work in process, incomplete, but getting close (or so we think).

We’re just getting started, really. Much more to come.

P.S. @mariana I see you popped in here: https://github.com/valueflows/valueflows/issues/31#issuecomment-150617691

Thank you very much!

(Lynn Foster) #20

Did you deefine “new” terms (properties and classes?). For instance the
Agent Vocan could be described with the FOAF RDF VOcab terms?

We’re using existing vocab terms where they fit, including FOAF:Agent. And adding new ones as needed.

Clarification: We aren’t using foaf or other vocabularies yet, but we want to and we continue to research which would be best, based on the general principle Bob stated. It’s a matter of timing and how our development workflow goes. We’re coming up on something like version .1… or maybe .01… :smile:

(Bhaugen) #21

Oops, sorry about FOAF, @mariana and @fosterlynn.

(Mariana Curado Malta) #22

Hi there,
I am studying the pictures.
We need to descriptions of each term (class and Data property or object property) in order to find what term/vocabulary you will use.

I am a bit confused with what you call vocabulary. As far as I understand you are defining a Ontolology. The vocab Agent is in fact a class Agent and all relations and propertiea of that class of the ontology…as well as all other vocabs are definitions of classes of the Ontology.

Am I seeing that wrong?

It would be so much easier to discuss this in a conf call. Maybe we could arrange that?

Mariana