#OpenBadges #Taxonomies and Shopping Lists

Kerri Lemoie (@kayaelle) has taken the ambitious task to lead the Open Badge community in exploring further the field of taxonomies. I was not able to attend the last conference call, but I took some time to go through the Etherpad of the meeting and here are my latest thoughts on the matter.

In Over 2 Millions Types of #OpenBadges ! Don’t you think that’s wonderful? I explored the typology of Open Badges and the idea of a taxonomy to conclude the inanity of any attempt at enumerating the different types of Open Badges. In a later post, The Celestial Emporium of #OpenBadges Taxonomies I concluded that, considering that a taxonomy would need to be finite to have any practical value, it is very likely that such a taxonomy would provide an over-simplified representation of the world, an illusion of understanding — as if the mere fact of naming things increased our understanding.

After exploring critically the concept of taxonomy, in this post I’ll try to explore a more practical approach. After all, if people feel the need for taxonomies, it might be interesting to know what the actual needs are and what are the possible solutions to satisfy those needs.

Taxonomy, Typology or shopping list?

In the discussion on taxonomies, we need to take into account that there is a difference between a typology and a taxonomy:

The etymology of both words gives clues to their differences. In Greek, táxos means an order, onom- means name, so the word “taxonomy” means naming genus or species. “Typo-” means a type of organism and -logy means a study. Nelson Orringer · University of Connecticut (source)

Moreover, proper taxonomies must respect some basic principles:

While the fundamentum divisionis is a property of any classification in its entirety, the level of generality of classes is a property of each single class concept, as of any concept. Two concepts (X and Y) are at the same level of generality when (a) the extension of X (total number of concrete instances classed under X) is not part of the extension of Y, and vice versa; and (b) the extension of X is not part of the extension of a concept Z which is at the same level of generality of Y, and the symmetric can be said for Y.

While the violation of requisite (a) brings about a grossly incorrect classification, requisite (b) is violated for practical purposes in several actual classifications, it must be strictly respected only in taxonomies. [highlighted by me]

in Classification, Typology, Taxonomy by Alberto Marradi Universities of Bologna and Florence (Italy) source

This last point is particularly important when constructing or analysing a taxonomy. For example, the so-called taxonomy of Open Badges published by Ilona Buchem violates the property of fundamentum divisionis and, to use the terms of Alberto Marradi, “brings about a grossly incorrect classification.”

1 Content-related categories (what the badge represents)
1.1 Achievement badges (demonstration of achievements)
1.2 Competence badges (demonstration of knowledge, skills, competence).
1.3 Potential badges (indicators of future performance)
1.4 Participation badges (evidence of participation, e.g. events)
1.5 Membership badges (represents membership, e.g. club)
1.6 Commitment badges (attitudes, values, beliefs)
1.7 Encouragement badges (good work stamps)

[…] source

If mutual exclusiveness is a property of any couples of classes of a well-formed classification, it is clear that the classification above does not respect that elementary principle as the total number of concrete instances under 1.2 is under 1.1: to demonstrate competency (with a ‘y’ and not competence, with an ‘e’) one needs to demonstrate some kind of achievements, achievements that have to take place within certain events (hence another problem of mutual exclusion with 1.4).

All that can be said about the pseudo-taxonomy above is that it is at best some kind of inventory, not even a typology as there is no logos, just taxons. We are more likely to find respect of mutual exclusiveness in a shopping list…

While the distinction might sound pedantic, it is not to be ignored. For example, in A foundational badge system design, Carla Cassilli who initiated the reflection on taxonomies wrote:

The last two years or so have found me investigating and contemplating many different types of badge systems. Along the way I’ve been wrestling with considerations of badge types, potential taxonomies, and conceptual approaches;

Carla Cassilli, as distinguished Open Badges expert, did not commit the mistake of conflating the study of types (typology) with taxonomies and concepts. They are different instances of intellectual activity. We still have to learn a lot about Open Badges, in particular to get rid of some of the misconceptions that prevailed at its origin, before reifying them into the straightjacket of categories that, if produced today, might tell more about our ignorance than our knowledge.

What I would expect from the newly formed working group on Open Badges and taxonomies is to be accurate with the use of terms such as typology, classification and taxonomies, in particular to be able to identify and disregard any pseudo-typology, pseudo-classification or pseudo-taxonomy.

Taxonomies of or for Open badges?

When ‘Taxonomies’ and ‘Open Badges’ are used in the same sentence, it is important to know whether it is addressing the issue of taxonomies for Open Badges or taxonomies of Open Badges. For example, my first post mentioned above was clearly related to taxonomies of Open Badges, i.e. starting from categories such as competency badges and value badges. The second post was also clearly related to taxonomies of Open Badges while starting the exploration of taxonomies that could be used for Open Badges.

Having already addressed the issue of taxonomies of Open Badges in the section above, I will simply state, along with the participants to the first meeting of the working group, that there are already many classifications and taxonomies (again, all classifications are not taxonomies) that are being used by different sectors, domains and industries.

So, my questions are: what is the Open Badge problem to which another classification/taxonomy is the solution? Why cannot we use already existing ones? What could it bring that existing ones do not already provide? With its corollary: why existing ones are not able to provide what is now expected for Open Badges?

To those questions I would add: if one accepts the idea that there is a taxonomy issue, how does that issue compare with other Open Badge issues, in particular the flawed Open Badge Infrastructure, its asymmetry and its organisation centredness? Is there a chance that by addressing this issue we will create a more balanced infrastructure or will it maintain or even reinforce the current asymmetry?

Considering that the need for classifications/taxonomies generally comes from institutions, not individuals (who create folksonomies) it is not unreasonable to believe that there is a serious risk to reinforce even further the current asymmetry of the Open Badge Infrastructure.

What is the problem we are trying to solve?

From my understanding, one major problem with the current structure of Open Badges is the fact that every badge issuer stores its own definition of the criteria (on a server or within the badge). So, if two different issuers of the very same badge use the same criteria, the definition will be in two different places. Moreover, the definition is usually a simple text, not even HTML, so it is impossible to use techniques such as micro-format or RDF (the format use by the semantic web) to structure the information.

Is a new taxonomy the solution to this problem? Not really.

The benefits of connecting criteria to the real world!

The benefits of connecting criteria to the real world

The main problem we have today with making sense of Open Badges is the poor information contained in the criteria. While it was acceptable at a time of a proof of concept to use simple text fields, this is clearly not sufficient to fully exploit the value of Open Badges. What is needed is to externalise the definitions of the criteria outside of the Open Badge itself and the Open Badge realm altogether. Open Badges will be most valuable when looking for something else than an Open Badge, e.g. a concept, a value or a place this piece of information will be connected back to one or more Open Badges. What will this information be? We do not know and do not want to know. We just have to assume that it could be anything. How will this piece of information connected to Open Badges? It could be through semantic chains of through inference/search engines.

Of course it is more complicated for the person writing a criteria to use external references. A first iteration one could use simple HTML and link the main keywords through hypertext links to, let’s say Wikipedia or DBpedia definitions. A second iteration could be to enrich the HTML code with micro-formats or RDFa metadata. At this stage, Open Badges would still hold the definition of the criteria internally, but the hypertext links would connect them to the rest of the world, to definitions shared by others that issuers of Open Badges.

The next stage of development could be the creation of a public database of criteria that could be shared across badge issuers, earners and consumers. This time, the badge would not contain the full definition but a simple link to the definition stored in a public repository. Such a repository would be useful to reduce the duplication of badges and definitions.

And all this can be achieved without any new taxonomy…

On the other hand, what is needed is that badge issuing platforms provide good criteria editors (supporting RDFa or similar) and entrepreneurs willing to take over the organisation of criteria repositories that should be more than the current badge silos.

I look forward to further contributing to the work led by Kerri Lemoie.

2 thoughts on “#OpenBadges #Taxonomies and Shopping Lists

  1. Kerri Lemoie says:

    Thanks for sharing your thoughts. Some points:

    1. Using existing classifications/taxonomies – I’d suggest that we should make several extensions specific to existing taxonomies. The overlap in terminology can be handled by clearly defining and referencing each term.

    2. Taxonomies OF Open Badges – If we use existing taxonomies, the extensions are not FOR Open Badges but references for issuers, earners and consumers to use.

    3. There should be a textual criteria. Even if it is abbreviated with a referencing url. We could do this with an extension for now and maybe it bubbles up in the next release of the specification.

    4. Even if the Criteria is textual rather than a link as it is now, and that content can be analyzed to make usable data connections, taxonomies provide definitions that can be referenced using extensions and JSON-LD.

  2. Serge says:

    This is a possible solution, how about
    1. having one single extension and something like a series of XSLTs to translate the different formats into a unified one. To work with a new classification/taxonomy, just create a new XSLT
    2. I do not understand that point
    3. Agree with textual criteria, but many be not plain text, at least HTML so we can enrich the data with microformat, RDFa, JSON-LD, etc.
    4. JSON-LD could certainly be (part of) the solution. It would be for for the issuing platforms to provide Criteria editors using JSON-LD in the background.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.