Monday, November 12, 2007

Why David Recordon (Six Apart) needs an Ontology!

David Recordon and I had noticed a similar issue. At Web 2.0 we connected and talked about it. The issue is one of how in every social network you have to manually re-declare your whole social network. Rather than repeat it here, the post can be read here:

Short version:
Why do we have to keep re-declaring our social networks for every single social network application (Facebook, Mix2r, Twitter, Plaxo, MySpace, Dopplr,

David and I had lunch in the UK with Matt (CTO, Dopplr) and James Governor (RedMonk) and all agreed that this is an issue facing the next social network. David made a great presentation of it at Web 2.0 Expo Berlin.

We seem to agree that the solution is a non-proprietary open social graphing application that can be used to suck social networks into various social sites. If we are to make an open social network provider service, there are a lot of answers that have to be figured out first. Foremost, real FOL and ontology work has to be done. The social networks of today have limited and immature binary relationships. Here is an example:

"A is a friend of B"

What does this mean? Is it asymmetrical or symmetrical (is B also a friend of A?). Does B know A exists? Can B traverse the binary relationship (see that A declares B is a friend)? If so, does B know the exact nature of the declaration on the relationship? Do either A or B survive if the relationship is dead? Does the relationship exist if either A or B cease to exist?

There are simply far too many of these types of questions that are not really answered by existing social networks. This sort of ontology work is sadly needed. Perhaps a group like the Ontolog Forum can become involved to help sort some of this out.

Davd is a smart guy - I cannot wait to see what he comes up with!

Thoughts? Comments?

BTW - I am on vacation for two weeks and will not respond after tonight CET.


  1. It's far too simplistic in my view to think that an ontology of social networking concepts would have allowed, or now allow or accelerate interoperation of the different social networking applications and services currently available.

    One can simply see that the current OpenSocial effort was introduced after the apparent success of these services and that there is already a variety of social networking Web applications and services with a multitude of focuses. For instance, LinkedIn focuses on business relationships whereas MySpace is focused on younger kids.

    This heterogeneity of services and cluster of focus implies that your example of "A is friend of B" has different meaning and implication in each platform while appearing to use the common concept of friend. I would agree that at a high-level a friend in Facebook, MySpace, or LinkedIn have some similarities, however, that does not translate into what is the core value of these services. That is, using such a high-level relationship would not help the site implement their core value propositions while enabling low value of integration.

    The problem is exacerbated by the fact that each of these services (e.g., MySpace, LinkedIn, Facebook, and so on) are also competitors for consumers and advertisement dollars and want to enter each others business without explicitly saying so.

    I do believe there is a time for trying to enable interoperation, but that happens at a later time, like now. Notice that all of the companies involved have their own platforms and services focused on some niche market. Since there seems to be a potential winner that is moving away from the pack, Facebook as of 2007, the rest of the players are reacting to try to consolidate via API and data models---and I would agree that part of this data model could be seen as an ontology. However, it is interesting to note that the effort is lead by the player who has the most to gain but has less of a footprint in the space...

    Heterogeneity is necessary for advancement of society, human activities, and business. We do not all think alike and creating a priory agreements (such as what is required by ontologies) is unrealistic early on and rarely happens. When they do, they typically are incoherent, take huge amount of time, and are second rate. With time they get better but not before being worse and the likelihood of success is also small.

    I believe the grassroots, Darwinian, and open-source model, that the Web 2.0 follows is the best prospect for a programmable Web and that semantic Web notions while interesting academically will never put a serious dent in the mix...

    Various early ideas and implementations are necessary for advancement and progress. This also matches well our free society while forcing a priory agreement never helps create markets, has no value for the early leaders, and take huge efforts with low-level of return on investment. The realities of society and the market are simply working against the need or value of ontologies plus the technical merits are also relatively low.

    See my short paper The Programmable Web: Agile, Social, and Grassroots for a more elaborated argument.

  2. Max:

    Great comments back! Thanks for taking the time. At the very least, ontologists could help by at least providing the FOL of (entity) and (relationship). Each can have distinct attributes that are optional based on the level of complexity of the social network. As you have inferred, the different social networks have misaligned levels of capabilities and it would be far too simplistic to assume they would all use the same taxonomy.

  3. I think there is an interesting proposition in building a framework for analytics on friend relationships. As you mention, there are different meanings of "friend" based on context. An ontology is a modelling mechanism that can enable the contextual definitions of "friend". However, perhaps this can also be represented in the Open standard.

    More importantly, as you point out, is the human motivation for standardization. Google is using it's name to support "open source" in an area where it has fallen far behind (I mean, who uses Orkut?)

    I would love the ability to define my own hierarchies of online "friendship". Perhaps Facebook will lead the way in defining more granular relationships. They seem to be on their way (in that, you qualify your relationship - "I went to school with this person" , "I worked with this person". ) If they represent these relationships in OWL - then we could write rules on the strength of relationships which could help us manage our own interactions as our networks inevitably get bigger. (e.g. roles engineering on broadcasting messages)

  4. This comment has been removed by a blog administrator.

  5. The above comment was removed as it was a filthy spammer trying to sell you something.

  6. This comment has been removed by the author.

  7. Pace and direction is what is dissembling the process of aligning relationship in social networks.
    If we look at the initiative taken by OpenID and how it is starting to be embraced as a loginID. There is the talk about decentralisation of twitter and how this can be done. In my opinion what is needed is a consortium and an agreement amongst the major players facebook , myspace, etc to feed in the profiles.
    A way to get the ball rolling is a firefox addon like roboform that gets filled, spreads the details for various social networks and the autofills. I agree this is not the best solution, but if the user opt-ins for it to be stored on a central location then maybe the social networks just need to add a button that will auto populate information

  8. This comment has been removed by a blog administrator.


Do not spam this blog! Google and Yahoo DO NOT follow comment links for SEO. If you post an unrelated link advertising a company or service, you will be reported immediately for spam and your link deleted within 30 minutes. If you want to sponsor a post, please let us know by reaching out to duane dot nickull at gmail dot com.