Announcement of Thingk.com private beta

Posted on January 18th, 2010 in News,ThingkBlog by admin  Tagged , , , , , ,

The Thingk.com private beta will open up in May 2010.  Here are some details surrounding the private beta:

  • This will be an invitation only beta, which means:
    • Some colleagues and friends will be asked directly to join – please feel no obligation but your feedback will be greatly appreciated if you can assist.
    • Throughout the last few years, I have been looking at interesting individuals who may be interested in patternsmithing and Thingk persistence – you will be receiving a hand-crafted email as a personal invitation – once again, no obligation please – only do this if it brings you value but if you do your feedback is greatly appreciated.
    • As expressed on this blog, if interested feel free to contact me at the contact form http://xri.net/=joel.kotarski and I will be able to accomodate you.  Also if you think you know someone who would be interested from the category above, please contact me here as well.
    • We’ll probably have limited invitations to spread around among private beta users – these will probably only be on a request-only basis, however, to control capacity initially.
  • For those in town, we will most likely host an in-person meeting where we will go over some details about the site and what is being accomplished – for out-of-towners we may also host a teleconference as a separate event.
  • Since we will be testing privacy and security (with potential bugs to be found), I ask that you put the kind of ideas you may spontaneously work on but not your actual work and certainly not anything you would consider intellectual property at this point.  Much of the data from the private beta will be wiped clean on launch so that we can fully test; however, I am planning to make an easy mechanism to tag content you wish to transition over to the full system – also, I will be soliciting feedback on making this process as painless as possible for those generally interested in getting value from the work they do in beta testing.
  • As always, being asked to participate should not imply obligation to participate, and participating should not imply obligation to give detailed feedback.  Anything you do is greatly appreciated and will be rewarded in some manner.

“Polymeric” user interface

Posted on January 2nd, 2010 in News,ThingkBlog by admin  Tagged , , ,

The user interface will be central to the success of the Thingk.com site.  We’ve gone through several iterations of user interface in an attempt to simplify how our users will persist their Thingks.  After several iterations, I believe we have reached a concept that will achieve all the goals we set out to accomplish – I am codenaming it “Polymeric” – a nod to combinatorial chemistry (more on this later).  This label may or may not make it beyond the source code and this blog/design meetings, but it usefully describes what is going on: I  want the user interface to not only assist those who have a Thingk ready to persist without getting in their way, but also those who are searching for a solution in a combinatorial manner (where the site becomes a source of information, components, and inspiration as well as a persistence mechanism).

Polymeric finally admits that the user interface should get out of the way of the user – that the best expression comes out of the user if they are guided and prompted along but in a non-intrusive way.  At the core of Polymeric is the ContextProvider – the context is used to provide suggested components, prompted questions, and directions for exploration.  The ContextProvider works at three levels : what the user has done before (what Thin(g)(k)s* they have created, what Thin(g)(k)s have inspired them, as well as the components of each), what they are doing right now (the current word they may be trying to type, what question they are answering), and where the engine can infer they are going (via pattern matching of current text we can infer related directions for exploration).

*: Thin(g)(k)s means Things, Thinks, and/or Thingks (or any combination thereof).

At the center is the Polyermic edit box – a text editor (with optional rich editing functionality but this is hidden away and only appears under certain circumstances) which is fed by and feeds the ContextProvider.  At the moment, we are planning to have users use WikiWord format to denote Thin(g)(k)s across the entire site – it is just a format that works well.  The ContextProvider loads several layers of information, but the first layer of information is a static list of WikiWords (that is, pointers to Thin(g)(k)s) which the user has had contact with (either by creating, being inspired by, or using as components in their other creations), primed for usage in this context.  Let’s look at the post I am writing currently on this blog – four WikiWords are at play and only three are Thin(g)(k)s.  The Thin(g)(k) WikiWords are : “ContextProvider”, “Polymeric”, and “Thin(g)(k)”.  I’ve repeatedly typed all three manually throughout this post, but if I was editing this post in Polymeric in the actual Thingk.com web site, these words would have been suggested to me.  This due to the first and second layers of information: “Polymeric” and “Thin(g)(k)” would have been loaded in context in the first layer (as my creations and/or creations I collaborated with others on), but the moment “Polymeric” was suggested to me and I picked it, the ContextProvider would have loaded all the components of the Thingk “Polyermic” into context, including (but not limited to) the “ContextProvider”.  And one of the other Thin(g)(k)s that would have been loaded is a reference to a project which inspired context suggestions and must be payed homage to: GoogleSuggest.  GoogleSuggest inspired many people as an autocomplete mechanism to use Ajax to give suggestions based on user inputs, but this is in the context of one singular input, whereas we will suggest individual parts of a larger body of text – this is of course nothing revolutionary – you can see the same feature on your cell phone.  The difference here is that a Thingk, behind the scenes, will be a RDF resource that follows the ontology defined for Thin(g)(k)s at the PatternSmithing Alliance, including linkages between Thin(g)(k)s at the RDF level.  Like most semantic web applications, we would never dream of having our users manually hand-type RDF; however, by choosing one of these WikiWords, the text of the Thin(g)(k) is not the only persisted but also the relationship/link to the Thin(g)(k) itself.  This was the crucial problem to solve – to give the user an intuitive mechanism for forging these relationships.

The Polymeric user interface has an additional set of features which are fed by the ContextProvider.  Several of these are based off the usage of WikiWords – for instance, it is possible to use a WikiWord for a Thin(g)(k) not currently on the site (for instance, my mention of GoogleSuggest would have started the process of creating an authorative reference on the site to Google’s project, pending Google’s permission to create such a thing) as well as suggestions to the user to turn a repeated pattern of text into a WikiWord and therefore a Thin(g)(k) on the site.  Additionally, the outer borders of the user interface will have bits of context floating in based on what is going on inside the editing area: a cloud of moving WikiWords that can be grabbed and dragged onto the editing surface that changes over time, and prompting questions that fade in and fade out  to guide the user to starting or evolving their work.

Once we open up the private beta, we will be soliciting feedback from some of you, so please contact me at http://xri.net/=joel.kotarski if interested.

  • Comments Off