In the process field, we call them “Event Streams”. These are streams of records indicating specific things that happened at specific times. In the Social Software world, they are called “Activity Streams”.
I was intrigued by a presentation by Martin Böhringer called “Using Enterprise Activity Streams to provide a Case Management solution” which is about emergent case management for ad-hoc processes. The idea is that micro-blogging offers a good deal of what you need to perform case management. The slide show spells this out by showing the activity stream from Twitter, Facebook, and Chatter and how they reference documents, other people, and context (cases). Another presentation shows the power of status sharing. My own experience has shown that sharing up to date status is a large part of enabling people to coordinate their work with each other — especially knowledge workers.
It seems that for two years there has been a large effort to provide a standard format for syndicating social activities around the web, known as “activitystrea.ms“. The site claims a large number of supporters, such as Facebook, MySpace, Windows Live, Google Buzz, BBC, Opera, TypePad, Gowalla, Gnip, Superfeedr, YIID, and many others. Yesterday Skype and Facebook announced an integration that allows you, among other things, to see and update your Facebook activity stream in your Skype console. Is that using the activity stream protocol? I was not able to verify, but clearly this shows the kind on potential we could expect if all such social systems implemented a consistent standard: you could exchange activity streams from anywhere to anywhere.
Where is this Going?
Paul Mathieson writes that this is “a paradigm shift where existing Business Process Management (BPM) methodologies and organisational structures are being enhanced by emerging social technology.” He has a point. There is a big change coming, which has been triggered by the social network systems. His post on Social Wiki to a Social Workflow System points to the need for additional flexibility that the more structure process systems fail to provide. Have we heard this theme of flexibility elsewhere ?
Adam Dean calls this a BPM Game Changer and he has a point. While BPM has worked for highly predetermined processes, Twitter can be use for highly unpredictable things that you never expected to happen. Today, his post on the new BPM Enterprise Twitters points to offerings from TIBCO and IBM with twitter-like features. My research of social systems also lead to the same sort of conclusion, particularly with Salesforce Chatter and Cisco’s Quad offerings.
Esteban Kolsky talks about something similar in the Long Term Future of Social CRM where he mentions “the Social Business evolution” and CRM becoming just part of the “Collaborative Enterprise”. We are seeing a big surge in enterprise social approaches and it is clear that this will eventually merge a lot of separate vertical collaboration spaces.
Given that Process Mining is a technique to recover the underlying process from a stream of event, it would seem that there is a perfect match with activity streams. The IEEE process mining group is attempting to define a standard event stream format (BPAF, XES). What is the difference? At the conceptual level, they are very similar. Activity streams however are not structured. As the presentation above points out, you can link to documents, link to people, and to context, but it is really hard to say what you really mean by those links. The event stream approached make it very clear what the context is, and what roles the other elements are playing. Yet, the lack of structure may be a large part of why social sites are so popular — and usable.
We should not yet summarily proclaim activity streams as “un-mineable”. Recent advanced in linguistic based search technique have show ability to find what you are looking for on the Internet from unstructured documents, and the same technique might be able to pull from the “tweet” just exactly what the person meant, and how that relates to the overall context of the case work. Farfetched? Yes, but not impossible.
The open question for me is where this is going in the future. Will event streams and activity streams become more similar over time. Will the 140 character limit remain sacrosanct? It is likely that we will see additional extensions to include other formalized data included, just as we have see the additional of geolocation. Don’t cringe, but one might call these “Enhanced Enterprise Activity Streams“.
The choice of name is a little unfortunate. The Activity Stream Wiki does define the terms they use:
- The activity in ActivityStreams is a description of an action that was performed (the verb) at some instant in time by someone or something (the actor) against some kind of person, place, or thing (the object). There may also be a target (like a photo album or wishlist) involved.
- The stream in ActivityStreams is a feed of related activities for a given person or social object.
Many people call these “events” because they occur at an instant of time. I think most people would say that an activity is something that extends over a period of time. In the process space we use activity to mean something that can be planned, anticipated, started, monitored, completed, and reviewed. In an activity stream those would all be separate “activities” if you will.
There seems to be plenty of momentum behind the term “Activity Stream” so I am not attempting to change that, just to make aware that this term means the same thing as an event stream. Add “activity” to the growing list of contentiously defined terms. 😦