What are the basic elements of Enterprise Collaboration?

A friend asked the question as she started the process of exploring her requirements for collaboration products, which she will use to support a social networking initiative she is about to get going.

Like many other practitioners before her, Lisa found that when it comes to enterprise collaboration, there is a huge difference between wanting to solve a problem and knowing what specific product features are relevant to your needs. In other words, if you want, say, “to empower collaboration in order to create alignment between highly distributed teams in order to improve the product cycle”, how does MySite functionality help you? If you want “to increase intimacy between partners and internal stakeholders”, is that something a blog, a wiki or a forum will produce? How relevant is a forms server to collaboration?

The thing is, it’s common to start requirement discussions these days and immediately start hearing from stakeholders “I want a blog/wiki/community/forum/tweeter/your-Web-2.0-buzzword-of-choice-here”;  everybody wants to jump into it, whatever that is, because the benefits of Web 2.0 technology are presumably very  compelling. The question is, are those benefits the ones you are after? Is your organization ready? And so on…

It is very rare, however, to get an articulated statement of need that goes from specific use cases to correlating common collaborative patterns inside the organization with the functionality needed to empower it and enable it. Add to that the pressure from vendors to define the field (which is in the middle of a land-grab period)

What Lisa was asking was some sort of basic semantics and ontology to start moving away fro the “wants” towards the “we need… because…”. Basic thought blocks… A basic functional taxonomy.

Of course, there is a ton of pre-digested stuff that I could (and will) point her to, matrices after matrices of basic function points, usually provided or ’seeded’ by vendors in the category, talking about portal management, forms, business intelligence, types of content supported, and so on. I am sure we will get to those at some point, but first we need to agree on basic “thought blocks” for us to manipulate and discuss as we approach her requirements.

On the personal side, I find myself going over these basic elements over and over, because this is a very common stage of collaboration projects, so I decided to put some extra work at it and construct a basic  taxonomy to use as a starting point for future requirements analysis with my customers (hopefully, everybody wins). I will improve the taxonomy after I write this article, and use it as a skeleton to tag and classify content in this site as it grows

The basic framework

There are many ambiguities in the definition of collaboration-related terms, thanks in great measure to the efforts of vendors to re-define the world in terms that fit their products (as opposed to the other way around), as well as the broad extent to which the term is applied (e.g., intellectual collaboration between several individuals and collaboration between countries). Nonetheless, a common-sense approach to understanding collaboration requirements makes some of those ambiguities irrelevant by

1. looking at how people in a given company have traditionally collaborated, 
2. identifying recent changes that affect collaboration  
3. extrapolating the above to understand what is needed to support and empower collaboration today

To avoid distractions produced by huge hype campaigns, over-anxious vendors and the infinite wave of convergence that produce one presumably revolutionary device every week, each one enabling equally presumed revolutionary collaboration capabilities, let’s start with traditional collaboration concepts (1960s to 1990s) whose validity is not bound to recent technologies. Once we agree on some basic building blocks for the concept, let’s keep building on it until we reach a satisfactory “present” status, and only then let’s speculate on what tomorrow will bring. 

The basic concept of collaboration

As for the word “collaboration” itself, let’s define it for now as an iterative process followed by multiple individuals,  all of them sharing a common goal or mission, in order to:

• Share their private knowledge and access others’, which is usually done by sharing and interacting with content (i.e., mediated knowledge, knowledge represented as content) as well as by  communications between individuals conducted via conversations, messages, collaborative editions of documents, etc. (i.e., conversational knowledge);
• Explore and discover possible scenarios and alternatives related to the common goal;
• Develop consensus about reality and courses of action (usually making that consensus permanent and more shareable by creating new content).

The definition above is quite operational, and is particularly effective inside corporate environments because it is generic enough to allow for multiple “flavors” of collaboration, from the more structured, control-driven ones (e.g., performance analysis and review sessions) to the more decentralized, socially centered and conversation-focused ones (e.g., strategy brainstorming).

If we were to simmer that definition even more, we should arrive to the fact that people collaborate by interacting and working jointly on knowledge, both conversational and mediated.

Simplistic collaboration patterns

Basic collaboration patterns emerge as soon as you put a group of knowledge workers in the same location (”working jointly” used to require being in the same place), make sure they use some set of basic artifacts that most of them are used to, give them intersecting objectives (that is, the success of one influences the success of others), and give them time to develop their own routines.

The individuals will soon reach a balance between individual (private) work and collective work, the later conducted by meeting somewhere, bringing with them the content and other artifacts they need for the task at hand, and collaborating (sharing knowledge, exploring possibilities together, developing consensus). Given choices, they will like to maintain close relationships between them, so that as new elements come up during projects, tasks, workflows and such, they can address them through ad-hoc, further collaboration. 

Finally, as they become more familiar with each other, individuals will develop and/or further multiple types of relationships amongst each other giving place to a myriad of specialized networks and sub-networks (if you like graphs, think n-tuple meshes). As the number of those networks grows, individuals will represent them (besides explicit lists), by attaching characteristics to individuals “profiles“. In layman’s terms, the group of Community Sponsors in your company may be represented by a list of the thirty or so individuals in it, as well as a variable in each user profile called Community role, which for those 30 individuals will have the value of “Sponsor”. 

In that simple context, collaboration happens as those workers start all sorts of interactions and collaborative activities supported by work artifacts (schedules, calendars, tasks, workflows, processes) and in the process consume and create many types of content (messages, recordings, presentations, documents, etc.). When they are not meeting, those workers will be working (alone, or in different meetings, on the same artifacts and content).

Personal networks will operate in a somehow “orthogonal” yet crucial plane relative to the other elements, by acting functionally as a selector (which networks participate in a particular collaboration), but more important yet, acting structurally to represent corporate groups participating in crucial corporate processes that are enhanced by collaboration. The  process that matches both functional and structural networks is crucial to the ROI for social collaboration.

The simple collaboration model described so far works only in very constrained circumstances, in which:

• workers are co-located (or near-located). That is, it is simple to move between the private workspace (where they work by themselves) and collective workspaces (where they work with others);
• they share a common contextual understanding of the task at hand;
• the participating individuals have relaxed agendas, so that most events can be performed in a synchronous manner (i.e., all participant are available at the same time for mutual interaction – a phone call, a meeting, a conference), as opposed to asynchronously (i.e., participants conduct their own activities at different moments in time – store-and-forward activities such as email, voice mail or any other type of messaging, editions of a document that are performed with persistent locking of versions, etc.).

Some of the key concepts identified above are directly correlated to ECP’s functional layers:

1. Individuals and groups (which will be usually translated to “users” of the collaboration platform and “networks”)
2. Artifacts that are used by all individuals, such as calendars, process diagrams and workflows, location and presence indicators, identifiers, etc.
3. Content, which are documents and other file-based representations of knowledge used by individuals
4. Channels of communication between individuals and groups
5. Interactions between individuals and their common and private knowledge.

More realistic patterns

Enterprise workers today work in conditions that rarely include collocation, social integration and synchronous collaboration; rather, enterprise workers today…

• Are hyper-connected to data networks (more frequently, to more networks, with higher bandwidth) and they use it to generate massive amounts of content;
• Have lots of computing power to spend for assistive technologies (like most of those in Smart Phones);
• Are (increasingly) used to technology-mediated interaction technologies, as well as devices and modalities, and have relatively low resistance to behavioral changes required to incorporate them;
• Multi-task intensely, and are usually involved in a multitude of projects, each one in connection with different groups, objectives and processes;
• Find themselvesy distributed across geographies, time zones, even languages;
• Are NOT socially integrated, but are instead increasingly process-integrated by back-office corporate applications (this is equivalent to saying that knowledge workers now interact not only with other knowledge workers, but also with applications and automated processes);
• Use an increasing number of devices, as different combinations of connectivity, memory, processing power, storage and interaction bandwidth are brought to market in convergent devices;
• Generationally, growing number of knowledge workers are increasingly subject to “Web 2.0 expectations”, which project current experiences with convergent devices, mobility and other interaction technologies to predispose them to try technology-assisted collaboration (and accelerate adoption in the process).

These circumstances represent both challenges and enablers to collaboration, and have been at the root of many industry movements to address them (knowledge management, business process integration, content management, document management and more). ECPs are therefore not radical, new technologies, but rather a confluence of older technologies and infrastructural conditions that were not present before, and which come to be known as Web 2.0.

Challenges to collaboration

• Geographic distribution and thin time slices mean that synchronous activities require a lot of assistance from technology in order to still take place, this time in virtual workspaces, whether it be a web-based conference or a presence-enabled call that is automatically routed;
• Personal availability for synchronous activities is reduced by orders of magnitude, and scheduling becomes proportionally harder. As a result, multiple modalities of asynchronous knowledge interaction emerge to replace synchronous ones (each of them creating its own management and integration challenges) and presence management becomes critical for communications;
• Group maintenance activities become overwhelming when the number of working groups multiplies; even keeping in mind the context of each group becomes almost impossible. As a result, the administrative functions related to collaboration grow more and more complex;
• Collaboration becomes yet another activity mediated by applications (collaboration platforms), creating resistance (”I already don’t have enough time, don’t ask me to do something else”) and confusion (too many interfaces between systems and individuals); extreme ease of use and seamless functional integration will become crucial to avoid failed deployments;
• Teams are rarely socially integrated, which means that each project has the potential for participants needing to re-develop basic trust and capabilities understanding. As a result, assistive technologies for relationship management become more important (still a very immature science, reduced to basic reputation management and relationship initiation support);
• The massive scale of most global enterprises, the content they create, and the number of individuals they involve requires innovative discovery, sharing and consumption mechanisms for knowledge and individuals.

Enablers of collaboration

• Ubiquitous networks make location more irrelevant, facilitate most virtual meeting interactions when combined with presence and other assistive technologies  (synchronous advantage)
• Convergent, and more intelligent devices make it easy for workers to utilize time slots that would otherwise be wasted, interacting in relatively new but highly productive modalities;
• Lots of extra computing cycles can be “wasted” in assistive technologies, such as voice, natural language, semantic layers, rich presence and location, and so on;
• Many routine activities related to tasks, auditing, record-keeping, and more can now be performed by systems without human intervention;
• Content integration brought about by Web 2.0 makes it easier and easier for workers to collaboratively create content and weave into it personal interactions of all sorts, reducing process friction and at the same time empowering richer, more complex documents (but at the same time contributing to the “content avalanche”).

In a nutshell

When we factor these new constraints and enablers into our definition of collaboration, out definition of what a collaboration platform needs to support do not significantly change. Collaboration platforms will still need to manage users, groups, artifacts, content, communication channels and knowledge interactions; however, in order to address the challenges above, collaboration platforms will need to:  

• deliver many different types of virtual workspaces on demand and in an ad-hoc manner, 
• facilitate the discovery, sharing and utilization of knowledge resources, 
• enable streamlined and unobtrusive communications via multiple channels and modalities and 
• support collaborative activities across all workspace activities, which requires making collaborative services available to back-office applications as well. 

However, collaboration platforms need to support those elements in a manner much more agile than traditionally done,

• leveraging technology and networking to replace physical proximity and laxer time availability, 
• creating assistive technologies that improve management of groups and networks (i.e., social networking),
• automating administrative tasks via shared calendars, project management, tasks, etc., and 
• addressing the issues created by the massive quantities of individuals, knowledge and interactions involved.

Factoring patterns into meaningful functional blocks

The challenges and opportunities above have created the conditions for solutions that enable modern knowledge workers to collaborate productively, which we will call Enterprise Collaboration Platforms (or ECPs for simplicity).

From a functional point of view, ECPs consist of several crucial components, which together deliver the experience of collaboration as introduced in the beginning of this article, and which can be described as conforming a layered functional pyramid:

Core Services

Core services unify and standardize (to different extents) knowledge workers’ access to all the core application behaviors, metaphors,  involved in creating the user experience. In terms of the succinct definition of collaboration that we developed in the beginning of this article, core services do most of the “heavy lifting” needed to support users (i.e., knowledge workers who can access the ECP), artifacts (by containing and/or integrating to the servers that allow those artifacts to be discovered, created, shared, modified and managed) and groups (i.e., multiple typed relationships between any number of users):

• User Management – All services that allow knowledge workers to access and use the ECP, including user authentication, administration, user directories, Single Sign On, etc.;
• Group Management – The creation, modification and maintenance of groups of any number of users, related by a common typed relationship (e.g., “is a participant in Project X” or “all content moderators outside of the company”, and such), and surfacing of those relationships via rich user profiles and attributes;
• Artifact Infrastructure – Servers (or interfaces to servers) that enable users to create, instantiate and use artifacts such as calendars, schedules, workflows, structured data (databases), phone calls, instant messages, and other objects of collaboration that are not directly correlated to files.

End users do not typically “see” or interact with the core services layer, but those services permeate the complete platform, and enable all users to access and use the same artifacts and services (user differentiation based on function access is implemented at a higher functional layer) as they collaborate. 

As you advance towards your requirements document, you will find that this functional block contains most of the IS- and IT-oriented use cases and requirements. Depending on your internal power structure, you may see systems that exceed the requirements in every other functional category being “shot down” because its core services don’t align well with IT’s environmental vision. 

“Rich” Relationship Management

Collaborative relationships are characterized by very large number of possible types of relationships, potentially as many as there are nuances that are important for people to relate to other people. From “people who like bicycles” to “taoists’ to “commuters” to “married” people to… the list is obviously close to never-ending: users of a collaborative space are connected by a very large of “typed” relationships (it’s not just a single type of relationship). If you were to represent it as a graph on paper, you would probably end with nodes (points, little circles) connected by a massive number of lines of close to infinite lines, of almost as many different kinds (e.g., colors).  At this point you would be approaching a “mesh”, a complex graph of very specialized properties.

There is line I don’t want to cross in this article, that of discussing architecture (I am not qualified), but the fact is that the traditional representation of using records representing users, with fields for different types of relationships (e.g., “Likes bicycles”) AS THE ONLY WAY to represent social relationships very quickly breaks down in implementation, as the records in question become “wider” (i.e., the number of relationships increases).

We will visit the specific requirements that can separate simplistic relationship management from rich relationship management in a separate article. For now, let’s state that there is a fundamental difference between true relationship management and the simplistic association of a few relational attributes to users.

What about Document Management?

Depending on which vendor you ask, the core of EC will contain more or less support for desktop applications and the documents they produce, in the form of Document Management; Microsoft, for example, with a vested interest in making MS Office even more of a standard than it already is, and pushing other technologies into the enterprise into the process (both valid competitive aspirations), has a heavy bias towards tight integration of MS Office applications into the Enterprise Collaboration platform (SharePoint WSS and MOSS), and thus has made document management the core of their EC product, because given the dominance of Office, that really means Office-document Management.

But such close relationship between desktop documents, eminently private in their creation, storage and management, and Enterprise Collaboration need not be the case. In “pure” Web 2.0 collaboration scenarios, all documents can be online documents with shared access (a wiki page, a posting, or a whole wiki), and full-blown collaboration can be reached without the user ever touching a copy of Word or Excel, or if she does, leaving the detailed document management outside of the scope of the collaboration product through common attachment mechanisms.

Which side you sit on, it’s a matter of opinion and religion. My recommendation to practitioners is to consider document management as complementary, but external to the collaboration system. I do that for two reasons:

1. Manageability: Because of the need to guarantee availability, resource consumption, application provisioning, compliance problems, and a million other reasons, desktop documents are much more of a management problem than online documents, with little added benefit. Further, collectively created content is usually much richer in context and knowledge than ultra-formatted private documents that contain only one brain’s output. Call me an optimist, but I believe that IS and LOB managers will realize these facts and increasingly move to online documents, as opposed to desktop ones;
2. Vendor tie-in:  I always recommend to customers to reduce vendor tie-in, not to increase it (which is a particular case of a higher end rule, that of not sacrificing independence of opinion and choice unless there is an absolutely life-or-death reason to). As a consequence, I always recommend on the side of keeping Document Management outside of the realm of EC (and proprietary document formats as well), building seamless conduits for documents to make it in and out of the collaboration platform, and avoiding private desktop documents as much as possible. That means that yes, you will occasionally need to import into your EC environment a few files that are part of special processes (e.g., mechanical design) or that are used as interface between your organization and others (the usual spreadsheet or slide deck), but most other documents should be kept as richer, shareable, continuously updated web content.

The three pillars of collaborative workspaces

The three columns introduced below each take care of three key collaboration objects: knowledge, communications and content.

Knowledge Management

This group of functions allow individuals and groups to represent, access, integrate, share, publish and consume knowledge, by leveraging different models of interaction and organization, whether it be wikis, blog, forums, micro-blogging, shared-note-taking, application-sharing, white-boarding, co-creation and co-edition, or any combination thereof, whether executed synchronously or asynchronously.

Knowledge is increasingly represented by meshes of objects (nodes and users) and large numbers of typed relationships between them; to accommodate this representation, this layer also includes knowledge management artifacts such as taxonomies, semantic tagging, synonym and glossary management, topic maps and ontologies.

Also critical to knowledge management is the representation of individuals’ trust and expertise, as a critical qualifier of knowledge; therefore, the knowledge management layer also includes (or in some cases should include) assistive technologies such as reputation management, network mapping and analysis automated relationship management for users. 

Typical functionality delivered by the knowledge management column include:

• Content interaction models – Asynchronous models reign today (Wikis, blogs, forums, chat, twitting, mobile note-taking), but synchronous modalities are also being explored with interest (e.g., simultaneous co-editing, white-boarding, others). Regardless of the results, the list will probably keep growing, as different interaction modalities are explored, and standards for the core models will emerge;
• Knowledge representation and discovery services – Semantic tagging, tag clouds, smart tags, topic maps, ontologies and other abstraction technologies  that improve on the limitations of flat text and links by automatically (or collectively) creating metadata about content, and thus facilitating its management;
• Expertise location – Assistive technologies that assist in finding not only the right piece of knowledge, but rather the right resources that may be used (i.e., instead of finding an article on item XYZ, find the person who writes the highest rated articles on XYZ and semantically related concepts).

Communications Services

Support for fluid communications between knowledge workers and/or themselves and/or applications, utilizing transport-based channels operating in both synchronous (VoIP, chat, conferencing, IM, etc.) and asynchronous (email, SMS, RSS feeds, streaming, faxing, multicasting, etc.) modes. 

In the previous paragraph, the inclusion of applications as participants in communications is not accidental, but rather crucial, as more and more components of corporate knowledge is managed directly by automated systems: the capability to integrate those information channels into the collaboration system is a crucial capability for collaboration (e.g., your group may want to make sure that data from your SAP or Oracle applications can seamlessly integrate into collaboration processes via web services, and that setting up such flows is as simple as possible).

Despite the fact that today’s commanding communication modalities are text-based, a quickly growing number of technology-assisted interaction modalities and channels will almost certainly change that, balancing text-based interactions with other types, most of them mimicking person-to-person interactions, but some of them originals in their own ways (e.g., twitting). Each modality in turn carries with it artifacts of its own (filters, feed subscriptions, buddy lists, etc.), which can make channel proliferation more of a problem than a solution.

Functionality that falls in the communications layer includes:

• Synchronous channels: Chat, desktop conferencing (voice, video and converged);
• Asynchronous channels: Email, RSS, notification services, unified messaging, etc;
• Web services: Capabilities required for integration of application data into collaboration processes;
  
• Contextual communication artifacts: Ability to initiate multiple possible channels of communications with specific individuals wherever mentions of those individuals is made (e.g., contextual menus in Outlook wherever a recognized user smart tag is displayed, including calendar info, telephone numbers, presence information, email addresses and such, all of them “live”). Also, unification of multiple communication modalities under “smart client” applications, in order to reduce the complexity introduced by too many communication channels to master and control;
• Other artifacts and services: Additional services that are needed in order to maximize the contribution of communications to collective knowledge, such as recording, playback and podcasting, participation and delivery tracking, filing, conversion and indexing, subscription and filtering, channel convergence and switching, and much more.

Since communication is so crucial to creating knowledge, it would be expected for a corresponding level of support by Enterprise Collaboration products and platforms. Unfortunately, that is not the case: even for mature channels such as email artifact management is very poor and spotty. That creates several adoption and implementation problems, as users have to learn and internalize different gestures and usage patterns for different channels, as if they all were different processes, when in reality they are ultimately all producing the same transfer of information across space and time. That is probably the reason why, for example, a very valuable communication mechanisms such as NetMeeting was part of Windows for years with minimal usage and impact (now in the process of being removed and replaced by a more generalized set of services).

There seems to be increasing awareness about this problem in the part of vendors, as reflected by the hype around unified communications, unified messaging and such, as well as by early awareness about possible “interaction platforms”. Unfortunately, vendor interest is almost certainly more related to gaining a seat in the banquet devoted to devouring the spoils of analog telephony (a.k.a. ‘unified communications’, UC2, etc.) than produced by an interest to advance the state of the art in communication standardization.

Content Management

This functional block supports all activities and workflows related to content discovery, access, creation, storage, sharing, modification, distribution, archival, publishing, and consumption. I use the term content with its generic meaning of “contained information”, making no explicit distinction between on-line or desktop containment, structured and unstructured, single- or multiple-file. Such wide definition follows my approach to this taxonomy: when people collaborate outside of the Enterprise Collaboration platform, they may use a statistical table, a picture, a beautifully formatted document, or a piece of music to collaborate. The same should be possible in enterprise collaboration.

As with communications channels, the coverage of different types of content by different content management functions is quite uneven: while 70% to 90% of corporate content is said to be unstructured, only structured data stored in database enjoys wide coverage. But the movement to standardized levels of management regardless of content type (i.e., towards the “content platform”) is being fueled by more and more stringent compliance requirements, and hopefully will become reality sooner than the “interaction platform”.

While waiting for the emergence of a unified content platform, there is one distinction between types of content that remains highly disruptive for collaboration platforms, that which separates desktop-stored content (from office documents to sound-casts, video, multimedia, collaboratively created content, CBT, and more) – which reside usually in individual computers – and online content (e.g., wiki pages, blog posts) that reside on servers. The characteristics of both types of content are very specific: while desktop-stored content is private, application-specific (and in many cases proprietary as a result), of linear flow and using fairly elaborate formatting, online content is shared, somehow not-application dependent (through the use of common tagging languages such as textile and its derivatives), possibly enriched by strong hyperlinking and non-linear flow, and simpler formatting. The distinction is quite important to the product selection process because Microsoft collaboration product, SharePoint, puts special emphasis on Office documents as the unit of content, whereas other collaboration vendors such as Atlassian, JIVE, and others excel at managing sophisticated and rich online content (with IBM somehow in the middle). 

The distinction, however, will grow increasingly blurry, as desktop alternatives emerge that deliver the best of both types and are identical across the desktop-server divide (See “Evernote: New collaboration modality?“), without the need to preserve proprietary differentiations. 

Column Dynamics

Convergence

We have said that the core of collaboration is people sharing, accessing and interacting with each other, as well as with each others’ knowledge; in real life, those activities are closely intertwined, and in some cases indistinguishable. For the same reason, the lines between the three functional blocks above sre not always clearly delineated: the three categories keep weaving into each other and converging into each other. Because of that convergence, some services, tools and infrastructure components participating in ECPs are hard to classify, and will become increasingly so. Particular examples are conferencing (collaboration and communications), presence services (communications and collaboration), unified messaging (Content and communications), etc.

Channel proliferation

As channels and collaboration modalities keep proliferating (and in most cases entering the enterprise via sneaker-ware – consumer adoption by its knowledge workers) collaboration platform vendors are faced with the need to integrate management of the new channel or modality into their products or leaving it outside of the collaboration process. This makes the planning process increasingly difficult; in general, the best response is to avoid single-vendor lock-in and favor standards-based approaches (easier written than done, unfortunately).

Context Services

Context Services is the functional block most visible to users, and a visible differentiation between enterprise collaboration products; this layer creates maintains for users a workspace metaphor that anchors those users’ experiences into a certain context, or perspective. Those contexts are equivalent to views into a database: many virtual views can be implemented based on the same core data, all of them different. That is particularly the case with EC products (which are, after all, database applications), because the context or metaphor used for collaboration is highly influential upon the type of collaboration that will result (if any).

Workspaces are the experience of virtual spaces presented by ECP’s to users, as a way of providing a physical metaphor that can both make the collaboration experience more intuitive and at the same time anchor other services (and specially the artifacts that represent them) in a common metaphor. For example, in social-centric workspaces usually favor views of the system centered about the user looking at them, her previous activities and her choices for the things she wants to be presented with. In document-centric workspaces, preferred views are those of sites where lists of documents, users and/or artifacts are presented, and collaborative-content-centric views favor rich views of content available for access. Finally, future “workspace 3.0″ work currently going on (the term is mine, and arbitrary) is creating virtual reality workspaces where the commanding metaphor is physical, and where artifacts somehow try to leverage physical metaphors as well: the artifact for a phone connection in such a workspace is the 3D representation of a phone. These “views”, just like database views, are totally independent and arbitrary, and not usually exclusive.

For example, if the main “view” (or context) for a given system is implemented as activity- and project-based workspaces, with strong ties to task management, calendar, GANTT charts, and milestones, that project-centric view will influence strongly how users leverage the system, and will to a great extent determine the user interface as well, which will become technical, centralized, hierarchical, with lots of controls (many attributes are important in project management) and unforgiving as it relates to lack of precision. A conversation-centric workspace, on the other hand, will clusters its functionality around relationships, friends, conversations, discussions, issue discussion and consensus, and such. The user interface may become much simpler, and the view for the user will be almost certainly self-centered even when still concerned with her activities, contacts, on-going tasks and so on. In one, the user will feel part of a structure, in the other she will feel the center of the space.

Considering that enterprise collaboration products are still quite immature, the current state of affairs as it relates to supported collaboration contexts seems to be either-or. It is common to read and hear from practitioners that “SharePoint is document-centric”, “Clearspace is conversation-centric”, “Confluence is wiki-driven”, “Connections is activity-centric”, and so on.

Today, when most collaboration products are benefitting from the hype surrounding collaboration to fight for territorial expansion (as many customers as possible, as fast as possible), each vendor touts the success of its selected context as proof that it is THE context that makes sense… but is that necessarily the way to go? Couldn’t or shouldn’t all collaboration products support ALL of them?

I believe (but have no way of supporting the opinion) that ultimately all collaboration products will support ALL contexts for collaboration, including some that are not generally available yet outside of specialized domains (e.g., virtual reality worlds used for online games).  I take my clue, again, from the way normal collaboration works: supporting only one modality would be equivalent to a person saying “No, I am sorry, but unless we follow strict project management methodology i will not let you help me wash the dishes”.

Some of the key functional categories delivered by the context layer include:

• Workspace management: The ability to instantiate new workspaces in an ad-hoc, planned and/or automated manner, populate them with adequate artifacts, communications channels, workflows and content resources, and maintain them for as long as needed, preserving their state across time in the process;
•  Templates, workflows and components for quick initialization of particular types of workspaces, based on the purpose of the collaboration (i.e., team meeting, light project management, consensus-building discussions, documentation writing, etc.);
• Best practices to enhance collaboration around common functions and processes;
• Roles and responsibilities support, through the automated creation of content structures and personal workspaces that reflect common needs of users with well known 

The contextual layer is the “magic” one in the collaboration stack, because it creates the “illusion” of co-location and synchronicity, even if individuals are continents and time zones away from each other. The trick doesn’t need to be detailed and immersive from a virtual reality point of view: we are all familiar by now to the intense emotions that can be experienced while reading a posting in a discussion forum or wiki, even if we have never met the other person, or know much about her: to the reader, she is present while he reads, no more or less than if they were in the same room.

However “magic”, it’s important to note that virtual workspaces are just starting to evolve; by any standards, today’s workspaces are primitive and cumbersome to use, based mostly on text, and mostly serial in workflow. Most of these limitations are caused by the limitations of today’s “universal” thin client and the lack of vendor commitment to “playing by the rules” of already available standards (specially for communications, such as XMPP, SIP, SIMPLE, etc.). As those limitations are eliminated in the next ten years, workspace metaphors will:

• Abandon today’s myopic attachment to emulating narrow-band interactions (just like the traditional example of building over-dimensioned steel bridges that sank under their own weight, just because all engineers knew was how to build with wood);
• Deploy assistive technologies and artifacts in standard ways that may become intuitive to all (as opposed to having to hunt all over the user interface to find the correct widget for the interaction modality that the user has in mind). 

At that point, EC products will start to realize their very unique potential to give birth to massively distributed collaboration, supporting collaboration at a scale, scope and demographics that are not nearly possible in co-located cases (i.e., “massively distributed collaboration”, a term coined by Mitch Kapor). I personally believe that, despite all the energy and attention wasted by the early hype on the transformational power of the Internet, massively distributed collaboration does have the potential for significantly transforming culture and society, and ECPs will carry those transformations inside the enterprise as well.

Collaborative Application Services

Despite the strong focus on end-user services, the “golden pot” of ECPs is the integration of collaboration services into back-office enterprise applications such as SAP, Oracle Apps, etc. I have already described as an attribute of the content layer the ability to bring into collaborations data from those app: this functional block reverts the direction and makes it possible for collaborative workflows and services to be intertwined into the operation and workflow of back-office applications, creating in the process hybrid applications.

The power of such hybrid collaborative applications was validated many years ago, as IBM rolled out Lotus/Notes as the first true enterprise collaboration product. Soon afterwards, users started leveraging and implementing Lotus/Notes integration into back-office apps: according to recent industry surveys are one of the most critical barriers to switching to other collaborative platforms for most users still holding to Lotus/Notes, despite willingness to do so. Not surprisingly, Microsoft’s is very busy promoting similar integration capabilities in their Collaboration Platform, and trying to create IBM-to-Microsoft migration scenarios in which the collaborative services of one are replaced by those of another.

The collaborative Application Services (CAS) later is defined in terms on integration points, API’s and web services that can be used to selectively invoke collaborative processes from other applications, using multiple client types. Thus, the important attributes of this layer are programmability, architectural separation and customizable behavior, transforming the individual components and services of the EC product into a toolbox to be leveraged by other enterprise applications.

Effectively, the CAS functional block marks the dividing line between enterprise collaboration products and enterprise collaboration platforms, because this layer makes it possible for collaborative activities to become part of the workspace, regardless of process or activity, as opposed to a distinct activity that users conduct.

Summary

The result of this exercise is a first-order taxonomy that looks as follows:

Enterprise Collaboration

• Core Services
  • User Management
  • Group and Relationship Management
  • Artifact Management
• Supportive Services
  • Knowledge
    • Content interaction models
    • Knowledge representation and discovery
    • Expertise and resource location
  • Content
    • Integration
    • Discovery
    • Access
    • Creation
    • Storage
    • Sharing
    • Modification
    • Distribution
    • Archival
    • Publishing
    • Consumption
  • Communications
    • Synchronous and asynchronous channels
    • Web services
    • Contextual communications artifacts
    • Other artifacts and services
• Context Services
  • Workspace Management
  • Templates, pre-defined workflows and components
  • Best practices
  • Roles and responsibilities
• Collaborative Application Services
  • Standards
  • APIs
  • WEB Services
  • Protocols

In a coming article, we will take it from here into more detail. Also, the writing of this taxonomy made me realize how immature the collaboration market is (and my understanding of it as well): I found it infinitely more difficult to try to create an abstract taxonomy than addressing the need with a pre-defined scenario in mind, as when I go over requirements for a collaboration deployment with a customer. As a result, I expect the taxonomy above to go through several improvement passes, and I would welcome your input in the matter.

This entry was posted on Sunday, July 20th, 2008 at 2:51 pm and is filed under enterprise collaboration, social networking, web 2.0. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.