Introducing OpenWorkspace®

The central idea behind the Information Processing Model is that as we go about our day we see and hear ‘stimuli’ that we process momentarily to decide whether or not to pay attention to it.  Applying the model to the use of the desktop, we might see (on our calendar) an event scheduled to begin in 10 minutes.  Or we might hear (by audio notification or by inner-voice) a reminder of it.

Within seconds, we decide whether to forget it (by discarding it, for now) or to give it more attention (by transfering it to ‘working memory’.  For instance, the upcoming call might not require any further thought or preparation (or it might not be attended at all), in which case the call reminder is ‘discarded’, and given no further thought until the next time related stimuli resurfaces.  On the other hand, if the call needs preparation or the desktop needs to be readied for it, the call reminder is transferred to ‘working memory’, where a mental image symbolizing the call (e.g., an image of the other call attendee’s LinkedIn profile) is used to aid in recalling (from long-term memory) the additional information needed for the upcoming call.

In working memory, the user constructs a ‘mental model’ made up of the collection of information relevant to the call, becoming a focused task that demands the user’s limited thought.  This circular process of (working memory) requesting and retrieving (from core memory) additional information needed for the call continues until a more comprehensive view of what’s needed has been arranged—mentally (in working memory) then literally (on the desktop).  When the call is complete, working memory does its best job to ‘encode’ the updated mental model so that it can be retrieved in its last-known state when the call or topic is revisted sometime in the future.  Only then can the mental model be discarded, freeing up working memory to do it all again for the next topic or task at hand.

The doing required.

Key ideas and assumptions made here and in future posts about what we do on the desktop are based on long-established models like the GOMS Model (Goals, Operators, Methods, and Selection) and the Keystroke-Level Model (KLM).  In future posts, we’ll illustrate the doing problem by deconstructing the manually-intensive operations required to use the classic desktop.

The Memex (short for ‘memory extension’, or ‘memory index’) was intended to work as a personal knowledge system, a digital scrapbook of sorts that would allow users to store, navigate, and retrieve vast amounts of information by association (think hashtags), much like how the human memory worked, in terms of retrieving and processing associated information in working memory then encoding it into core memory, as theorized by the Information Processing Model and its supporting research.

Upon recieving input from the user (for instance a keyword, like “bow & arrow”), the Memex would surface a subset of information related to the topic, then display the information on multiple monitors in order to give the user a holistic view of the topic. The user could then page through the collection of information by using mechanical controls like knobs, keyboards, and levers, in order to add more information to the collection or to edit or reorder its contents.  Once the collection of information (or the ‘trail’ of information, as Bush referred to it as) was updated to the user’s liking (or once a new topic called), the collection could then be ‘saved’ to storage in the form of a reel of microfilm that could be ported from Memex to Memex in order to share knowledge and distribute collections of associated information among individuals.

Though the Memex never became a reality (it was technologically impossible at the time), it became a primary influence for how the personal computer came to operate three decades later.  Namely, it was foundational in establishing the construct of file systems, along with the mechanical concepts that became fixtures of computers (the processor, disc-based storage, keyboard, the embedded display).

That said, its intention to augment human information management was notably never fully realized by commercial operating systems, where users spend an excessively high cost of time, energy, and cognition to overcome information management constraints imposed by application-centric navigation and file management.

1. 1. Hierarchical File Systems: Modern desktop operating systems predominantly rely on hierarchical file systems. Information is stored in a structured manner—files within folders, and folders within directories—rather than in a web of associative links as envisioned by Bush. This model limits the fluidity and non-linear access to information that the Memex concept aimed to facilitate.  The fragmented nature of tabs, windows, and applications underscores a lack of continuity in how information is stored and accessed. The constant switching between contexts hinders the creation of a cohesive, associative network of information akin to the Memex.
   2. Search Mechanisms For Single Resources: Modern search tools in operating systems rely heavily on keyword-based searching rather than associative recall.  While helpful, this type of keyword-based search surfaces individual files, apps, and tabs, requiring users to incrementally repeat the process of searching and surfacing each resource one-by-one until a cohesive collection of information is arranged in view on the desktop. This approach often leaves the user with disjointed results rather than with a coherent trail of resources related in context, as the Memex would have provided.
   3. Limited Associative Linking: While hyperlinks within documents or web pages allow for some form of associative linking, this capability is highly limited within the operating system itself and not inherent to how users interact with the classic desktop that was never designed for the kind of cross-referencing and trail-following that Bush imagined.  Today we string together trails of associated information manually, our options limited to ‘tab groups’ and single applications that integrate information from other applications into a single window traveresed by side-navigation. Meanwhile, at the system-level , users cannot easily link files or pieces of information across different applications in a way that Memex proposed would mimic human thought processes (like constructing collections of related information in working memory).

Existing solutions address this problem within the application layer, where many web apps integrate (programmatically connect) with unrelated apps to bubble information up into a single-window user interface.  For example, project management or chat apps typically integrate with document management apps (or vice-versa).  And the chat app may also be connected to an app or ecosystem where documents are managed, or where employee payroll is handled, or to some enterprise resource planning app, for example.

OpenWorkspace® addresses this problem more holistically, at the system-level, across the application layer, both within and outside the web browser, tying together collections of information from disparate sources so that the user can open all of it at once, much like opening a file.