Following on yesterday's blog, I wanted to reflect on a more specific Wiki innovation. My dear friend, and software genius, and wonderful human being, Dan Bricklin, in his understated way has launched wikiCalc, which allows the user to run a spreadsheet on a distributed basis, and enables the tracking of all versions, and changes to the spreadsheet over time. In addition, you can populate a "cell" with anything from a number to an active web link, and beautifully it has both a server side, and independent client side application, which can be run independently. Put another way, you can make a change when sitting on a plane and publish it to the common spread sheet when you land. Why do I think this is such a big deal? Well, spreadsheets are the key, interdependent control system used by large organizations. GE, BP, American Express, and all large organizations have thousands and thousands of financial models in Excel, which are the basis for budgeting, and day to day management of the company. One of the big challenges of this situation is that there is no "compare" function for spreadsheets, the way that there is for word documents. If you want to understand what the nature of the changes are you need to have some fantastic annotation, or you are just out of luck, and figuring out what is "going on" in a spreadsheet is an exercise that is akin to figuring out the workings of a LISP program.
These financial models are very important to the running of large organizations, as they serve a role to both articulate what the given function or division will do (e.g. what sales are you projecting at what margins), but they are also often the control system by which senior management reviews progress, or lack of progress and uses it to guide the business on a day to day basis.
The mental fragmentation that occurs in today's spreadsheets is huge, and creates a compartmentalized version of "what is, is" as former President Clinton famously remarked. Many meetings are about what the facts are, not just what to do about the facts. Much of this fragmentation comes from the fragmentation of the numbers, and the expectations and definitions in spreadsheets -- large and small.
Enter wikiCalc. If a large organization could use wikiCalc to allow each relevant party to update their part of the spreadsheet, and track the changes that each has made, it would make the job of documentation (which is automatic) and integration much, much easier. Now, according to Dan, the current calculation engine can handle a few thousand cells of data, but if it is used for models of that size, it should only be a matter of time before a larger, more robust calculation engine is built, and the large spreadsheets could be done as well.
Susan Leigh Star, created a theory of mental objects in large organizations that sit at the nexus of two or more social worlds. She called them "boundary objects", because they exist at the boundary of two mental models of the work. Medical databases are one example of a boundary object that sits at the nexus of doctors, nurses, hospitals, insurance companies and patients. She documented that before medical databases were in place, there were an infinite number of diseases. Medical databases classified diseases into categories and thereby helped all parties make "decisions" in the very process classification. All together, these classifications created a database, which became a boundary object. Computer Aided Design/Computer Aided Manufacturing (CAD/CAM) systems serve a similar translation function between engineering, manufacturing and marketing.
In large social organizations, boundary objects serve a number of vital functions, they are a repository for extant decisions that have been made, and they provide a translation function across social worlds. So, if we imagine wikiCalc as a boundary object (which I think it will be, and which the financial models Excel often are already), then it can help "translate" the needs of the sales department into the needs of the human resources department, into the needs of the financial department. All these departments don't need to understand all the decisions and rules, and data at the same level of detail, they simply need to understand "their part" and believe that the model does a good job of translating across the different social worlds.
Who cares? Well, given the tremendous amount of time that organizations spend creating, maintaining and managing against spreadsheets, I think that the creation of a distributed, self-organized, easy to edit spreadsheet that can track all versions across all time would finally give organizations a tool to meet the interdependent needs of budgeting and control in a robust way. The problem with doing this with something like SAP or other enterprise requirements planning tools is that they are so rigid, and centrally controlled. This is why Excel continues to be the planning tool of choice, despite the almost universal adoption of ERP by large companies (even at MIT!). But, Excel's problem is that it is too decentralized, and has no easy way to allow for the "publishing" of a new version as part of a common model, or common "space".
Imagine that a financial model in an organization should be more like the WikiPedia, than like a set of independent, free floating models that then get shoehorned into a centralized control system like SAP. It would be much more efficient, accurate, and easy to use if it were self organized and decentralized. It is for these reasons that I think wikiCalc is a very big deal.
Furthermore, as is true of all beautifully designed, open tools, the unforeseen uses will be amazing, and probably swamp the expected ones. Who knows what people will do when they have a spreadsheet representation to organize, calculate and distribute any type of local information, any data or news feed, and display it in the most universally understood of all computer-based representations, the spreadsheet. It is cross language and global. The mind reels.