For a long time I have pondered a potentially worthwhile project: a database that contained the locations (and probable strengths, current commanders, etc.) of the Union and Confederate forces down to the regimental level. The data could then be displayed on an interactive basis. I would think that this would be hugely helpful in any strategic, campaign-level analysis.
With appropriate software, one could easily visualize the various units on each side as they streamed across the countryside and coalesced for battle. This should be especially useful for such campaigns as Gettysburg, Vicksburg, Tullahoma, Chickamauga, and Jackson in the Valley and it may well trigger some new theses about what happened and answer old questions as to why certain commanders acted as they did. It should also be of interest to many people on a purely genealogical basis.
It could only feasibly be done Wiki-style, in all likelihood, based on information gathered by individuals who knew particular regiments, batteries, or other such units. These would include reenactors, keepers of regimental websites, small-unit historians, et al. The resources would include the Official Records, published regimental histories, diaries, letters, and published secondary sources, along with other materials.
This data-gathering would probably be more difficult, however, for the Confederate armies, given the relative dearth of all types of information. But with the geography and time period restricted to just one campaign, as a start, it might only take a rather limited number of people a limited amount of time to accomplish something that may be hugely beneficial as an historical tool.
Over time, and with further increases in computing power and decreases in cost, more information could be input: naval forces, fortifications, census and other demographics, cities, railroads, roads and bridges (in existence or destroyed), topography, waterways, vegetation, weather, and on and on.
Beginning work on such a project would be years and years away, but the internet, wiki-software, and computerized cartography and display devices make this, I think, a reasonable undertaking.
When I worked with GIS I thought how cool it would be to create regimental level maps using GIS software but I was always intimidated by the scale of the effort. I had never thought of doing it wiki style. Brilliant.