I think it makes sense to start with a very basic textfile search capability first and then expand it if and when demand occurs.
Two rather non-complex systems could be:
1. comma separated data. Example:
tag1 data1 data2 data3
tag2 data4 data5 data6
When the user inputs "info tag1" FARR would output "data1 data2 data3" (or "data1 \n data2 \n data3")
Mousers example ("info phone hospital") in comparison adds one level of complexity. But how often is that needed? I suspect that it for much lookup usage is enough to output all the info related to a "first level tag" i.e. "info hospital" would display adress, phone number, email etc in FARR. Two drawbacks with the more complex example is that the user must memorize all the second level categorizes and must create all his/her new data in a very well formed and consistent way. I definitely prefer something less complex and less (human) memory hungry.
-Nod5
Also like this idea of getting a simple yet scalable info retrieval plugin out now and improving it as time allows instead of trying to create the ultimate plugin right from the get go and never getting it off the ground. The only thing i would add to the idea above would be to include some kind of
optional meta data or options that farr could interpret for display/functional purposes that would be listed at the beginning of each file. Using the example above:
@type: csv
@displaymode: grid
@defaultsort: asc
@headers:name, number, address, phone
tag1,data1,data2,data3
tag2,data4,data5,data6
....
A great example for this functionality could be an address book feature which allowed you to search for a name and have it display all the relevant info (name, address, etc) in the upcoming grid mode. You could also then extend the features down the road to include files that are formatted differently (xml etc) as well as different display modes and support for additional meta data/options. I think as a first step support for csv files should be a primary goal as most info from popular apps (ms office) can be exported into this format.
Either way cant wait to see what you come up with. Hopefully sooner rather than later :D