[dtsingletary]

I think the point you're missing is including a technical one: for a good number of users, getting a tweet history "hits the disk." The limited history is in part due to what's in memory on their servers and easiest to serve.

The historic tweet providers don't have this problem-- they (currently) can monetize to offset the cost of the servers to store the data in a necessary way, or by being for analysis only, the fetch-speed (if from disk) isn't as much of a concern.

Building a business on Twitter doesn't require using one of the other providers, unless that business is analytics.

Rate limits aren't some boogeyman. They're a financial and technological necessity to dissuade abuse and plan for capacity. Serving data isn't free (as in money). Have we had a semi-public API with as much data served in the past? How do you offset that "cost-center" right now?