This answer summarizes some of my comments.
HFT is certainly a very hot topic these days, but it's hard to point to any one reason. A large part of it is the mystery and the profits, but also part of it is the relative novelty. Note that there is no lack of papers about medium and low frequency strategies, it's just that they are not labeled as such. Medium and low frequency strategies had their day in the limelight back in the early to mid 2000s.
Medium/low frequency strategies, to name just a few, include momentum, reversal, earnings quality (accruals), post-earnings announcement drift, (low) volatility effect, as well as options strategies such as dispersion, volatility risk premium and futures/FX strategies such as the carry trade, momentum (again). So you see, there is no one type of medium/low frequency strategy, but most HFT strategies are relatively similar.
As to how "high" frequency must be to be considered high frequency, opinions tend to differ on this point (see earlier question), but I doubt most participants would label a trading strategy which must execute within a second and holds for a day to be "low frequency." In fact, it may not be very high or ultra high, but most investment managers would broadly consider it to be "mid-to-high" frequency. Some surveys show (see Shane's answer) a significant minority of managers (about 15%) consider even a week holding period to be high frequency. Back in the early days of academic research, high frequency was anything using daily (rather than monthly) close data.
As for data input, most low frequency strategies use daily or even less frequent data. Although it is hard to define, IMO medium frequency requires intra-day data for execution, but only at the point at which market microstructure can be ignored (generally 5-15 minutes, depending on liquidity). Anything which requires data at a frequency lower than 5 minutes must necessarily take microstructure into account, and that qualifies it as high frequency. As an aside, I believe that the myriad of additional issues that arise when dealing with tick data make it entirely not worth the trouble for anything but truly high frequency strategies.