The main point of SSDs in the mainstream market are to accelerate operating systems and programs, not to deliver fast bulk storage. The performance pr cost for these applications (and databases) should be measured in IOPS/$, not bandwidth/$.
What i find most interesting of SSD is their combination of high IOPS/$ combined with really high IOPS/W.
Harddrives deliver around 1-5 IOPS/$ and 10-50 IOPS/W, while SSDs, like x25-M (80GB) can deliver 150-175 IOPS/$ read and 40-60 write, and roughly around 10.000 IOPS/W read and 3000-5000 IOPS/W write.
In addition, the accesstime of SSD are a couple of orders of magnitude faster than harddrives.
I believe we will se massive adoption of SSD as OS-drives throughout 2010, and the trend has already started, but has been halted a bit by low avalibility of x25-M gen2 and high prices for Indilinx based drives compared to Intel.
The tipping point for mass adoption will be when drives that can hold OS+progs (30-80GB depending on the user) with good consistent performance and no major drawbacks reaches 100-150$. I believe Kingston 40GB V (intel based with 5 channels used) will be able to satisfy the lower-capasity end of this, and deliver performance sufficient for HTPCs, netbooks, and "office computers", at cost of aquirement of 100$ or less.