I previously posted when I found that the largest number you could use in Google Calculator is 2^1024 [1]. This is the smallest number with exactly n divisors [2], where n is the answer to life the universe and everything. The result of a search for the most precise form of that number that Google would accept is 1.79769313486231580e+308 at Everything2 [3]. At the time, Google Calculator didn't like working with a very precise decimal representation, but now it is just fine with using 307 out of the 308 digits! [4] It will not allow you to work with the binary representation of this number as it is too long (1,022 digits).

At this point I wondered why there is such a limitation for the scientific notation when there is a virtually unbounded decimal representation. I believe the answer to this question is that 1.79769313486231580e+308 is the largest number you can fit in a C++ double, and that python supports arbitrarily long numbers in its long data type. Certain routines would have to be rewritten using python's long data type in order to support a more precise representation in scientific notation. This is just a WAG :) [6]

[1] http://blogoscoped.com/forum/11949.html
[2] http://www.research.att.com/~njas/sequences/A005179
[3] http://www.google.com/search?client=safari&rls=en-us&q=1.797693134862e%2B308..1.797693134862e%2B308&ie=UTF-8&oe=UTF-8
[4] http://www.google.com/search?hl=en&lr=&client=safari&rls=en-us&q=17976931348623159077293051907890247336179769789423065727343008115773267580550096313270847732240753602112011387987139335765878976881441662249284743063947412437776789342486548527630221960124609411945308295208500576883815068234246288147391311054082723716335051068458629823994724593847971630483535632962422413721-0&btnG=Search
[5] http://www.python.org/doc/2.2.3/ref/types.html
[6] http://acronymfinder.com/af-query.asp?Acronym=WAG&Find=find&string=exact