Please describe a summary of the new feature (in a paragraph):
Silk is useful for object serialization, however, when the text processing
performance and the size of
data to be transferred are matter, we should use some binary format of Silk
data.
GZipping a silk data would be a quick solution, however, it still needs the
same parsing cost with
the text-based Silk after unzipping the data, and text parsing cost (including
readline()) is known to
be a major performance bottleneck for manipulating Silk data.
Please provide any additional information below:
Original issue reported on code.google.com by taroleo on 27 May 2010 at 3:25
Original issue reported on code.google.com by
taroleo
on 27 May 2010 at 3:25