[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
Re: [xml-dev] Is it time for the binary XML permathread to start up again?
- From: noah_mendelsohn@us.ibm.com
- To: "Pete Cordell" <petexmldev@tech-know-ware.com>
- Date: Mon, 23 Jul 2007 08:53:37 -0400
Pete Cordell writes:
> I wouldn't be surprised if FI + gzip was actually faster than gzip
alone.
Me either, but gzip isn't primarily about speed. This is a little like
saying: if you hook a race car to a Honda Fit (Honda's small economy car
in the US), it may go a bit faster, and you may still use less gas than
the race car would running at full speed. Right. If your primary point
is to save gas (I.e. get compression) then it's possible that running FI
first well help some with that. My point is that you'll almost surely
wind up in a speed range (a bit faster than gzip) that is way off from the
reason you'd use FI if you wanted speed (analogy to race car).
It still seems to me that the main reason you want FI is either pure
speed, or that speed with modest compression. Fine that's quite useful.
It still seems to me that the main reason you want gzip is for much better
compression when speed isn't a big issue. Fine, that's quite useful.
As best I can tell, the FI+gzip combination is what you want when you were
after what gzip gives you, and are willing to deal with some extra
complexity to slightly modestly alleviate its speed issues.
--------------------------------------
Noah Mendelsohn
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]