From: Peng Yu (pengyu.ut_at_[hidden])
Date: 2008-07-11 13:25:40
On Fri, Jul 11, 2008 at 10:13 AM, David Abrahams <dave_at_[hidden]> wrote:
> on Fri Jul 11 2008, "Peng Yu" <pengyu.ut-AT-gmail.com> wrote:
>> On Thu, Jul 10, 2008 at 2:47 AM, Sergei Politov <spolitov_at_[hidden]> wrote:
>> You understood me correctly. But it seems the transform from my code
>> to your code is a little complicated. Is there any more direct analog
>> between boost::fusion and the runtime algorithm. In particular, how to
>> directly use two iterations with the second iteration start with
>> something like it + 1 in my original code?
> If you want something more direct, you may need to stick with the
> runtime world.
> What is your goal in using fusion? If it isn't because you need to run
> faster, you might look at
> http://stlab.adobe.com/wiki/images/c/c9/Boost_poly.pdf, which could give
> you some ideas for removing the coding burden of runtime polymorphism.
I do want it fast. The application is double dispatch. Using static
polymorphism instead of dynamic polymorphism should give at least a
few time runtime improvement. Do you mean that it is not possible to
write the code in fusion as easy to read as my original code?
I didn't quite catch what the key point in the slides that you
pointed. Would you please let me know what is about and how it could
apply to my problem?
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk