On Mon, Aug 4, 2008 at 3:48 AM,
<Sebastian.Karlsson@mmorpgs.org> wrote:
Hi,
I'm considering using boost::serialization but have a few questions which I would love to get answered.
1) In the overview section performance is nowhere to be seen as a goal, which for my use case is very important. If I were to use the binary archive, how well would it perform in comparsion to a hand crafted optimized serialization aproach?
>>>> When I compared a bruit stupid << stream
overload >> method of serialization it was about 2-2.5 times
faster the serialization. I firmly believe that I loose performance to
increase flexibility, and depending on your requirements this may or
may not be acceptable.
I've seen in the examples that strings seems to be used to identify data, won't this create a large overhead for both deserilzation and storage?
>>>> This depends on your implementation. It's your choice on how you wish to "chunk" your data.
2) I saw the thread about the reason for boost::any not being serializable. But I don't really care about portability, so in that case, is it possible to implement a non intrusive serialization aproach to boost::any? Would really appreciate some little snippet for that if that is the case. If that's not possible my last option would naturally be to make an intrusive implementation, I see no obvious aproach here since I can't make a virtual template function, so I'm kind of hoping someone could push me in the right direction if it comes to this.
>>>> Someone else would have to chime in here, but you may want to take a look at tuple serialization examples.
Besides this I can only say that the library have a very slick interface and I really dig the design :)
// Sebastian Karlsson
_______________________________________________
Boost-users mailing list
Boost-users@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-users