Re: [boost] [Boost Review] Property Tree Library

From: Gennadiy Rozental (gennadiy.rozental_at_[hidden])
Date: 2006-04-24 10:53:34

"Jody Hagins" <jody-boost-011304_at_[hidden]> wrote in message 
news:20060424102955.2c83d6b2.jody-boost-011304_at_atdesk.com...
> On Fri, 21 Apr 2006 10:54:57 -0400
> "Gennadiy Rozental" <gennadiy.rozental_at_[hidden]> wrote:
>
>> I could achieve the same with multi_index serialized/deserialized from
>>
>> different archives. And I do not see any advantages in your solution.
>
>
> Gennadiy,
>
> You have stated several times that this library is not needed because it
> is easy to replace with some combination of multi_index, serialization,
> and program_options.  Could you please provide us with an example?

You better pick up the pace. We way beyond this already ;)) Try Problem 
Domain thread  for more on my views.

> I think this is especially important because both MI and serialization
> have large learning curves, and a number of us have never really used
> either in real work.

I would assume respective library authors may disagree.

> For myself, I've played with serialization only enough to realize that
> it is way too slow for any real work we are doing.  This type of stuff
> is probably not performance prohibitive from a serialization POV.

I don't see a performance to play any role in this discussion. Once you 
chose DOM model you already through performance out of the window.

> I have also tried program_options, but it is just too awkward to use for
> anything but trivial applications.

Oh. Here we are in complete consensus.

Gennadiy