Re: [boost] Variant (aka Discriminated Union) implementation

From: Douglas Gregor (gregod_at_[hidden])
Date: 2002-06-25 08:32:25

On Tuesday 25 June 2002 02:19 am, Ruslan Zasukhin wrote:
> > I'm glad to see work on a variant type. It's been discussed a million
> > times, but we've never had a variant type come up for review (and I've
> > needed one for a LONG time).
>
> I have work on similar task in Valentina database not so far.
> But I have made code using Interface way.
>
> So I wonder (note, I am new to boost)
>
>     Why boost org always try go by template way only ?
>     Why I do not see solutions based on interfaces ?

You do see solutions based on interfaces, but the interfaces are resolved at 
compile time using templates instead of at run-time using virtual functions. 
With a variant type, for instance, this allows complete type checking to 
occur at compile-time instead of pushing all checking to run-time.

        Doug