On Tue, Aug 10, 2010 at 5:21 PM, Eric Niebler <e...@boostpro.com> wrote:
> On 8/10/2010 11:14 AM, Robert Jones wrote:
>> Well, as a complete novice to code of this sophistication I
>> understood that piece perfectly, as far as it goes. Naturally, as the
>> opening piece of a series it raises far more questions than it
> That's great feedback, thank you.
Let me follow up. I like it too!
>> It also scares me somewhat. This stuff could mark an absolute
>> explosion of complexity in the code your average jobbing programmer
>> is expected to get to grips with, and in my experience the technology
>> is already slipping from the grasp of most of us! When you get this
>> stuff wrong, what do the error messages look like? Boost.Bind &
>> Boost.Lambda errors are already enough to send most of us running for
>> the hills,
> A great point! (I've held back a whole rant about how long template
> error messages are library bugs and should be filed as such. That's a
> whole other blog post.) I sort of address this when I say that a good
> dsel toolkit would force dsel authors to rigorously define their dsels,
> leading to "better usage experiences". That's pretty vague, though. I
> could be more explicit. But certainly the intention here is that proto
> makes it easier for dsel authors to give their users more succinct error
I think we see a great improvement with static_assert in C++0x!
And we are potentially able to reduce error message if SFINAE is
applied more often, with the disadvantage of losing information on
>> and tool support is somewhat lacking as far as I know,
>> being pretty much limited to STLFilt.
>> Maybe I'm just too long in the tooth for this!
>> Still, great piece, and I look forward to subsequent installments.
> Eric Niebler
> BoostPro Computing
> proto mailing list
proto mailing list