Boost logo

Boost :

Subject: Re: [boost] [pimpl] Mini Review
From: Vladimir Batov (vb.mail.247_at_[hidden])
Date: 2011-05-27 20:10:12


> "Krzysztof Czainski" wrote
> ...
> If my tone was found impolite too, my apologies - that was not intended at
> all. I just wanted to express my thoughts about this library, which in this
> case were negative ones, but I assure you this isn't always the case ;-)

Look, Kris, you have nothing to apologise for. I took my frustration out on an
innocent passer-by... As it often happens in our not-so-fair life. So, apologies
are all mine.

Getting back to your emails about Pimpl, I firmly believe that critisism is
healthy and in my experience it always results in better software. I always
strive to address whatever suggestions/criticism/feedback I can get. What I
found somewhat frustrating with your comments was that I could not address/fix
things which you seemingly was not happy about.

Say, you say "Pimpl library tries to solve too many things at once". It's the
critisism which I as the developer want to address... but cannot. What things
exactly you think do not belong in Pimpl? What am I supposed to do to make you
happy? As it is, it's a one-sided show -- you criticise and I just sit there as
a "whipping boy" unable to do anything about it -- not to fix it not to defend
it. Not fun.

Or you say "Individual use cases might want those, not care about them, or
specifically not want them". My interpretation is like "it can do this, I do not
care for this, so it's bad". In such a setting how a library developer is to
satisty more than one customer if customer A says I do not care for the features
needed by customer B and customer B is saying the opposite.

If on the other hand you say I do not think safebool() belongs with Pimpl, then
we can talk about it. In the end I might jump sides and admit my original design
was wrong or I might have certain use-cases which you do "not care about" but I
need to support to keep another customer happy. In the end we might agree to
disagree but that'd be after a productive two-way discussion. I hope you see my
point.

Best,
V.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk