Re: [boost] Re: New design proposal for boost::filesystem

From: Carlo Wood (carlo_at_[hidden])
Date: 2004-08-21 18:27:47

Let me start with an appology for the rather unfriendly tone
of my previous message.  Fortunately people on IRC pointed this
out to me; I didn't intend the post to be the way they read it,
but I was wrong nevertheless.

The point is that I just don't have time for (nor I am interested
in) long discussions about design issues, and I think that actually
I shouldn't even have started this thread ;).  Due to experiences
in the past, on other mailinglists mostly, I have gotten over
sensitive to remarks like a short "no" and "that doesn't work"
and the likes - without back-ups, examples or constructive remarks.
I think I was in error to see Davids request in this way (as if he
was making my life unnecessary hard), while in fact he was just
asking for well - examples :).

On Sat, Aug 21, 2004 at 06:33:10PM -0400, David Abrahams wrote:
> Carlo Wood <carlo_at_[hidden]> writes:
> 
> > On Sat, Aug 21, 2004 at 07:39:43AM -0400, David Abrahams wrote:
> >> When proposing design changes, first you have to describe the
> >> problems your changes are designed to solve, or nobody will
> >> understand the motivation.
> >
> > If you are not aware of the problems with boost::filesystem
> 
> I am aware of some problems I had, but I'm not sure those are the
> same ones you're trying to solve.
> 
> > then I just give up immdeately, instead of wasting my time
> > by trying to convince you about the need for a redesign.
> 
> Easy, big fella!  I'm long since convinced that something should be
> changed, but until I know what you're trying to address with your
> changes I'll have a hard time deciding whether the problems are
> important and whether your proposal effectively addresses them.
> 
> > I've seen several people already agreeing that there is a design
> > problem with the current boost::filesystem.
> 
> I'm among them.  However, I have a hard time articulating what I found
> difficult about the library, other than that the path portability
> checks seemed to get in my way.  If there's something deeper going on,
> I'd like to hear your opinion of what it is.

I am afraid that I too have a hard time to point out any clear
error that is so clear that everyone will immdeately agree that
it is an error.  This thing is hard to grasp, and THAT is the reason
that I think that it will be extra-ordinairy hard to convince enough
people on this list that a change is not only needed, but that a
particular change will be the one that is needed.  If not impossible.

I therefore, unfortunately, will have to stick to my opinion that
it is too unlikely that this discussion is going to lead to an
actual change of boost::filesystem, to put more time into it.

The only approach that might work that I can think of is a re-newed
analysis of *a* boost::filesystem - as if without looking at the current
implementation.  Then, if we go step by step, we might come to an agreement.
However - only if everyone would participate from the beginning.
And in practise that will not happen.  A few people would agree, one
would write an implementation and THEN 10 others will disagree because
<argumentation using the current implementation> blah blah.  Brrr.
Ok, I might be afraid of something that wouldn't be happening,
but I think that I can only TRY it if I can see boost::filesystem
as one of my projects - and really, it is not, I am LOADED with
other work as it is, already.

I already give a few ideas in other posts.  If _anyone_ likes it,
then feel free to defend it towards others and go through the process
of dicsussion and refinement.  If any additional clarifications
are needed then of course I will be happy to give them.

-- 
Carlo Wood <carlo_at_[hidden]>