Re: [boost] Is Boost-CMake Dead or Alive?

Subject: Re: [boost] Is Boost-CMake Dead or Alive?
From: Vladimir Prus (vladimir_at_[hidden])
Date: 2010-09-01 02:59:51

Eric Niebler wrote:

[moving to -devel, where it really belongs]

> On 8/28/2010 3:12 AM, Vladimir Prus wrote:
>> David Abrahams wrote:
>>> At Thu, 26 Aug 2010 13:41:46 -0400, Eric Niebler wrote:
>>>>
>>>> Even if we achieve our vision, it's not clear that Boost as an
>>>> organization will adopt it.
>>>
>>> <ahem>
>>>
>>> A small correction: even *when* we achieve our vision, there's no
>>> *guarantee* that Boost will adopt it.
>>>
>>> [But Boost will adopt it.  Just promise not to sue us in case they
>>> don't :-)]
>> 
>> I find the above statement to be a very interesting advertisement approach
>> for a project that addresses requirements that were never discussed on
>> this list, and which itself was never discussed much. In fact, this statement
>> sounds like you have made some secret arrangements (*), or even that you had
>> some revelation from supernatural powers -- which is probably not what you
>> was going to communicate.
>> 
>> I think Eric statement is much more balanced.
>> 
>> - Volodya
>> 
>> (*) In fact, ryppl gave an impression of some secret arrangements from the
>> very start.
> 
> Dave was expressing confidence about the value of our work, no more.
> There have been no secrets and no back-room negotiations. 

Eric,

unfortunately, I must have been taught either wrong English or wrong logic,
because surely "I am confident that XXX will YYY" conveys quite a different
meaning than "XXX will YYY". While now and here, you can clarify the meaning,
I doubt you'll be around when any given person reads such statements in his
email client.

And of course, when a representative of a third party having commercial interest
in the matter makes public statements to the effect "XXX will YYY" (without
any "I am confident" qualifier (*), this surely makes the matter look even more
suspicious.

> Boost's
> scalability problems have been discussed endlessly here and in open
> sessions at BoostCon, 

I think "endlessly" is a very accurate word here. The discussions have failed
to clarify, prioritize, or even list the issues. Say, recently Dave has announced
that he'll no longer maintains his libraries (**) because SVN is slow and library
organization is a "mess". While other folks agree that there are some issues,
they don't seem to have that drastic effects in general. That thread has died,
and we still don't know why SVN performance (and other issues) is enough to make
one person stop maintaining his libraries completely, while another person is
happy to have extra latte while SVN does one-a-quarter merge to release branch.

> but no decisions about Boost's future were made or
> will be made by fiat. Boost doesn't work that way.
> 
> It's true that the requirements of, or even the need for, ryppl was not
> debated here on this list. But that's how most Boost libraries start.
> Someone has a vision and builds a thing. Then they bring it to Boost and
> it's improved and accepted or rejected. Trying to get quorum on a thing
> that isn't built yet is just hard. "If you build it, they will come" --
> so we hope.

It is clear that each project involves a vision (very hard to share), 
design (somewhat shareable) and review (share as you like). And it's
perfectly fine to have a vision that involves fixing all issues with
Boost development that you are aware of. However, saying "Boost will
adopt ryppl, no we won't bother to even list the issues it will fix"
sound bizarre -- because it's an attempt to share a vision -- which
is not shareable.

- Volodya

(*) Check recent boost-cmake postings for an example. In fact, said party
even has marketing materials which *already* mentions Boost. 

(**) http://old.nabble.com/Maintenance-suspended-td29100510.html