Re: Replacing the ML with a forum

From: Robert Ramey (ramey_at_[hidden])
Date: 2023-04-28 18:20:01

On 4/27/23 8:11 AM, Vinnie Falco via Boost wrote:
> On Thu, Apr 27, 2023 at 7:56 AM Robert Ramey via Boost
> <boost_at_[hidden]> wrote:
>> I fail to see what is to be gained by replacing the current system.
> 
> The current system is not being "replaced" it is being enhanced, with
> the OPTION (but not the requirement) to interact with the mailing
> lists using a modern, forum-based interface. Just as there are folks
> who have expressed that they would not participate in Boost
> discussions outside of the mailing list interface, there are as many
> (or more) people who have expressed that they will not interact with
> Boost through the mailing list at all. You don't hear from them
> because well, they don't post to the list (the dog that didn't bark).
> 
> You ask, what is to be gained? 

> Well, the largest existential threat to
> Boost is that we do not refresh the foundations of our social
> technology and our membership. 

I don't believe this.

> Library membership in Boost no longer
> carries the cachet that it used to.

Perhaps -

> Authors of new libraries
> increasingly prefer to go directly to the C++ standard instead of
> being thoroughly vetted in a Boost review. 

True - Because its much easier to get something into the standard than 
it is to get it into boost. For one thing Boost requires a working 
implementation, test suite, and user level documentation. For the 
standard, all one needs is to convince a small group of one's buddies 
that the world needs the authors proposal.  This is especially effective 
when the proposal is too opaque for any normal person to understand - 
thus guarenteeing that there were be no real objections.  That is until 
its features get added into the standard and real people have 
difficulties implementing and/or using them.

If we want to continue to
> maintain Boost's extraordinary relevance, we must do the following:

> 1. Make the mailing list more appealing to younger generations
> (millenials and the TikTok youths)

I don't think this is a real problem. I'm trying not to sound like a 
curmudgeonly old person, but really - are millenias and TikTok youths 
our target audience?

> 2. Increase participation in the Boost Formal Review process

To do this we need to attract better libraries and work harder to 
promote the review.  Also the 1-2 week window for reviews is a big 
problem.  I might want to do a review but I can't right now.  One of the 
features of my "Blincubator" project was to permit one to post reviews 
long in advance of the official review period.  In fact.  We might delay 
the formal review period until there are some magic number of reviews 
already in the queue.

> 3. Spark the imagination of the new generations of library authors
>      - to write new libraries
>      - to submit their libraries to Boost for review
>      - to become active maintainers upon acceptanc >
> 4. Inspire people to contribute to Boost
>      - by using the libraries, and providing feedback
>      - by contributing to individual projects
>      - by helping with documentation
>      - by helping other users with questions or problems

I think the best way to do this is for boost to create better software. 
This would mean less bugs/design flaws, better documentation, less 
quirky tools.  Why can't we apply the standards we use for user 
libraries to boost infrastructure?

This would mean: boost tools like the following:

a) documentation chain
b) build system - Bjam
c) build system - CMake integration
c) CI
d) GitHub integration
e) website development and testing
f) probably some others I forgot

would be subject to similar requirements as boost user libraries.  That 
is, Test suite, documentation, Formal review.  None of the boost tools 
have been subjected to these requirements - and anyone who tries to use 
them notices.  This is probably the single greatest obstacle to 
continued progress of boost.

> The mailing lists must never go away, for as long as there are even a
> few people who prefer them. But this cannot be the sole method to
> participate in the Boost social space. To stay current we must provide
> additional choices. The need for a forum interface is obvious.

Don't the #boost and #boost-user forums on Slack currently fulfill this 
requirement?

Robert Ramey