Re: [boost] [beast] Request for Discussion

Subject: Re: [boost] [beast] Request for Discussion
From: Vinnie Falco (vinnie.falco_at_[hidden])
Date: 2016-09-23 12:25:39

On Fri, Sep 23, 2016 at 8:49 AM, Paul A. Bristow
<pbristow_at_[hidden]> wrote:
> I'm not an expert on the subject, but I think you are right to be cautious
> about trying to be everything to all men.
> ...
> I think you should propose what you have now for formal review.

I appreciate the kind words and support for a formal review. I have
carefully studied all of the review comments provided during the
review of Boost.Http in 2015. In particular, a comment from Robert
Ramey:

from http://lists.boost.org/Archives/boost/2015/08/224700.php
"To me it's very unfortunate that so much effort has had to be
expended by a library author to produce a library that is not
accepted. It's also unfortunate that so many reviewers need to spend
this much time to dig up enough information to reach this consensus.
This illustrates my motivation behind the design of the boost library
incubator."

I think the idea behind the incubator is sound, and I agree with
Robert's assessment. Looking over the formal review comments for
Boost.Http, about half of the reviewers admitted they did not have
enough time to do a deeper dive. I would like to avoid these mistakes.
My approach is to get a higher awareness of Beast through discussions
on this mailing list, and through face to face meetings with
stakeholders and reviewers (for example at CppCon 2016, and C++ Now
2017). In the meanwhile, the number of users continues to grow.

The odds of success during a formal review can only go up by
collecting and addressing feedback from stakeholders before the review
begins. These are the set of individuals who participated in the 2015
review of Boost.Http:

Agustin K-ballo Berge
Antony Polukhin
Bjorn Reese
Darren Cook
David Sankel
Edward Diener
Glen Fernandes
Lee Clagett
Niall Douglas
Robert Ramey
Rodrigo Madera
Roland Bock
Tom Kent
Vinicius Oliveira
Vladimir Prus

I would like to hear from as many of these individuals as possible,
about their experience looking at Beast and taking it for a test
drive, before starting the formal review process.

Of special importance is to get feedback from Vinicius Oliveira, the
author of Boost.Http. We spent several hours talking about our
respective libraries with each other. Beast provides the lowest level
interfaces, while Boost.Http tries to offer the higher level
functionality that was recently requested. Specifically, Boost.Http
attempts to provide a turn-key server solution. I've heard that
Vinicius is busy working on a client, and has also implemented his own
parser.

I think public feedback on Beast from the author of Boost.Http is
crucial, since Beast claims to offer uncontroversial low level
functionality which Boost.Http would have to use if Beast is accepted
to Boost. Perplexingly, I have seen no evidence of any public dialogue
from Vinicius about his current client interface goals, or any plan to
refactor Boost.Http to use Beast. I'd like to hear his reasoning for
why a HTTP library needs to "support multiple backends" (what does
that mean?), a theme which came up often in his responses to review
feedback. Is this a feature that serves the needs of users?

Thanks