[boost] Fwd: [beast] (In)formal review

Subject: [boost] Fwd: [beast] (In)formal review
From: Michael Caisse (mcaisse-lists_at_[hidden])
Date: 2017-07-11 08:29:00

Forwarding to the list:

Received 1:19 AM my time... still July 10th in Hawaii ... and I'm still
up reading Beast reviews so it seems good enough for me.

Thanks for the review!

-------- Forwarded Message --------
Subject: 	[boost][beast] (In)formal review
From: 	Rüdiger Berlich

-------------------------------------------

Dear Michael, dear Vinnie,

having read the very lively discussion during the review, I am not sure
I can add much to the technical side. Please feel free to either count
this voice or leave it out, as you see fit, particularly as this mail
didn’t come in time for the formal review. Hence I am sending this as PM.

In my own applications, I limit myself to Boost as an external
dependency. Having Websocket-functionality as part of Boost (the „pure“
HTML-side is of less significance to me) would be a god-send and would
save me countless hours, allowing me to focus more on what I’m good at.
Currently I just build directly on ASIO.

I feel that Beast fills a very big niche and, having followed the
development of the library for a while now, I am also confident that it
is sufficiently advanced to be included into Boost. 

Thus, from a mere user’s perspective, I’d like to strongly vote in favor
of ACCEPT. 

==============

 - What is your evaluation of the design?

I have worked with ASIO for a while and appreciate that Beast follows
the same principles. 

 - What is your evaluation of the implementation?

I have browsed though some of the code and feel that it is well
structured and certainly easier to read than some of the other
Boost-code I have come across

 - What is your evaluation of the documentation?

Beast does not intend to be a high-level library, and for the intended
audience, the documentation is well-designed and complete. I would hope
that, if and when Beast becomes part of Boost, there’ll be more
tutorial-like information similar to what Boris Schaeling does in his
books (e.g. here: https://theboostcpplibraries.com — could well become
part of that).

 - What is your evaluation of the potential usefulness of the library?

Immense. C++ is lacking a LOT of functionality on the library side, and
this is one of the essential building blocks needed by people intending
to create the higher-level libraries. I believe that we will see many
new library proposals on top of Beast.

 - Did you try to use the library? With which compiler(s)? Did you
   have any problems?

I have followed the development of Beast for a number of months and have
„played“ with it in different contexts. I am in the process of
integrating it into my own (client-server type application). In
comparison to other libraries, it seems to be much more vivid and
fast-paced. I have used it on MacOS and Linux (—> clang and g++).

 - How much effort did you put into your evaluation? A glance? A quick
   reading? In-depth study?

No in-depth study, but in summary there is certainly more than a day of
involvement.

 - Are you knowledgeable about the problem domain?

Having done quite a bit of „do it yourself“ on top of ASIO for my own
application, I appreciate what Beast can do for me (on the Websocket
side) and am in dire need of this functionality. 

As a library-USER, I actually do not intend to be an expert for the
implementation side and would be very thankful if it became available as
part of Boost.

I’d like to thank Vinnie for the huge amount of work that went into
Beast and the Boost-community for the superb collection of libraries
that have kept C++ alive.

Kind Regards,
Beet

P.S.: I do not like the name „Beast“, but do not think it is important.
Certainly this is up to the author to decide :-)