Show HN: ButterflyNet – Fully async networking framework for Python 3.4

Posted by SunDwarf on Sunday November 22, 2015 @10:05AM
from the show-hn dept.

https://github.com/SunDwarf/ButterflyNet

( 10 comments | 25 points | on HN )

Re:Show HN: ButterflyNet – Fully async networking framework ... by softwarelimits on Monday November 23, 2015 @07:59AM (#10614304)

Hehe, it would be very interesting to read a quick analysis about why you think that "Twisted is a clusterfuck".

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by omginternets on Monday November 23, 2015 @10:27AM (#10615074)

Hyperbolic though that statement may be, Twisted is complex. Even after the initial learning-curve, there's quite a bit to juggle...

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by SunDwarf on Monday November 23, 2015 @11:23AM (#10615421)

I guess that's a bit too harsh and overstating. I'll change it, I just never really got Twisted, it always seemed to try and do too much for me.

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by optimuspaul on Monday November 23, 2015 @09:23AM (#10614696)

yeah, that seems hyperbolic. Twisted can have a steep learning curve, but it is proven and pretty solid. Not sure why one would think it a clusterfuck.

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by rspeer on Monday November 23, 2015 @02:10PM (#10616525)

I think that Twisted is too large and complex, and I don't approve of its "tutorials only" approach to documentation.

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by asdfaoeu on Monday November 23, 2015 @08:39AM (#10614475)

This seems like a very thin wrapper around asyncio and ThreadPoolExecutor.

Also I'd recommend dropping the AGPL license.

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by xxpor on Monday November 23, 2015 @11:12AM (#10615359)

+1 on dropping AGPL. Most people won't go near that with a 100 ft pole.

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by SunDwarf on Monday November 23, 2015 @11:25AM (#10615432)

It is a little barebones for a v1 release, but IMO I think it's better than working with the normal asyncio high-level functions.

What's wrong with the AGPL? I just thought it'd be the best fit for something like this.

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by e1ven on Monday November 23, 2015 @01:39PM (#10616309)

The project you're working on is essentially a library - It's linked to by whatever software another developer might want to use.

I'm certainly not a lawyer, but my understanding is that the AGPL an version of the GPL, with a clause that requires developers to share their code if they host the service, not just if they distribute the code.

Since the AGPL is based on the GPL, rather than LGPL, it would mean that any projects that used your library would count as Derived Works - They'd need to release their whole project if they used your code.

Typically, one might see the AGPL used for services, rather than component libraries. This has caused some companies, such as Google, to ban using any AGPL code.

See: http://www.theregister.co.uk/2011/03/31/google_on_open_sourc...

[ Parent | on HN ]

Re:Show HN: ButterflyNet – Fully async networking framework ... by asdfaoeu on Monday November 23, 2015 @06:01PM (#10617979)

AGPL is designed for software like yours, however because of the legal hassle most companies will just avoid your software. As a comparison your competition both tornado and twisted use very permissive licenses without any issue.

[ Parent | on HN ]