Re: Project idea

HTTP2 and SPDY4 should be similar in most respects (common framing, etc),
with SPDY4 adding some features for which we really need data before it
makes sense adding to HTTP2.
Implementing either would be a good choice :)
-=R


On Mon, Mar 25, 2013 at 7:56 AM, Gábor Molnár <gabor.molnar@sch.bme.hu>wrote:

> Thank you for the answers!
>
> I'm glad to hear that the spec will settle by the early summer!
>
> As Ilya mentioned, there's a spdy module for node. Do you think that it
> would worth the effort to implement a separate http2 module? Or should I
> expect HTTP2 to be very close to SPDY? I know that it is based on SPDY, but
> have no idea how far it will diverge from SPDY.
>
> Regards,
>   Gábor
>
>
> 2013/3/14 Ilya Grigorik <ilya@igvita.com>
>
>> Hi Gabor.
>>
>> Just in case, make sure you check out:
>> https://github.com/indutny/node-spdy- works very well.
>>
>> ig
>>
>> On Thu, Mar 14, 2013 at 5:51 AM, Mark Nottingham <mnot@mnot.net> wrote:
>>
>>> Hi Gábor,
>>>
>>> Our current plan is to have something suitable for implementation by
>>> early summer, so this is feasible.
>>>
>>> Of course, that's not a guarantee; we'll need to choose a header
>>> compression algorithm by then. However, if you architect well, this
>>> shouldn't be a showstopper, since you can plug new compression algorithms
>>> in later.
>>>
>>> If you decide to proceed, please do keep us informed; I'm sure people
>>> here would be interested and willing to provide feedback.
>>>
>>> Regards,
>>>
>>>
>>>
>>> On 14/03/2013, at 7:58 AM, Gábor Molnár <gabor.molnar@sch.bme.hu> wrote:
>>>
>>> > Hello all,
>>> >
>>> > I'd like to hear your opinion about a project that would result in an
>>> early prototype implementation of HTTP2 (as a node.js module) by the end of
>>> the summer. I am a student, and this is a project idea for the 2013 Google
>>> Summer of Code, but I'd like to do some preliminary research about the
>>> feasibility of the idea before contacting any mentoring organization with
>>> it.
>>> >
>>> > The project would look very similar to Tatsuhiro Tsujikawa's great
>>> sdpylay suite (it would include a simple server as well as a command line
>>> client and debugging tools besides the core library) except that it would
>>> be written in JavaScript for node.js (that would make prototyping much
>>> easier compared to  C or C++). The emphasis would be on easy-to-understand
>>> and well documented code instead of performance.
>>> >
>>> > I know that the current state of the standard is not suitable for
>>> implementation yet. Do you think that it will mature fast enough to start
>>> implementing an early prototype it in the summer?
>>> >
>>> > Please share your thoughts about the idea.
>>> >
>>> > Thanks,
>>> >    Gábor Molnár
>>> >
>>> >
>>>
>>> --
>>> Mark Nottingham   http://www.mnot.net/
>>>
>>>
>>>
>>>
>>>
>>
>