Re: moving forward—with a plan

Hi Markus (and list),

  

I absolutely agree that having one or more implementations are way more
valuable than a spec. Several months ago, I posted to this list with my
attempts to implement a Ruby version. I found the easiest approach for me
would be to write a test suite that would exercise the Hydra example server
and my code, and highlight the differences. When I started digging in to the
spec itself, though, I found it had diverged from the example implementation,
which made it much harder for follow.

  

Having a reference implementation to follow is way more valuable to me when
I'm learning the intricacies of Hydra. The spec is only useful to me as a user
and an implementer when I come across edge-cases, but I need to have learned a
lot before I can reach the point where I can understand what the spec is
telling me.

  

I'm happy to contribute a Ruby client/server implementation if that's the
direction that is decided.

  

Thanks,

Paul Sadauskas

On May 25 2016, at 3:14 pm, Ruben Verborgh <ruben.verborgh@ugent.be>
wrote:  

> Hi Markus,

>

> > we have are missing commitment and contribution...

>

> +1

>

> > and, probably most importantly, lack of code that is driving and
informing the discussions.

>

> +1, that's a good observation

>

> > This made it hard to assess whether a design proposal was good or bad.

>

> Indeed!!

>

> > So, my proposal is comparatively simple: we start creating a Hydra
client (or a couple of them) and use it to drive the further development of
Hydra.

>

> Makes sense; probably good to have milestones also,  
so that we become more or less in sync.

>

> "a couple" seems important; otherwise, we don't really need a spec.

>

> > The server side is obviously important as well, but it is comparatively
simple to mock or fake so I would mainly focus on the client-side.

>

> I wouldn't disregard the server entirely;  
we can mock or fake, but real-world use cases are important.

>

> …and the survey is a great idea!

>

> Ruben