Re: Hydra Status

Great point. Comparing Hydra to its alternatives would help people 
understand its strengths and weaknesses. Currently I've had people ask 
me "Why do I need Hydra?". They feel like the existing technologies 
fulfill their needs. And there is a lot to choose form, Swagger/RAML, 
JSON Schema, all those hypermedia media types...

On 2016-11-10 21:45, Brian Sletten wrote:
> One of the main concrete examples I am excited about wrt Hydra and Linked Data (compared to Swagger et al) is that, being backed by a real data model, it is trivial to connect the APIs we describe with the organizations that develop them. Here is the endpoint, here’s what it does, here’s what it returns, here’s who developed it and the organization she works for and a link to the repo its in and the continuous integration tests and uptime metadata and the Jira issue that she worked from and the requirements as gathered by this person, etc. It becomes possible to mine resources and APIs that are exhibiting stability and value to the organization and the people that were involved. Some of this can be done with the established terms and models for Swagger and RAML, but not much and extending them isn’t as easy.
>
>> On Nov 10, 2016, at 12:36 PM, Graham Conzett <conzett@gmail.com> wrote:
>>
>> Regarding linked data, I've done very little semantic web/linked data
>> stuff in the past so I'm coming at this from a purely hypermedia API
>> perspective, but I think it's great that we get that "for free" by
>> using Hydra.
>> I'd love to take a stab at writing up some step by step examples
>> starting with the very basics - if for no other reason than to get
>> feedback and better understand the concepts. Right now the docs seem
>> kind of intimidating because if you're not familiar with JSON-LD it's
>> a lot to take in all at once.
>>
>> On Thu, Nov 10, 2016 at 3:33 PM, Brian Sletten <brian.sletten@gmail.com> wrote:
>>> I’ve been meaning to write to the list and pledge to get some things done. I’ve had a long, rough summer and fall so I won’t be able to do much until the end of the year but I would very much like to see all of these things move forward and plan on helping with examples, implementations, clients, etc.
>>>
>>> Not much more to offer than that now, but I just want to share my enthusiasm for moving things forward as soon as possible.
>>>
>>>> On Nov 10, 2016, at 11:51 AM, Tomasz Pluskiewicz <tomasz@t-code.pl> wrote:
>>>>
>>>> Hi Graham
>>>>
>>>> Your timing is very good. I was just about going to write a similar email just days ago but postponed it somewhat.
>>>>
>>>> You may have found the thread from I think around June where the future of Hydra was discussed. I'm not particularly happy with the outcome and the drop-off does seem to prove a point here. There were some declarations about creating various clients but little has happened on the mailing list since.
>>>>
>>>> I personally did propose a move towards a more "modern", community-driven approach. There is a #hydra Slack channel on [1] as well as quite an interest in JSON-LD and also Hydra on the #general channel. That is partially driven by myself and Asbjorn from this list (among other who are less active here - sorry if I miss someone out). I also did propose a move to GitHub issues as much as possible. I find these emails hard to follow, often chatty and unwieldy. Not to mention the pain of going back to old threads and linking between them. GitHub is much more flexible in that regard.
>>>>
>>>> Tooling aside, I think that the nail to the coffin here was more about how the issues are usually approached.
>>>>
>>>> The first problem is that oftentimes drift towards very specific and sophisticated solutions rather that something more simple that can be improved upon later. Something good enough to solve immediate needs without sacrificing the extensibility. Instead I frequently saw nitpicking about corner cases and hypothetical "intelligent agents", ideal world, etc. I'm not saying it's not an important direction to head towards but the effects are evident: Some features were being discussed for weeks or even months on end without reaching a consensus.
>>>>
>>>> Second, I think that the disconnection between the mailing list and GitHub issues/repository is just plain confusing. Not to mention there is that W3C wiki which I can never find the bloody link to. And most people probably don't even know exists. And so, the effect is that even features that reach consensus here are not implemented in the spec. And even worse, the issue is closed yet spec outdated (please post a link by the way). I'm not sure whether this point is also a tooling issue. Maybe more a leadership problem where we happily discuss stuff here and then expect it to magically appear in the specification.
>>>>
>>>> To sum up this lengthy email, unlike last time, I strongly urge us to make a move towards a more open and agile community. There is demand for Hydra, a Hydra which will solve practical problems of an average developer. There is also little competition in the API space yet we still fail to deliver. We cannot wait much longer! If we do, Hydra will eventually become DOA. Dismissed by those familiar with as too much ideals-driven and made obsolete some other technology (God forbid inferior :wink:).
>>>>
>>>> The steps I would perform would include first and foremost:
>>>>
>>>> 1. cleaning up GitHub issues
>>>> 2. bringing the spec up to date
>>>> 3. implementing crucial features which are offered by the alternatives (c'mon paging and filtering is still in flux!)
>>>>
>>>> Only then I would proceed with new and more "Semantic Web" feature we all want Hydra to be known for.
>>>>
>>>> Thanks,
>>>> Tom
>>>>
>>>> [1]: http://slack.httpapis.com
>>>>
>>>> November 10 2016 8:15 PM, "Graham Conzett" <conzett@gmail.com> wrote:
>>>>> Hello all,
>>>>>
>>>>> After evaluating a lot of different hypermedia solutions, I'm very
>>>>> interested in using Hydra! It seems to offer the right mix of power,
>>>>> simplicity and extensibility without being overwhelming and requiring
>>>>> a huge shift in the way our existing REST APIs work. However, it looks
>>>>> like there has been a significant drop-off in mailing list
>>>>> participation and the website doesn't seem to have been updated in
>>>>> quite some time. There also seem to be several GitHub discussion
>>>>> marked as "resolved" that are not reflected in the core vocab draft.
>>>>>
>>>>> I'm a little worried about going all in on something that isn't being
>>>>> actively developed anymore. That being said, is there anything we
>>>>> could help out with? Would writing up a series of examples be useful?
>>>>> Any interest in moving more of the discussion to GitHub as opposed to
>>>>> the mailing list to get more eyes on things?
>>>>>
>>>>> Thanks!
>>>>>
>>>>> - Graham
>>>>
>>>>
>>>
>
>

Received on Friday, 11 November 2016 07:54:19 UTC