Stirring up this old thread, I chatted briefly with Steve Klabnik who just
pitched http://jsonapi.org/ at All Things Open in Raleigh. Seems like there
are similarities here with GraphQL and Falcor as well.
On Mon, Oct 12, 2015 at 2:01 PM David Robinson <[email protected]> wrote:
> My take on GraphQL is almost identical to Matt F's, even to the remarks on
> how its name is confusing.
>
> To convert a GraphQL statement like the following example into a TP3 query:
>
> {
> user(id: 3500401) {
> ID,
> name,
> isViewerFriend,
> profilePicture(size: 50) {
> uri,
> width,
> height
> }
> }}
>
> graph meta data has to be applied, but that alone might not be enough. In
> the example, ID could
> be a property on user, another vertex or even an edge label or property.
> But since there is nothing
> that forces ID to be unique, it could be all the above in a graph, making
> the query, even with
> meta data ambiguous. If something was "hard coded" for a particular
> query/graph set of scenarios
> this might work. But as Matt said, it doesn't seem good for general
> purpose graph queries.
>
> On Mon, Oct 12, 2015 at 7:35 AM, Stephen Mallette <[email protected]>
> wrote:
>
> > Daniel, it sounds like we need to quickly turn and correct any
> conversation
> > that introduces the perception that "non-individuals" are doing stuff in
> > our project (even when that is not really what the intention is). Your
> > cartoonish example, as you called it, made sense to me.
> >
> > I think HipChat is a good place to start as well. It sounded like a lot
> of
> > folks wanted to join to hear what was going on here, it seems like
> > scheduling a time when a talk could happen would be good. Jon, maybe you
> > could throw out a few days when you might be able to join the HipChat
> room
> > and we could converge on something that works.
> >
> >
> > On Oct 11, 2015 11:05 AM, "Marko Rodriguez" <[email protected]>
> wrote:
> >
> > > Hey Jon,
> > >
> > > Perhaps lets start slow and surefooted by jumping into TinkerPop's
> Apache
> > > HipChat room sometime this week. The public URL for anyone to join and
> > chat
> > > is here:
> > >
> > > https://www.hipchat.com/gz5JlgJBC
> > >
> > > This way Gruno can monitor the discussion and if he feels comfortable
> > with
> > > how things are unfolding, then we can increase the bandwidth
> accordingly.
> > >
> > > I'm in there all day 5 days a week, but if you want to propose a
> specific
> > > time, please do.
> > >
> > > Marko.
> > >
> > > http://markorodriguez.com
> > >
> > > On Oct 11, 2015, at 5:44 AM, Daniel Gruno <[email protected]>
> wrote:
> > >
> > > > On 10/11/2015 03:30 AM, Jonathan Schneider wrote:
> > > >> I hope this is simply another case of misunderstanding.
> > > >
> > > > If anything, it might be a misunderstanding of what I wrote back.
> > > > Companies can contribute to ASF projects, they can collaborate with
> > > > projects, but both they AND we must do so as individuals - checking
> > > > their company privileges, rules and hats at the door. The TinkerPop
> > > > project cannot collaborate with a company, as we simply don't do that
> > > > (ever), but they can collaborate with individuals of that company
> > acting
> > > > as individuals. My reply was as much a clarification of terms
> > > > (contributors and collaborators) as it was a guideline.
> > > >
> > > >>
> > > >> I do not mean to imply that there should be some relationship
> between
> > > >> Netflix as an organization and the Tinkerpop project, merely that
> > there
> > > are
> > > >> a number of parties that are working on different sides of this
> > problem
> > > and
> > > >> it would be beneficial to have a quick live discussion about whether
> > > >> Tinkerpop fits into this or not.
> > > >>
> > > >> The word "call" may have implied a private meeting as well, which
> was
> > > not
> > > >> at all the intent. Can we host something on Apache Hipchat,
> Hangouts,
> > > or
> > > >> whatever other medium is available that would allow anybody to join?
> > > >> Perhaps recorded?
> > > >
> > > > Again, this depends on the independence of the people you are talking
> > to
> > > > and the outcome. If this is an Apache (or MIT/BSD-style) licensed
> open
> > > > source project, then you are of course welcome to add support for it,
> > > > provided you handle the press/marketing/announcement around it in a
> > > > neutral manner, so as to not give the impression that any favors have
> > > > been played here (I'm quite well aware of the _why_ here, but public
> > > > perception is also a key element in staying neutral. Not only must we
> > > > act neutral, we must also tell it in a neutral way).
> > > >
> > > >>
> > > >> Daniel - I hope you don't mean to imply that there can be no
> > > collaboration
> > > >> between ASF projects and open source projects that happen to be
> > steered
> > > by
> > > >> private companies. That seems painfully limiting if true.
> > > >
> > > > That was not what I was implying, no. But, as stated above, you
> cannot
> > > > collaborate with a company, only the individuals in that company.
> > > >
> > > > GOOD: We just collaborated with committers from the XYZ project
> > > > BAD: We just had a conf call with $company folks, and their boss
> > says...
> > > >
> > > > The above is a bit cartoonish, but I hope it illustrates my point.
> > > >
> > > > With regards,
> > > > Daniel.
> > > >
> > > >>
> > > >> Jon
> > > >>
> > > >> On Sat, Oct 10, 2015 at 5:04 PM Daniel Gruno <[email protected]>
> > > wrote:
> > > >>
> > > >>> On 10/10/2015 07:43 PM, Marko Rodriguez wrote:
> > > >>>> Hello Jon,
> > > >>>>
> > > >>>>> Netflix has recently released a similar library called Falcor
> > > >>>>> <http://netflix.github.io/falcor/>. They are optimized for
> > slightly
> > > >>>>> different uses, but are both angling at solving the over/under
> > > fetching
> > > >>>>> problem and are more or less equivalent.
> > > >>>>
> > > >>>> Gotcha.
> > > >>>>
> > > >>>>> Netflix has a need for a solid binding between Falcor/GraphQL and
> > > >>> Tinkerpop
> > > >>>>> immediately and I'm sure several of us could help. Could we set
> up
> > > an
> > > >>>>> introductory call with the Netflix UI Engineering, Platform, and
> > > >>>>> Engineering Tools teams that are all hacking at this from
> different
> > > >>> angles?
> > > >>>>
> > > >>>> Excellent. Yes, lets have a call. However, before I commit lets
> wait
> > > for
> > > >>> an answer to:
> > > >>>>
> > > >>>> MENTORS: Is it "okay" to have a phone call (thus "off list") with
> a
> > > >>> commercial entity to talk about their use of TinkerPop and about
> > > potential
> > > >>> collaborations?
> > > >>>>
> > > >>>
> > > >>> It's not okay for the Tinkerpop project, and it's not okay to
> discuss
> > > >>> this on the dev list.
> > > >>> If you, as an individual, want to make that phone call, fine, but
> as
> > a
> > > >>> project, Tinkerpop must be completely vendor neutral.
> > > >>>
> > > >>> Companies cannot - I repeat cannot - collaborate with a project.
> They
> > > >>> can sponsor items (without their branding on), and they can pay
> devs
> > to
> > > >>> program as *independent individuals* (provided they gain the merit
> > for
> > > >>> this), but companies themselves cannot contribute to an ASF project
> > and
> > > >>> likewise, the ASF cannot contribute to or collaborate with
> companies.
> > > >>>
> > > >>> With regards,
> > > >>> Daniel.
> > > >>>
> > > >>
> > > >
> > >
> > >
> >
>
