Re: Update on Shindig's java OpenSocial RESTful support

[Chris Chabot <ch...@xs4all.nl>]

Hey Guys,

We had a bit of time pass since these emails, hows the weather on the  
RESTful spec compliance side now? I kind of started laying the  
groundwork today now that i caught up with the email & patches backlog  
after the google I/O and my fingers are itching to get started on  
this :)

	-- Chris

On May 19, 2008, at 10:16 PM, Cassie wrote:

> So the format right now... isn't right... so I wouldn't try to start
> coding it on the php side. I am going to try to write some detailed
> java tests in the next couple of days that will match the restful spec
> exactly (for gets, not posts nor deletes yet). As soon as all of those
> tests pass then the js will be good to go spec wise.
>
> I don't anticipate it taking too long to clean up the json format, but
> we are definitely not compliant yet so its good that you asked :)
> I'll ping when the js is good to go.
>
> - Cassie
>
>
> On Mon, May 19, 2008 at 7:52 PM, Chris Chabot <ch...@xs4all.nl>  
> wrote:
>> Well that goes without saying (or so i would have hoped), but you  
>> have to
>> admit in general it is easier to develop when you know what your  
>> code is
>> linked against (javascript libs in this case), is supposed to be  
>> working and
>> fully correct to spec :)
>>
>> I have no idea how far the assumptions in the js code are from the  
>> spec,
>> maybe not at all or not in a way that it would influence a  
>> implementation,
>> thats why i was asking :)
>>
>>       -- Chris
>>
>> On May 19, 2008, at 7:38 PM, Kevin Brown wrote:
>>
>>>
>>> I think a better goal would be "all versions exactly match the  
>>> spec".
>>>
>>
>>

Re: Update on Shindig's java OpenSocial RESTful support

[Chris Chabot <ch...@xs4all.nl>]

We will hopefully start on this any-day-now, as soon as i get the  
basic REST support done :)

	-- Chris

On Jun 9, 2008, at 10:36 AM, Kevin Brown wrote:

> - OAuth service provider (for RESTful) -- Dave Primmer and others are
> actively working on this for Java; I'm not sure if Chris or anyone  
> else is
> working on it for PHP.

Re: Update on Shindig's java OpenSocial RESTful support

[Kevin Brown <et...@google.com>]

On Mon, Jun 9, 2008 at 1:27 AM, Rohit Ghatol <ro...@google.com> wrote:

> Hi,
> I wanted to know whats the plan for oauth, any idea when we can see this
> coming in Shindig. Also, where can I read more about social tokens?


There are two sides of OAuth in Shindig:

- OAuth consumers (makeRequest) -- this is already implemented.
- OAuth service provider (for RESTful) -- Dave Primmer and others are
actively working on this for Java; I'm not sure if Chris or anyone else is
working on it for PHP.

The security token is simply an efficient and secure way to pass credentials
to a gadget; it's implemented in the SecurityToken class. It's a cookie
equivalent, storing some encrypted key-value pairs for authentication.


>
> Regards,
> Rohit
>
> On Fri, Jun 6, 2008 at 6:25 AM, David Primmer <da...@gmail.com>
> wrote:
>
> > Hi Chris,
> >
> > Do you mean how compliant with the spec the java social-api server is?
> > That's a tough one to answer, since it turns out that the spec itself
> > is not really very specific on a number of issues. See my recent posts
> > to the spec list. We've found it pretty tricky getting the Atom format
> > correct with all the 'hoisting' rules. There's a lot of little things
> > to do. Json is much easier. The java server's json format is better
> > than the atom format, it is pretty much read-only right now. Doesn't
> > do oauth, but does check social tokens. It doesn't do any of the
> > optional stuff in the spec and it uses large end-to-end tests instead
> > of unit tests but those give it decent coverage.
> >
> > I think you'll find that you need to invent a lot of stuff to get a
> > system that works like a real social network. Shindig, as you know,
> > does not come with one. But maybe if you have a real container it's
> > not that big of a deal. I've found it hard to code without a real
> > fully modeled container.
> >
> > davep
> >
> > On Thu, Jun 5, 2008 at 1:35 AM, Chris Chabot <ch...@xs4all.nl> wrote:
> > > Hey Guys,
> > >
> > > We had a bit of time pass since these emails, hows the weather on the
> > > RESTful spec compliance side now? I kind of started laying the
> groundwork
> > > today now that i caught up with the email & patches backlog after the
> > google
> > > I/O and my fingers are itching to get started on this :)
> > >
> > >        -- Chris
> > >
> > > On May 19, 2008, at 10:16 PM, Cassie wrote:
> > >
> > >> So the format right now... isn't right... so I wouldn't try to start
> > >> coding it on the php side. I am going to try to write some detailed
> > >> java tests in the next couple of days that will match the restful spec
> > >> exactly (for gets, not posts nor deletes yet). As soon as all of those
> > >> tests pass then the js will be good to go spec wise.
> > >>
> > >> I don't anticipate it taking too long to clean up the json format, but
> > >> we are definitely not compliant yet so its good that you asked :)
> > >> I'll ping when the js is good to go.
> > >>
> > >> - Cassie
> > >>
> > >>
> > >> On Mon, May 19, 2008 at 7:52 PM, Chris Chabot <ch...@xs4all.nl>
> > wrote:
> > >>>
> > >>> Well that goes without saying (or so i would have hoped), but you
> have
> > to
> > >>> admit in general it is easier to develop when you know what your code
> > is
> > >>> linked against (javascript libs in this case), is supposed to be
> > working
> > >>> and
> > >>> fully correct to spec :)
> > >>>
> > >>> I have no idea how far the assumptions in the js code are from the
> > spec,
> > >>> maybe not at all or not in a way that it would influence a
> > >>> implementation,
> > >>> thats why i was asking :)
> > >>>
> > >>>      -- Chris
> > >>>
> > >>> On May 19, 2008, at 7:38 PM, Kevin Brown wrote:
> > >>>
> > >>>>
> > >>>> I think a better goal would be "all versions exactly match the
> spec".
> > >>>>
> > >>>
> > >>>
> > >
> > >
> >
>

Re: Update on Shindig's java OpenSocial RESTful support

[Rohit Ghatol <ro...@google.com>]

Hi,
I wanted to know whats the plan for oauth, any idea when we can see this
coming in Shindig. Also, where can I read more about social tokens?

Regards,
Rohit

On Fri, Jun 6, 2008 at 6:25 AM, David Primmer <da...@gmail.com>
wrote:

> Hi Chris,
>
> Do you mean how compliant with the spec the java social-api server is?
> That's a tough one to answer, since it turns out that the spec itself
> is not really very specific on a number of issues. See my recent posts
> to the spec list. We've found it pretty tricky getting the Atom format
> correct with all the 'hoisting' rules. There's a lot of little things
> to do. Json is much easier. The java server's json format is better
> than the atom format, it is pretty much read-only right now. Doesn't
> do oauth, but does check social tokens. It doesn't do any of the
> optional stuff in the spec and it uses large end-to-end tests instead
> of unit tests but those give it decent coverage.
>
> I think you'll find that you need to invent a lot of stuff to get a
> system that works like a real social network. Shindig, as you know,
> does not come with one. But maybe if you have a real container it's
> not that big of a deal. I've found it hard to code without a real
> fully modeled container.
>
> davep
>
> On Thu, Jun 5, 2008 at 1:35 AM, Chris Chabot <ch...@xs4all.nl> wrote:
> > Hey Guys,
> >
> > We had a bit of time pass since these emails, hows the weather on the
> > RESTful spec compliance side now? I kind of started laying the groundwork
> > today now that i caught up with the email & patches backlog after the
> google
> > I/O and my fingers are itching to get started on this :)
> >
> >        -- Chris
> >
> > On May 19, 2008, at 10:16 PM, Cassie wrote:
> >
> >> So the format right now... isn't right... so I wouldn't try to start
> >> coding it on the php side. I am going to try to write some detailed
> >> java tests in the next couple of days that will match the restful spec
> >> exactly (for gets, not posts nor deletes yet). As soon as all of those
> >> tests pass then the js will be good to go spec wise.
> >>
> >> I don't anticipate it taking too long to clean up the json format, but
> >> we are definitely not compliant yet so its good that you asked :)
> >> I'll ping when the js is good to go.
> >>
> >> - Cassie
> >>
> >>
> >> On Mon, May 19, 2008 at 7:52 PM, Chris Chabot <ch...@xs4all.nl>
> wrote:
> >>>
> >>> Well that goes without saying (or so i would have hoped), but you have
> to
> >>> admit in general it is easier to develop when you know what your code
> is
> >>> linked against (javascript libs in this case), is supposed to be
> working
> >>> and
> >>> fully correct to spec :)
> >>>
> >>> I have no idea how far the assumptions in the js code are from the
> spec,
> >>> maybe not at all or not in a way that it would influence a
> >>> implementation,
> >>> thats why i was asking :)
> >>>
> >>>      -- Chris
> >>>
> >>> On May 19, 2008, at 7:38 PM, Kevin Brown wrote:
> >>>
> >>>>
> >>>> I think a better goal would be "all versions exactly match the spec".
> >>>>
> >>>
> >>>
> >
> >
>

Re: Update on Shindig's java OpenSocial RESTful support

[David Primmer <da...@gmail.com>]

Hi Chris,

Do you mean how compliant with the spec the java social-api server is?
That's a tough one to answer, since it turns out that the spec itself
is not really very specific on a number of issues. See my recent posts
to the spec list. We've found it pretty tricky getting the Atom format
correct with all the 'hoisting' rules. There's a lot of little things
to do. Json is much easier. The java server's json format is better
than the atom format, it is pretty much read-only right now. Doesn't
do oauth, but does check social tokens. It doesn't do any of the
optional stuff in the spec and it uses large end-to-end tests instead
of unit tests but those give it decent coverage.

I think you'll find that you need to invent a lot of stuff to get a
system that works like a real social network. Shindig, as you know,
does not come with one. But maybe if you have a real container it's
not that big of a deal. I've found it hard to code without a real
fully modeled container.

davep

On Thu, Jun 5, 2008 at 1:35 AM, Chris Chabot <ch...@xs4all.nl> wrote:
> Hey Guys,
>
> We had a bit of time pass since these emails, hows the weather on the
> RESTful spec compliance side now? I kind of started laying the groundwork
> today now that i caught up with the email & patches backlog after the google
> I/O and my fingers are itching to get started on this :)
>
>        -- Chris
>
> On May 19, 2008, at 10:16 PM, Cassie wrote:
>
>> So the format right now... isn't right... so I wouldn't try to start
>> coding it on the php side. I am going to try to write some detailed
>> java tests in the next couple of days that will match the restful spec
>> exactly (for gets, not posts nor deletes yet). As soon as all of those
>> tests pass then the js will be good to go spec wise.
>>
>> I don't anticipate it taking too long to clean up the json format, but
>> we are definitely not compliant yet so its good that you asked :)
>> I'll ping when the js is good to go.
>>
>> - Cassie
>>
>>
>> On Mon, May 19, 2008 at 7:52 PM, Chris Chabot <ch...@xs4all.nl> wrote:
>>>
>>> Well that goes without saying (or so i would have hoped), but you have to
>>> admit in general it is easier to develop when you know what your code is
>>> linked against (javascript libs in this case), is supposed to be working
>>> and
>>> fully correct to spec :)
>>>
>>> I have no idea how far the assumptions in the js code are from the spec,
>>> maybe not at all or not in a way that it would influence a
>>> implementation,
>>> thats why i was asking :)
>>>
>>>      -- Chris
>>>
>>> On May 19, 2008, at 7:38 PM, Kevin Brown wrote:
>>>
>>>>
>>>> I think a better goal would be "all versions exactly match the spec".
>>>>
>>>
>>>
>
>