Beta 2, Beta 3 / Shindig 1.1 / OpenSocial 0.9 & 1.0(???)

[Mark D Weitzel <we...@us.ibm.com>]

First of all a big thanks to everyone that's been driving the Beta 1, 2, 
and 3, effort. From our standpoint, this is a big help. One of the things 
I'm trying to get a picture of is the overall road map so that I can help 
advise the right set of folks internally. To this end, I've got a couple 
of questions that hopefully make sense....

Now that we've got Beta 2 wrapping up, do we have an idea of what the time 
frame is for Beta 3 and ultimately Shindig 1.1?
Does it make sense to try and use this recent set of work to establish a 
regular interval for releases? This would be a big help for planning 
purposes where it takes a while to integrate/test drivers with other code.
>From reading earlier posts, Shindig 1.1 will be an implementation of 0.9. 
I haven't seen too many new requests for 0.9 Next on the OpenSocial lists, 
so is it worth figuring out what we need to do to stabilize/clean up the 
spec and get it to a 1.0 level and have Shindig 1.1 match that? If the 
spec doesn't change too much from 0.9, then could we reasonably manage 
this? (Please forgive me if I'm rehashing long heated discussions--I 
apologize that I've not chimed in much before this but I've been a bit out 
of pocket.)

Thanks for thinking about this. Again--apologies if I'm whacking a 
hornet's nest.

-Mark W.

Re: Beta 2, Beta 3 / Shindig 1.1 / OpenSocial 0.9 & 1.0(???)

[Mark D Weitzel <we...@us.ibm.com>]

Kevin,

Thanks for the response. Given that we are thinking we'll have a few 
releases a year of Shindig, do we have a time line for when we'd like to 
get 1.1 released? Is the thought that we'd have few set release dates, 
e.g. March & Aug? 

The reason I'm asking is so that I can work with our product teams to 
figure out what the right version of Shindig is for us to use. We have 
internal processes we need to following when working with code that 
originates in open source and have to account for that in our development 
cycles. Almost always, we are required to use an officially released 
version of the code. Right now, that is 1.0. However, there are some 
capabilities in the current beta streams that we'd like to be able to 
leverage. If we knew when 1.1 was scheduled to be released, we could plan 
accordingly.

Thanks again for all your help.

-Mark W.





From:
Kevin Brown <et...@google.com>
To:
shindig-dev@incubator.apache.org
Date:
08/13/2009 01:54 AM
Subject:
Re: Beta 2, Beta 3 / Shindig 1.1 / OpenSocial 0.9 & 1.0(???)



On Wed, Aug 12, 2009 at 4:07 PM, Mark D Weitzel <we...@us.ibm.com> 
wrote:

> First of all a big thanks to everyone that's been driving the Beta 1, 2,
> and 3, effort. From our standpoint, this is a big help. One of the 
things
> I'm trying to get a picture of is the overall road map so that I can 
help
> advise the right set of folks internally. To this end, I've got a couple
> of questions that hopefully make sense....
>
> Now that we've got Beta 2 wrapping up, do we have an idea of what the 
time
> frame is for Beta 3 and ultimately Shindig 1.1?
> Does it make sense to try and use this recent set of work to establish a
> regular interval for releases? This would be a big help for planning
> purposes where it takes a while to integrate/test drivers with other 
code.
> From reading earlier posts, Shindig 1.1 will be an implementation of 
0.9.
> I haven't seen too many new requests for 0.9 Next on the OpenSocial 
lists,
> so is it worth figuring out what we need to do to stabilize/clean up the
> spec and get it to a 1.0 level and have Shindig 1.1 match that? If the
> spec doesn't change too much from 0.9, then could we reasonably manage
> this? (Please forgive me if I'm rehashing long heated discussions--I
> apologize that I've not chimed in much before this but I've been a bit 
out
> of pocket.)


There are literally dozens of people involved with the specification that
don't follow what's going on with Shindig, and it is not Shindig's role to
define the opensocial specification. That is a job for the opensocial
foundation.

Shindig's releases aren't intended to be tied to any specific version of 
the
spec. We may release many major updates between spec revisions, and we may
wind up with one release that happens to work for multiple spec versions.
It's much the way that web browsers don't necessarily match HTML versions.

Realistically the gap between OS spec updates will become very long in the
near future, most likely measured in terms of years. Shindig will still 
have
releases at least a few times a year, though.

Shindig 1.1 will most likely also just be a 0.9 compliant release, as will
1.2, 1.3, and probably a few versions after that. Shindig's version is 
based
on our overall architectural choices, not on specification compatibility.
That is to say, versions 1.0 through 1.x will all have the same basic
architecture (if not full backwards compatibility), but going from 1.x to
2.x will likely mean that we made some massive architectural changes.

Obviously major changes to the specification can possibly result in major
changes to Shindig, but this is something that will always be addressed on 
a
case by case basis. If we can reasonably implement new versions of the
specification without needing to make big changes to Shindig, we'll do 
that,
and if we must make drastic changes then we will update our major version
appropriately.


>
>
> Thanks for thinking about this. Again--apologies if I'm whacking a
> hornet's nest.
>
> -Mark W.
>
>

Re: Beta 2, Beta 3 / Shindig 1.1 / OpenSocial 0.9 & 1.0(???)

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

On Wed, Aug 12, 2009 at 4:07 PM, Mark D Weitzel <we...@us.ibm.com> wrote:

> First of all a big thanks to everyone that's been driving the Beta 1, 2,
> and 3, effort. From our standpoint, this is a big help. One of the things
> I'm trying to get a picture of is the overall road map so that I can help
> advise the right set of folks internally. To this end, I've got a couple
> of questions that hopefully make sense....
>
> Now that we've got Beta 2 wrapping up, do we have an idea of what the time
> frame is for Beta 3 and ultimately Shindig 1.1?
> Does it make sense to try and use this recent set of work to establish a
> regular interval for releases? This would be a big help for planning
> purposes where it takes a while to integrate/test drivers with other code.
> From reading earlier posts, Shindig 1.1 will be an implementation of 0.9.
> I haven't seen too many new requests for 0.9 Next on the OpenSocial lists,
> so is it worth figuring out what we need to do to stabilize/clean up the
> spec and get it to a 1.0 level and have Shindig 1.1 match that? If the
> spec doesn't change too much from 0.9, then could we reasonably manage
> this? (Please forgive me if I'm rehashing long heated discussions--I
> apologize that I've not chimed in much before this but I've been a bit out
> of pocket.)


There are literally dozens of people involved with the specification that
don't follow what's going on with Shindig, and it is not Shindig's role to
define the opensocial specification. That is a job for the opensocial
foundation.

Shindig's releases aren't intended to be tied to any specific version of the
spec. We may release many major updates between spec revisions, and we may
wind up with one release that happens to work for multiple spec versions.
It's much the way that web browsers don't necessarily match HTML versions.

Realistically the gap between OS spec updates will become very long in the
near future, most likely measured in terms of years. Shindig will still have
releases at least a few times a year, though.

Shindig 1.1 will most likely also just be a 0.9 compliant release, as will
1.2, 1.3, and probably a few versions after that. Shindig's version is based
on our overall architectural choices, not on specification compatibility.
That is to say, versions 1.0 through 1.x will all have the same basic
architecture (if not full backwards compatibility), but going from 1.x to
2.x will likely mean that we made some massive architectural changes.

Obviously major changes to the specification can possibly result in major
changes to Shindig, but this is something that will always be addressed on a
case by case basis. If we can reasonably implement new versions of the
specification without needing to make big changes to Shindig, we'll do that,
and if we must make drastic changes then we will update our major version
appropriately.


>
>
> Thanks for thinking about this. Again--apologies if I'm whacking a
> hornet's nest.
>
> -Mark W.
>
>