You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@abdera.apache.org by David Primmer <da...@gmail.com> on 2008/01/11 02:05:21 UTC
Announcing Google FeedServer
Hello,
I'd like to announce a little project that we've been working on for a while
at Google and would like the feedback of the Abdera community. We're really
excited about the potential of Abdera and have chosen it as the foundation
for an open-source Atom Publishing Protocol Server. We hope that the
approach we've taken is helpful to those who would like to create read/write
Atom feeds for existing data sources.
There are more details below, but I just wanted to give a quick update on
where we'd like to go from here. We have lots of plans for the FeedServer,
and would like to contribute back to the Abdera project. The FeedServer was
conceived as a separate product and could possibly be an Abdera sub-project
when it graduates from incubator. We are also considering how we will use it
for the Data API portion of the Shindig project. For now, we have a project
on Google's Codesite that should give people the opportunity to look it over
and offer feedback:
http://code.google.com/p/google-feedserver/
We think we have some good docs up on the Codesite wiki. To dive in and see
what we've done, we've created an intro to the source code for developers
http://code.google.com/p/google-feedserver/wiki/SourceCodeWalkthrough
Feel free to discuss here or join the Google
group<http://groups.google.com/group/google-feedserver>for the
FeedServer.
Special thanks to Jeff Pitman, Jun Yang, Vasu Nori, Chris Anderson, Sanjay
Baberwal, Jeff Scudder, Brandon Nutter James Snell, DeWitt Clinton and Mark
Stahl.
davep
Re: Announcing Google FeedServer
Posted by Vasu Nori <va...@gmail.com>.
On Jan 11, 2008 10:14 AM, David Primmer <da...@gmail.com> wrote:
> Dan, we had long discussions about the CollectionProvidor approach and we
> do
> think it's a good approach. We wanted to get what we had out (some written
> before we saw your code) and then discuss the differences. We anticipate
> using both.We wanted a system that didn't require the dev to re implement
> all the basic atom crud for each collection and only specify the data
> translation rules.
This is referring to the Adapter approach.
one of the goals of google-feedserver was to separate the 2 concerns: atom
protocol implementation and data operations (CRUD) on the datasources.
all classes in com.google.feedserver.server package implement logic to
handle atom protocol - of course, brains behind this whole code is Abdera.
all classes in com.google.feedserver.adapter package implement the logic to
do CRUD on data from the datasources. Hopefully, this part is something
other Abdera developers/users can utilize..
vasu
Re: Announcing Google FeedServer
Posted by David Primmer <da...@gmail.com>.
Dan, we had long discussions about the CollectionProvidor approach and we do
think it's a good approach. We wanted to get what we had out (some written
before we saw your code) and then discuss the differences. We anticipate
using both.We wanted a system that didn't require the dev to re implement
all the basic atom crud for each collection and only specify the data
translation rules.
WorkspaceProvidor is an example of an area that we didn't work in because we
anticipated using your stuff. So this is just a start of our Abdera work.
It's not finished or finalized or anything.
davep
On Jan 11, 2008 9:59 AM, Dan Diephouse <da...@mulesource.com> wrote:
> Hiya,
>
> Were there any thoughts on why the CollectionProvider/ServiceProvider
> code was not a suitable approach? Did you look at possibly bringing the
> two approaches together?
>
> - Dan
>
> James M Snell wrote:
> > Hello Dave,
> >
> > Very happy to see this. After quickly reading through the docs and
> > browsing the source, there looks to be some very valuable function
> > provided, much of which I think could easily become part of Abdera.
> >
> > The Adapter mechanism looks particularly good but I would like to see
> > if we could get some closer alignment between this and the work that
> > Dan Diephouse has been doing, and the new StreamWriter, StreamBuilder
> > and (poorly-named) EntityProvider mechanisms. For performance
> > reasons, it would really be ideal if Adapter implementations could use
> > the StreamWriter interface for serializing feeds and entries.
> >
> > I do notice that the FeedServerProvider is currently rather limited in
> > function (e.g. no service document, no media post support, no feed
> > paging). I assume these functions will be added over time?
> >
> > Oh, and are you aware that Abdera already has a MethodOverrideFilter?
> :-)
> >
> > - James
> >
> > David Primmer wrote:
> >> Hello,
> >> I'd like to announce a little project that we've been working on for
> >> a while
> >> at Google and would like the feedback of the Abdera community. We're
> >> really
> >> excited about the potential of Abdera and have chosen it as the
> >> foundation
> >> for an open-source Atom Publishing Protocol Server. We hope that the
> >> approach we've taken is helpful to those who would like to create
> >> read/write
> >> Atom feeds for existing data sources.
> >>
> >> There are more details below, but I just wanted to give a quick
> >> update on
> >> where we'd like to go from here. We have lots of plans for the
> >> FeedServer,
> >> and would like to contribute back to the Abdera project. The
> >> FeedServer was
> >> conceived as a separate product and could possibly be an Abdera
> >> sub-project
> >> when it graduates from incubator. We are also considering how we will
> >> use it
> >> for the Data API portion of the Shindig project. For now, we have a
> >> project
> >> on Google's Codesite that should give people the opportunity to look
> >> it over
> >> and offer feedback:
> >>
> >> http://code.google.com/p/google-feedserver/
> >>
> >> We think we have some good docs up on the Codesite wiki. To dive in
> >> and see
> >> what we've done, we've created an intro to the source code for
> >> developers
> >> http://code.google.com/p/google-feedserver/wiki/SourceCodeWalkthrough
> >>
> >> Feel free to discuss here or join the Google
> >> group<http://groups.google.com/group/google-feedserver>for the
> >> FeedServer.
> >>
> >> Special thanks to Jeff Pitman, Jun Yang, Vasu Nori, Chris Anderson,
> >> Sanjay
> >> Baberwal, Jeff Scudder, Brandon Nutter James Snell, DeWitt Clinton
> >> and Mark
> >> Stahl.
> >>
> >> davep
> >>
>
>
> --
> Dan Diephouse
> MuleSource
> http://mulesource.com | http://netzooid.com/blog
>
>
Re: Announcing Google FeedServer
Posted by Dan Diephouse <da...@mulesource.com>.
Hiya,
Were there any thoughts on why the CollectionProvider/ServiceProvider
code was not a suitable approach? Did you look at possibly bringing the
two approaches together?
- Dan
James M Snell wrote:
> Hello Dave,
>
> Very happy to see this. After quickly reading through the docs and
> browsing the source, there looks to be some very valuable function
> provided, much of which I think could easily become part of Abdera.
>
> The Adapter mechanism looks particularly good but I would like to see
> if we could get some closer alignment between this and the work that
> Dan Diephouse has been doing, and the new StreamWriter, StreamBuilder
> and (poorly-named) EntityProvider mechanisms. For performance
> reasons, it would really be ideal if Adapter implementations could use
> the StreamWriter interface for serializing feeds and entries.
>
> I do notice that the FeedServerProvider is currently rather limited in
> function (e.g. no service document, no media post support, no feed
> paging). I assume these functions will be added over time?
>
> Oh, and are you aware that Abdera already has a MethodOverrideFilter? :-)
>
> - James
>
> David Primmer wrote:
>> Hello,
>> I'd like to announce a little project that we've been working on for
>> a while
>> at Google and would like the feedback of the Abdera community. We're
>> really
>> excited about the potential of Abdera and have chosen it as the
>> foundation
>> for an open-source Atom Publishing Protocol Server. We hope that the
>> approach we've taken is helpful to those who would like to create
>> read/write
>> Atom feeds for existing data sources.
>>
>> There are more details below, but I just wanted to give a quick
>> update on
>> where we'd like to go from here. We have lots of plans for the
>> FeedServer,
>> and would like to contribute back to the Abdera project. The
>> FeedServer was
>> conceived as a separate product and could possibly be an Abdera
>> sub-project
>> when it graduates from incubator. We are also considering how we will
>> use it
>> for the Data API portion of the Shindig project. For now, we have a
>> project
>> on Google's Codesite that should give people the opportunity to look
>> it over
>> and offer feedback:
>>
>> http://code.google.com/p/google-feedserver/
>>
>> We think we have some good docs up on the Codesite wiki. To dive in
>> and see
>> what we've done, we've created an intro to the source code for
>> developers
>> http://code.google.com/p/google-feedserver/wiki/SourceCodeWalkthrough
>>
>> Feel free to discuss here or join the Google
>> group<http://groups.google.com/group/google-feedserver>for the
>> FeedServer.
>>
>> Special thanks to Jeff Pitman, Jun Yang, Vasu Nori, Chris Anderson,
>> Sanjay
>> Baberwal, Jeff Scudder, Brandon Nutter James Snell, DeWitt Clinton
>> and Mark
>> Stahl.
>>
>> davep
>>
--
Dan Diephouse
MuleSource
http://mulesource.com | http://netzooid.com/blog
Fwd: Announcing Google FeedServer
Posted by Vasu Nori <va...@gmail.com>.
---------- Forwarded message ----------
From: Vasu Nori <vn...@google.com>
Date: Jan 11, 2008 10:08 AM
Subject: Fwd: Announcing Google FeedServer
To: vasunori@gmail.com
>
> The Adapter mechanism looks particularly good but I would like to see if
> we could get some closer alignment between this and the work that Dan
> Diephouse has been doing,
Yes, we haven't done that yet. Hope to touch base with Dan after he gets a
chance a to review the code.
> and the new StreamWriter, StreamBuilder and
> (poorly-named) EntityProvider mechanisms. For performance reasons, it
> would really be ideal if Adapter implementations could use the
> StreamWriter interface for serializing feeds and entries.
totally agree. Streaming is fantastic and we will have to implement it.
I do notice that the FeedServerProvider is currently rather limited in
> function ( e.g. no service document, no media post support, no feed
> paging). I assume these functions will be added over time?
agreed.
Regarding service document, I'd like to add a point though.
The way Google feedserver is architected allows new feeds to be added at
runtime (by simply adding a new adapter, adapter properties file & sending a
new feed URI to the feedserver). Initially, when the system comes up, there
are no known feeds to the server; i.e., service document = empty.
As each new feed URI comes in, the corresponding adapter is instantiated.
so, the service document gets built up overtime.
If at any point in time, a request for service document is received, google
feedserver can return the service document known UNTIL that point in time.
This is yet to be implemented though.
>
> Oh, and are you aware that Abdera already has a MethodOverrideFilter? :-)
>
aha. didn't know that.. :)
thanks for the comments/suggestions.
> - James
>
> David Primmer wrote:
> > Hello,
> > I'd like to announce a little project that we've been working on for a
> while
> > at Google and would like the feedback of the Abdera community. We're
> really
> > excited about the potential of Abdera and have chosen it as the
> foundation
> > for an open-source Atom Publishing Protocol Server. We hope that the
> > approach we've taken is helpful to those who would like to create
> read/write
> > Atom feeds for existing data sources.
> >
> > There are more details below, but I just wanted to give a quick update
> on
> > where we'd like to go from here. We have lots of plans for the
> FeedServer,
> > and would like to contribute back to the Abdera project. The FeedServer
> was
> > conceived as a separate product and could possibly be an Abdera
> sub-project
> > when it graduates from incubator. We are also considering how we will
> use it
> > for the Data API portion of the Shindig project. For now, we have a
> project
> > on Google's Codesite that should give people the opportunity to look it
> over
> > and offer feedback:
> >
> > http://code.google.com/p/google-feedserver/
> >
> > We think we have some good docs up on the Codesite wiki. To dive in and
> see
> > what we've done, we've created an intro to the source code for
> developers
> > http://code.google.com/p/google-feedserver/wiki/SourceCodeWalkthrough
> >
> > Feel free to discuss here or join the Google
> > group<http://groups.google.com/group/google-feedserver>for the
> > FeedServer.
> >
> > Special thanks to Jeff Pitman, Jun Yang, Vasu Nori, Chris Anderson,
> Sanjay
> > Baberwal, Jeff Scudder, Brandon Nutter James Snell, DeWitt Clinton and
> Mark
> > Stahl.
> >
> > davep
> >
>
Re: Announcing Google FeedServer
Posted by Vasu Nori <vn...@google.com>.
>
>
> The Adapter mechanism looks particularly good but I would like to see if
> we could get some closer alignment between this and the work that Dan
> Diephouse has been doing,
Yes, we haven't done that yet. Hope to touch base with Dan after he gets a
chance a to review the code.
> and the new StreamWriter, StreamBuilder and
> (poorly-named) EntityProvider mechanisms. For performance reasons, it
> would really be ideal if Adapter implementations could use the
> StreamWriter interface for serializing feeds and entries.
totally agree. Streaming is fantastic and we will have to implement it.
I do notice that the FeedServerProvider is currently rather limited in
> function (e.g. no service document, no media post support, no feed
> paging). I assume these functions will be added over time?
agreed.
Regarding service document, I'd like to add a point though.
The way Google feedserver is architected allows new feeds to be added at
runtime (by simply adding a new adapter, adapter properties file & sending a
new feed URI to the feedserver). Initially, when the system comes up, there
are no known feeds to the server; i.e., service document = empty.
As each new feed URI comes in, the corresponding adapter is instantiated.
so, the service document gets built up overtime.
If at any point in time, a request for service document is received, google
feedserver can return the service document known UNTIL that point in time.
This is yet to be implemented though.
>
> Oh, and are you aware that Abdera already has a MethodOverrideFilter? :-)
>
aha. didn't know that.. :)
thanks for the comments/suggestions.
> - James
>
> David Primmer wrote:
> > Hello,
> > I'd like to announce a little project that we've been working on for a
> while
> > at Google and would like the feedback of the Abdera community. We're
> really
> > excited about the potential of Abdera and have chosen it as the
> foundation
> > for an open-source Atom Publishing Protocol Server. We hope that the
> > approach we've taken is helpful to those who would like to create
> read/write
> > Atom feeds for existing data sources.
> >
> > There are more details below, but I just wanted to give a quick update
> on
> > where we'd like to go from here. We have lots of plans for the
> FeedServer,
> > and would like to contribute back to the Abdera project. The FeedServer
> was
> > conceived as a separate product and could possibly be an Abdera
> sub-project
> > when it graduates from incubator. We are also considering how we will
> use it
> > for the Data API portion of the Shindig project. For now, we have a
> project
> > on Google's Codesite that should give people the opportunity to look it
> over
> > and offer feedback:
> >
> > http://code.google.com/p/google-feedserver/
> >
> > We think we have some good docs up on the Codesite wiki. To dive in and
> see
> > what we've done, we've created an intro to the source code for
> developers
> > http://code.google.com/p/google-feedserver/wiki/SourceCodeWalkthrough
> >
> > Feel free to discuss here or join the Google
> > group<http://groups.google.com/group/google-feedserver>for the
> > FeedServer.
> >
> > Special thanks to Jeff Pitman, Jun Yang, Vasu Nori, Chris Anderson,
> Sanjay
> > Baberwal, Jeff Scudder, Brandon Nutter James Snell, DeWitt Clinton and
> Mark
> > Stahl.
> >
> > davep
> >
>
--
Vasu Nori
ext: 41643
650.214.1643 (direct) 925.209.0647 (c)
vnori@google.com
Re: Announcing Google FeedServer
Posted by James M Snell <ja...@gmail.com>.
Hello Dave,
Very happy to see this. After quickly reading through the docs and
browsing the source, there looks to be some very valuable function
provided, much of which I think could easily become part of Abdera.
The Adapter mechanism looks particularly good but I would like to see if
we could get some closer alignment between this and the work that Dan
Diephouse has been doing, and the new StreamWriter, StreamBuilder and
(poorly-named) EntityProvider mechanisms. For performance reasons, it
would really be ideal if Adapter implementations could use the
StreamWriter interface for serializing feeds and entries.
I do notice that the FeedServerProvider is currently rather limited in
function (e.g. no service document, no media post support, no feed
paging). I assume these functions will be added over time?
Oh, and are you aware that Abdera already has a MethodOverrideFilter? :-)
- James
David Primmer wrote:
> Hello,
> I'd like to announce a little project that we've been working on for a while
> at Google and would like the feedback of the Abdera community. We're really
> excited about the potential of Abdera and have chosen it as the foundation
> for an open-source Atom Publishing Protocol Server. We hope that the
> approach we've taken is helpful to those who would like to create read/write
> Atom feeds for existing data sources.
>
> There are more details below, but I just wanted to give a quick update on
> where we'd like to go from here. We have lots of plans for the FeedServer,
> and would like to contribute back to the Abdera project. The FeedServer was
> conceived as a separate product and could possibly be an Abdera sub-project
> when it graduates from incubator. We are also considering how we will use it
> for the Data API portion of the Shindig project. For now, we have a project
> on Google's Codesite that should give people the opportunity to look it over
> and offer feedback:
>
> http://code.google.com/p/google-feedserver/
>
> We think we have some good docs up on the Codesite wiki. To dive in and see
> what we've done, we've created an intro to the source code for developers
> http://code.google.com/p/google-feedserver/wiki/SourceCodeWalkthrough
>
> Feel free to discuss here or join the Google
> group<http://groups.google.com/group/google-feedserver>for the
> FeedServer.
>
> Special thanks to Jeff Pitman, Jun Yang, Vasu Nori, Chris Anderson, Sanjay
> Baberwal, Jeff Scudder, Brandon Nutter James Snell, DeWitt Clinton and Mark
> Stahl.
>
> davep
>