Google Reader clone?

[Dave Winer]

Any interest in developing a Google Reader clone?

I have all the back-end tech needed, and more.

Need good Ajax user interface people to work with.

Dave

PS: Obviously we'd establish an open JSON format, so both ends could
be replace, just as with the River view.

[Steve Moyer]

I'm game ... Reader is written in GWT which is a great way to build an
AJAX application (and how I've been doing 90% of my paid work
lately). Do you have any feature requests other than "Google Reader
clone"? If it's an open format and we adopt JSONP, you could load the
application onto your computer (using HTML5 Off-line application
Caching - http://www.w3.org/TR/offline-webapps/#offline) and save the
server configuration locally (using HTML5 Web Storage -
http://dev.w3.org/html5/webstorage/#the-sessionstorage-attribute) and
you could actually connect the reader to more than one server. I
guess the only reason you even need the server is to bypass the
SOP ... something that may be possible with the user's permission in
the near future.

So ... are we just talking about a non-Google controlled but browser
based feed reader or do you have more features in mind?

ciao, smoyer

[Dave Winer]

Thanks for taking up the offer!

Here's what I mean.

Create a piece of software that:

1. First and foremost, a Google Reader user would immediately feel comfortable with. That means it has to adopt its model of feeds.

2. It's a three-pane browser. Left pane contains lists of lists of feeds. Upper right pane contains the contents of the current selected feed. Lower right pane contains the contents of the selected feed item.

3. Navigation works as in Google Reader.

Now here's where it's different from Google Reader.

1. Does not attempt to be plug-compatible with Google Reader. In other words, the client communicates to the server through a protocol that is developed openly and then frozen. There is no vendor or standards body that can break apps. Very key point.

2. Lots of implementations. Open source, proprietary, running on all OSes, whatever you like.

3. I get to put it on my own server, that I pay for.

The way to start this project, imho, is for someone to build a very simple three-pane JavaScript browser with mocked up JSONP data. Then we start plugging back-ends into it.

Dave

[Steve Moyer]

Okay ... I just dumped my Google Reader subscription list to an OPML
file (thanks Google) for some raw data ... first I'll build a quick
mock-up so we can agree on the format of the page and then I'll dive
right into some code. Maybe I've got Google Reader configured
differently than you do, but I only have two panes ... the left
contains a tree with my feeds and the right contains a continuously
scrolling set of feed items. I'm using the "expanded" view. With the
"list" view, there's still only one right pane but only the item's
titles are shown and if I click on an item, only that one is expanded
(and any previously expanded item is contracted).

What I'm envisioning from your description is how my Outlook e-mail
client was configured (back when I was in the corporate world).
Folders were on the left, when I selected a folder (feed) the folder's
contents would appear in the right top pane and if I selected one of
the e-mails (usually showing sender, subject and date) it would show
me the e-mail body in the lower right panel. I this what you have in
mind? Is there a setting in Google Reader that rearranges the UI so
it's three panels? If there is, I couldn't find it!

I'm also wondering why the same reader couldn't handle both feeds and
rivers ... wouldn't it be nice to pick a river and see it continuously
scroll by too?

And I have one final question that I'll ask out of curiosity ... I'm
seeing how this all fits together, but do you have a name in mind for
the whole system? Or official names, etc. for any of the individual
parts? It's much easier to keep a consistent lexicon and branding if
you start with it early. Anyway ... keep up the good thinking!

smoyer

> > Caching -http://www.w3.org/TR/offline-webapps/#offline) and save the

[Paul Clark]

Steve,

 

If it's helpful I wrote a page that will process all the feeds in your OPML file, retrieve each RSS feed and convert the items to river format JSONP. I used the tool to generate JSONP for testing my river UI.

 

http://river.havagan.com/JsonGenerator.aspx

Paul

[Dave Winer]

Wow!

I'm almost feeling like I could just kick back and watch this happen.

Keep going guys..

Dave

[Dave Winer]

Shows how little I use Google Reader. :-)

When in doubt just do it the way GR does it.

Dave

[Dave Winer]

I don't have a name in mind yet, but I wasn't expecting it to start so quickly.

Let me give it some thought. I understand the need for it. You need to name your source files something, and if the project had a name, that's what you would use. Right?

Dave

[Steve Moyer]

Okay ... if we default to "do it the way Google Reader does it" then I
won' t bother putting together a mock-up. It's more fun to just start
coding anyways! ;) Since I'll have to put together some JSOs to
represent the different types of data, I'll do so with an eye towards
providing a straw-man of the open interface you discussed.

[Steve Moyer]

Very nice ... I didn't connect anything to the JSONP callback, but I
dumped a bunch of JSON I can use for prototyping.

Thanks

[Stephen Hogan]

I would like to join in on this endeavor!