Search This Blog

Saturday 20 January 2007

Six things I'd like to see in Google Reader.

I really like Google Reader - it has allowed me to expand the number of feeds that I can read and it provides me with an interface that allows me to do that much faster and with more control than before. I seem to be reading somewhere between 1.5million and 3 million words a month via its interface - that's about the same as 15-30 novels - so it must be pretty good. However there are a couple of things that I'd really like to see added or changed.

[1] Search and Filters. Apart from the simple madness of the fact that this is a Google product for gathering data that doesn't have Search built in there is much more that I want than a simple search box. A Custom Search Engine based on your feed list OPML can be a partial solution but a built in "Search and Do something" facility is top of my list of sought after changes. I would really like to be able to search my "feedspace" and treat the result like a labeled set of items. I could then:
  • Search and read the items (obviously).
  • Search and mark the item(s) (as read, unread, with a new label, for sharing etc...).
  • Search and mark (as read\as unread\ with a label) the source feed(s).
  • Save the search as a dynamic label so that new articles would automatically get added if they met the search criteria.
  • Search and mail. Ideally used with the aforementioned dynamic labels. I could then use Reader to keep an eye out for really important stuff and send a mail to my phone when something I'm really interested in pops up.
[2] Some clarity and development on the role of Labels. Are they "Item Labels" (where they function as tags on individual articles and are applied individually to articles - starring and sharing items is biased towards this view of Labels) or "Feed Labels" (where the Labels serve as general purpose desriptors that apply a management hierarchy to your subscriptions and are automatically applied to each new item that arrives). The interface is very unclear at the moment about what the differences are between the two views, if any. I see these as separate things with some common aspects and would like to see a change that made that fact clear but I'd be happy for someone to teach me why I'm wrong.

[3] Automated detection and removal of dupes across different feeds. I see identical articles in many feeds (e.g. the Engadget PDA feed often contains items that are also in the Engadget GPS feed). Once such a dupe is read in one spot it should be marked as read everywhere even if the feed source is not identical. Clearly some care needs to be taken with this but it's the sort of incremental intelligence that Google is usually brilliant at so I have hope.

[4] Mobile interface. I use the mobile interface more than anyone else I'm aware of and I love it. Unlike others I find that the current ultra minimalist interface design is almost perfect for me for the simple browsers and cramped screens on my mobile devices but I want to see a handful of extra features.
  • Authentication. Entering a Google Account password is a major headache on a mobile device. After seeing how Microsoft Live for Mobile handles this I think Google could make things dramatically simpler and safer as follows:
    • Give your mobile number to Google Reader (they have mine already for Google Calendar alerts) and choose a short easy to enter mobile access PIN.
    • Google Reader then sends an SMS to that number containing a URL to page that only requires that PIN to authenticate you to your Google Reader Mobile page.
    • Expire the PIN after use\after a short time just to keep us all nice and safe.
  • Add the ability to add labels to either an item or a feed.
  • Add the ability to edit or add new labels and subscriptions.
  • Add a "Share This" hot key. The interface already support the following in its Item read view: 1 - See Original , 0 - Next in Your Reading List, * - Add a Star (duh!), # Keep Unread, plus links to your home page, tags, subscriptions and signing out. It's a no brainer to add 2 - Share This, and 3 - Add a Label to this item. doing so would not clutter the display.
[5] More Stats and access to the data behind those stats. I love the new stats that but I'd like some more.
  • I'd like to know what fraction of my activity had been carried out via the Mobile interface.
  • Word counts would be interesting. Feedback into the Search and Filter features at the top and you have a way to manage your feeds based on activity.
  • A live interface to the raw stats data data would enable some interesting mashups (especially if others, like Amazon, were to do the same).
[6] Feed Suggestions. The data mining potential of our reading patterns could enable some very effective suggestions (similar to the "Interesting Items for You" Google Gadget).

5 comments:

Ithiad said...

[7] I'd like Reader not to drop old unread items. So I like to mark stuff unread that I want to think about later and then don't do this thinking until months later. I know I'm a bad person. But it isn't like losing old marked unread items is anything but very poor.

[8] I love the roll pane. I'll even suffer [7] for a while for the toilet roll interface. EverNote does the same. So buy EverNote and mash it with Reader and Gmail. Then write a rollodex contact manager in place of the brain deformed thing that's there. It's all personal information management. Put it in the same interface.

[9] I browse some tags in list view and just know I won't read some of the stories. Let me mark them read directly from the line instead of having to expand the item. I don't want to read it but I still get to scan it before I can mark it read? Dumb.

[10] The option to automatically read items as you roll over them is good. Put the check box on the main page instead of deep in the settings so I can race over my newspaper feeds and then take time with my music links.

[11] Let me expand the starred items by tag. Or no better... yeah, let me search properly.

[12] Release the API. It's annoying to watch the pain on the discussion board.

[13] Automatically generate feeds for pages that don't have them. Well, I mean diff against the last google cache when the page is reindexed, tidy a little and send to an RSS feed.

Joe Mansfield said...

Good points.
[7] When do old unreads get flagged for removal? I haven't noticed it (yet) but I don't keep much unread - I clear everything out when I've read enough.

[8] As far as the contact manager comment is concerned I think you're right that they really do need to up the standard there - I have better contact managers in my Windows Smartphones. The new contact manager buried in Google Video might do the trick but my recall of that is that it was incomplete and it's been disabled since GOS put it up.

[11] I think that would be resolved by my "search and do something" suggestion but as a matter of course that should apply to any aggregated view. The confusion on the role of tags causes this IMO - if that was clear this would not be a problem.

[12] Releasing the API is a no brainer but being still stuck in Google Labs terrain probably keeps that from happening for now. I must peruse the discussion group too though.

[13] That's an interesting idea - ideal for infrequently refreshed sites too, the sort that tend to get "forgotten" but are still worthwhile.

[14] Editable CSS for the shared page is a major failing and I'm annoyed that I missed that as the locked in formatting really does bug me.

Ithiad said...

[7] I seem to lose unread items after about two weeks or so. They're still marked unread, though, and you can find them if you scroll far enough through the all items view of a feed. But they don't count in the unread items count or appear in the show only new view. The best effort at finding them is probably with this link:
http://www.google.com/reader/view/user/-/state/com.google/kept-unread
But you can't manipulate the read status of the items from there which makes it extra annoying.

[8] 'better contact managers in my Windows Smartphones' is damning.

[12] There's been some sort of unreleased API around for about a year. It's not properly documented or, I think, really useful enough.

[13] This could be done easily enough outside Reader. I can almost see it now... you want to subscribe to a McSweeney's RSS, you point yourself at DaithiIsGod.com/www.mcsweeneys.net/daily.xml
Could build the initial feed out of the current page, the google cache and the wayback machine. The properly interesting part is doing a DOMmy diff that's not stupid and removing the advertising and junk. Prehaps the person subscribing to the feed could be used to indicate the interesting content as part of the initial subscription process.

Then we stick our AdSense into the feed and watch the money roll in before we are sued.

Ithiad said...

[15] Add to calendar link like in Gmail.

Ithiad said...

[16] Be more accurate with the counts. The 100+ is nice but not helpful when you have 1000+ unread. Round counts to nearest 100 if they exceed 100.

[17] Make my Gmail atom link read properly in Reader. It should be embarrassing that it doesn't work.