Agent FeedSeeder Says ‘Hi’

If you see: FeedSeeder/Development (; TEST Furrow in your logs, that’s me. Well, technically, the T-minus project I’ve been talking about.

Hi. How are you? Things are good. Thanks.

I’m pretty happy with how things are behaving, even at this yet early stage. Just the poking around I did tonight felt very comfortable. I’m looking forward to when I can live in it for a while.

RSS 2.0 feeds are no problem – just as I suspected. Atom, RDF, and some Feedburner feeds are yet to come.

T-60: Thoughts on the Browser-based Feedreader

Chrono Cracker wrote up his wishes for a browser-based feed reader, and since I’m working on one, I thought I’d respond with my own 25 cents.

  1. Automatic Feed Finding
    Yea, I agree – typing in fewer characters rules. Still seems like needing to type a URL in the first place is a workaround.
  2. Rate the Post
    I’ve talked about this before. As time goes on, I see less and less value from it. Single-dimensional ratings don’t actually provide any meaningful data on relevance (simply because we don’t all rate on relevance). There’s an interesting idea in there though, I think only is has even made a first step in this useful direction.
  3. Tagging/Inline Search & Link & Site Genre
    Every attribute is a tag. Read, Time read, keywords, number of keywords, author, source, publication date, number of vowels in the title. Everything. By that logic, a collection of posts matching a specific tag is more important than their source.
  4. Smartness – Artificial Intelligence
    In a pool of 200+ feeds, no reader has yet solved the problem of ‘What should I pay attention to right now’. This is a problem far larger than a lowly feed reader. Many of the points we’ve already talked about start to solve this problem
  5. Design
    Agreed. All the browser-based readers I took a look at are dang ugly. Though, it could be that I’m such a fan of NetNewsWire my judgement maybe cloudy.
  6. AJAX
    Studies have shown that Duck tape is actually really bad at fixing ducts, and really good at removing warts.
  7. Server-side
    Seems to me this is where the action is in browser-based feed readers
  8. Order
    Some ordering presentations are far more useful than others. We’ve yet to find the one that really works.
  9. Direct Comments/In-line Comments/Trackbacks
    Agreed. Commenting and feedback not posted to the source isn’t useful in most circumstances. There are a rare few where it is.
  10. Blog/Email
    Yes, and this interaction is so often executed so poorly.
  11. Top and Related
    This goes back to the bits about attributes, tagging, and sorting and yes, why aren’t subscriptions easier to share? Darn good question.
  12. Mark all Read/Backup Copy
    This is only a useful solution if the opposite was a problem. What if it wasn’t?
  13. Keyboard Shortcut
    Fersure. Bloglines does this very well.
  14. Alerted
    I thought that’s how RSS worked best?
  15. Bookmarklet
    Now we’re solving the problem we talked about up in ‘Automatic Feed Finding’
  16. Translation
    I’ve read Douglas Adams auf Deutsche, he wasn’t as funny.
  17. Email
    Seems like a temp-fix for newletters that aren’t yet RSS. Just checking.

T-68 and Counting

I made enough progress on the core functionality this weekend, that I can actually start thinking about how to implement the interesting bits.

T-71 and Counting

Lots of progress the last two days. The absolute core functionality is in – enough to move forward to the interesting bits.

Yes, it’s a feed aggregator, River of News-style.

T-75 and Counting

PodcastNYC lists out 9 Ways to Kill Podcasting. Articulating some of the ‘this feels wrong’ parts of the iTunes, Yahoo podcast directories.

On the other side, thanks to Dave Winer for comparing the river-of-news vs. portal presentations. Personally, I see time-based navigation as a good way to see how a story evolves – though, I think the story itself should be the primary navigation method.

Any individual item in an rss feed has a number of attributes. From my perspective pubDate, source and author are of secondary to title, link, description.

Oh, and if your interested in getting early previews of this project – drop me an email.

T-79 and Counting

No direct progress today, mostly a ‘cleaning the desk’ day. Think found a solution to the roadblock I mentioned yesterday – not enough spare cycles to verify.

T-80 and Counting

Hit a fairly big roadblock today. I’m sure there’s a number of workarounds. I’m guessing this difficulty is a hint that I need to think about the mechanics of the project differently. I very well could be building it the hard way – and the roadblock is telling me that.