web affordances · 2005-06-27 00:09

We’re at a strange stage of web development where programmers have a far better grasp or view of what’s possible and available for many web services that normal users. Web services (program to program) have clear apis, documented holes, bugs and changes. The people using the website version of web services have nothing like this.

A case in point is Bloglines, my aggregator of choice (I fell off the wagon – on schedule – more on this later). There are two hidden inbuilt rules that change how you use the service once you know of them:

1. feeds that disappear disappear

This one is pretty criminal to users: if a feed is not where it should be for 3 days, it automagically unsubscribes everyone from the feed. Without any notification. Bloglines is one of my outboard brains, my memory. If I’ve trusted Bloglines to remember what I want to see on the web, it should remember. Maybe flag the feed, or move it to a sinbin, but at least make the change visible. I have no idea who I’ve been unsubscribed from – I know my own feed was down for more than 3 days when crackers and a server move coincided. Who knows whom you’ve been unsubscribed from without warning. Export those blogrolls now.

2. 200 is your limit

Each feed carries a maximum of 200 entries. If I’ve been away for a couple of weeks, Engadget, The Guardian and similar feeds have easily hit this. It used to be, I think, that Bloglines always had the latest 200 – now, weirdly, it has the first 200 until you hit the limit (they store far more than 200 for bookmarks/Keep New, but the entry does eventually disappear, but the Keep New stays as a blank entry). I personally want the latest 200 – or there not to be a limit. After all, they’re serving others with later entries – why can’t they be appended to mine?

In computer games, I’m very much a player that really just tries to feel the limits, poke at how it all works and see what I can do. I shouldn’t have to do this for web services.

(I don’t know if this is covered anywhere in the help – and after all, normal users shouldn’t have to read the manual. Personally I’d love for the user flows and business logic to be published, though.)

home