About one week ago I wrote a post with a similar title on my Italian blog. It was to announce the “beta” of Filo, a small service (well, more like a weekend project) that I developed a while ago, and that turned out to be a good testbed for experimenting a bit with design ideas and development practices.
What I’m giving you here is an introduction to the project and an overview of its features, I’ll update my dev blog with more in-depth articles about the techie stuff.
What is Filo?
Filo is a website that allows you to keep track of what you want to read (and don’t have the time to do it right now). It was heavily inspired by Instapaper, a web service by Marco Arment (one of Tumblr’s developer) and basically started as an exercise to bend Instapaper’s behaviour to my needs.
Wasn’t del.icio.us / ma.gnolia / $othersocialbookmarkingsite enough?
Yes and no. Other existing services are mainly aimed at people who want to share their bookmarks with somebody else or who want to tag / archive with their own tags.
This is awesome, but requires that you actually know where to put that document!
Filo instead provides you a simple, strictly first-in-last-out list of items you want to remember.
How does it work?
Items to read in Filo are called knots (that’s a little linguistic joke, as filo means line in Italian). You can create Knots either manually (using a form accessible from everywhere on the site) or using a bookmarklet that you can drag in your browser’s bookmark bar.
Using the bookmarklet is very easy: you just browse to a page you want to “remember” and click the bookmarklet. It will contact Filo in the background and create the new knot automatically.
Once a knot is created it will be available from the website and in your personalized RSS feed.
Once a knot is accessed (either clicking on it on the website or clicking it’s title on the feed), it will be marked as read and archived. It is possible to mark archived knots as “to be read again”, as it is possible to trash knots entirely. At the moment, there is no way of bringing items back from the trash (but it will be possible in the future).
How to access the service
The procedure to sign up and sign in have been reduced to a bare minimum: when accessing the site, you’ll be prompted for an email, just type in yours; if it’s recognized as an existing user’s, you’ll be asked for your password, otherwise a new user will be instantly created so you’ll be able to start generating knots!
You’ll also receive an email to confirm your address. You’ll need to click on the link provided in it to fully activate the account (and be able to log in again in the future).
Note, the über-simplified registration process was first presented as an idea by Davide Casali at the recent ExperienceCamp.
Localization
As of today, Filo supports English and Italian.
Mobile version
Filo is already designed to be used from mobile devices (well, actually the CSS still needs a bit of love). Moreover, there’s an iPhone/iPodTouch version available at http://filo.m.bzaar.net/ (well, you can go there with any browser, but with the iPhone is cooler ;) ).
Reading from a feedreader
Filo creates a personal RSS feed for each user. This feed may be imported in any feedreader. Every time you access a knot form your feed reader, it gets automatically archived in Filo (and will disappear from the unread feed at the next refresh).
Boring Technical Details
Filo is written in Ruby, builds on the Ruby on Rails framework, and stores its data in a couple of MySQL tables.
Front end logic is powered by jQuery and the whole thing should degrade gracefully. The iPhone version uses the iui library.
Potential Troubles
. Filo is hosted on DreamHost. Now, don’t get me wrong, DH is great and considered what I pay for the hosting, the service I get is just awesome. BUT it’s not really inteneded for hosting Rails applications: we’re running on Apache + FastCGI and, well, the whole thing tends to “feel” quite slow.
The good site of it is that I put some decent effort in optimizing the code, so when eventually Filo will move to a more rails friendly service it will possibly scream (well, maybe just whistle :) ).
. The whole thing (and especially the integrated login + registration) is not thoroughly tested on a number of different environments (e.g. IE and javascript-less)
. CSS needs some love, and the same is true for the user settings page.
Where next?
Some ideas for the future:
. OpenID support
. “social” stuff, as being able to read and comment your friends knots.
. suggested reads
. offline storage of long articles