Freie Freude

Thanks to Igalia, I am in Berlin this weekend enjoying the Desktop Summit!

Today I have been reading about the work done by the Distributed Computing group at the Zurich Institute of Technology. These guys are developing complex techniques for combining “audio social information” (tags and listening habits, taken from Audioscrobbler) to discover relationships, thus generating a social audio space where songs/artists are placed and that the user can navigate. An opposite (or maybe complementary?) approach to this one would be the use of audio features to establish relationships between songs/artists, but this might ignore some of the semantic and social information that is so relevant to our understanding of music. They have developed a set of prototypes for Android (with video) and Amarok. The science looks interesting and these kind of approaches could well be the next step in the evolution of music players now that the average collection is often too big to be dealt with using only long text lists.

Ok, back to the paper now (Social Audio Features for Advanced Music Retrieval Interfaces, if you are interested).

Today I want to think about a possible mobile application to fulfil the goal of getting “a varied stream of my music that suits a general theme and that requires little interaction from my side”. For the particular case of mobile music:

• Music can be in the background, allowing the user’s attention to be directed elsewhere.
• Portable music devices are often pocketed away. Direct interaction and user input are therefore sparse.
• Mobile music has usually only one listener.
• Fine management of the specific songs being played and their sequence might not be required.

These are all assumptions, but I will go with them for now.  Bear in mind that this is just a bit of exploratory design and so far hasn’t been validated.

The main idea is that the user should be able to just select the musical “theme” and get music along it. Since both tags and genres have a subjective component, I thought that using music artists as the “seed” for the radio stream might work well. If this sounds similar to Last.fm‘s radios, yes, it is similar.

Last.fm uses streamed content and relies on crowdsourcing to generate the relationships between artists, tags, songs, etc… One of the constraints that I had in mind when I thought about this application was that the music should come from the user’s personal collection and not depend on having an Internet connection. More importantly, Last.fm’s radios offer a wide range of choices and therefore require complex UIs, ; one of the questions that I am wondering here is if one could strike a compromise between providing music that satisfies the user’s needs and doing so with an interface that is simple, quick and pleasant to use.

This is a quick sketch. In this hypothetical app., the user is presented with a list of the artists and bands in their collection.

An image of the band could be used to make the application more visually attractive and easier to recognize. It is worth noting that the list layout shown in the image is not necessarily the best one for this particular case. The proper thing to do if we were serious about developing this would be to prototype, test and evaluate different possible list layouts in a systematic way: one column, two columns, images, no images… This might make a good starting point for a future post about testing.

It is worth repeating that this is supposed to work on the user’s music collection on a mobile device, which on average [citation needed] would be limited to dozens/hundreds of artists, most of which the user can recall or at least recognize [citation needed]. A different problem altogether would be to explore a vast collection that has been compiled by somebody else, as would be the case with e.g. Jamendo.

When the user selects one of the items, the application begins to play a stream of music that is related to the artist selected. The definition and calculation of similarity in music is a complex engineering topic that I will cover in future posts. Determining the “suitability” of the algorithm selected would not be trivial since what we are looking for is to provide the selection of music that better fits subjective desires and expectations, which change from person to person.

The “now playing” screen is intentionally simple: you can pause the current song, skip to the next one or go back to the list in order to select a new band. I am not even fully convinced that the “settings” item makes sense there, but it could act as a place to keep functionality that does not need to be always present (e.g. play only music from the selected band and not from related artists).

That’s all for today. Stay tuned for more stuff coming in the next days!

The first problem that I want to talk about in this new version of the blog is that of managing, browsing and consuming media content, with an initial focus on music and mobile devices. A good summary of current technologies and research can be found in Tim Langer’s paper “Music Information Retrieval & Visualization”.

Thinking about the systems’ user interfaces, we could imagine a graph displaying the level of control about the concrete songs being played versus the attention and frequency of interaction required (before and during playback). Different ways of enjoying music could be displayed in this graph, for instance:

• at one corner we would find the traditional radio, which requires no interaction except for selecting the station and offers no control of the music being played;
• Last.fm radios would be nearby: they require a more careful selection and offer a bit more control during playback (skipping songs);
• on the other end of the graph, we could think of a music game where the user can not select the music being played but is required continuous interaction to win (the interaction being the whole point of the system);
• the extreme combination between control and need for interaction would be to individually pick one song after the other;
• creating a playlist and playing it would be similar, but the required interaction is concentrated at the beginning, when the list is created;
• selecting an album would be similar but does away with the need to create the list in the first place;
• etc…

More variables could be taken into account. Of great importance is the perceived, subjective “quality” of the musical stream, how much it conforms to expectations, fits the user’s mood and follows an overall theme. Another one is the discovery of new music, which very naturally leads to a social component when we want to share our discoveries. The social facet of music does not stop at sharing: music is often used as a means of self-expression, as one of the ways that we use to create and present an image of ourselves to others. These concepts and more may determine the choices and trade-offs made during design of an interactive solution.

The ideas outlined here so far are based on some reading, personal experience and a few assumptions. Assuming is not necessarily bad but one needs to keep in mind that an assumption is not the same as hard knowledge: eventually, assumptions need to be challenged and tested. I will talk about this testing in future updates.

Just a quick note to let the world know that today I am officially becoming a teleworker and moving to Barcelona. I am really thankful to Igalia for giving me this flexibility; it is amazing to be able to make your job and your personal life work together instead of having one determine the other.

Finally! I have returned to this blog after a year and a half of absence. My last entry was written in the Canary islands during the GCDS and lots of things have happened in the meantime.

On October 2009 I took a year off from work to take part in the MSc. in Human-Centred Interactive Technologies at the University of York. This is a full-time MSc. that focuses on providing a theoretical basis and research methodology for the design and evaluation of interactive systems. The experience was great, I learnt a lot, met many good people and even got a distinction out of it. The whole thing began as a crazy idea but now I am very happy that I was able to do it.

I returned to Igalia a couple of months ago and am back at working on free technologies. Now that I have some formation on interaction design and usability, I plan to gradually orient my work more and more towards these topics. I really hope to be able to write here more often.

Here is my account of the things I saw and learn during the second day of the Gran Canaria Desktop Summit, a day later than intended thanks to the fine folks at Nokia who invited us to a party  which began at the sensible and European (and therefore quite alien for a Spaniard) time of nine and a half.

I attended many talks yesterday, of which the three that I found more interesting where the ones about Tracker, Zeitgeist and GnomeShell. It is true that the sum is sometimes bigger than the parts alone, and we might well be before one of such occasions.

Tracker indexes your information items (on the filesystem and on the Internet), and records properties and values about them. In the future, such values could be not only basic types but also objects themselves, thus creating graphs of information that represent more accurately the complex ways in which we interact with our data.

Zeitgeist strives to solve a complex problem with an elegant solution. The problem is identifying what the user is really doing at a given time, understanding the activities that are going on even if they involve several different applications, files and whatnot. The solution, at its simplest, is just logging events and then analize them in order to extract the history and the current situation of the user’s interaction with the system. Events can be of two types: activities that the user does (i.e. open a file or URL) and notifications that just happen (i.e. incoming email or changed location).

The most interesting point about Gnome Shell is not its concrete user interface, but the technologies behing it. The actual code is just a few thousand lines long, thanks to the use of JavaScript and several libraries to make it able to access Clutter, the window manager and other GObject-based libraries seamlessly. Its use of the two other projects mentiones above offers an exciting playground for developers to try their crazy ideas about how the interaction with our computers should be.

And now, time to run to the conferences again. If you are in Gran Canaria, don’t miss the Gnome party tonight!

This is my first post about this year’s Desktop Summit in Gran Canaria, and also the first entry in this blog for a long long time. My intention is to keep a diary of the conferences here, if I can find time between attending the talks and going to the beach.

The first talk of the day was given by Robert Leftkowitz. He made the case for software development as a liberal art, and Free and Open Software as the kind of software “that a gentleman would use”.  The concepts he used go back a long, long way. It is extremely uncommon to hear ancient Greek philosophers being quoted in a keynote. It is always reassuring to think that two or three thousand years ago human being just like us encountered the same problems and moral dilemmas that we face today. By the way, I had never before heard someone say that software is rethoric, and it is a suggestive idea.

The second keynote was given by Walter Bender, who gave a overview of the Sugar learning platform and the humane and altruistic motivation behind it, always keeping a “low floor and no ceiling” attitude. I specially remember one simple sentence that has hidden depths:

“Nothing in our children’s future resembles 1970’s office work”.

The third keynote was given by Richard Stallman, and it was weird. I have seen Stallman a few times, but never before had he managed to fit a history lesson, a flamewar about C#, an auction for a gnu and the geekiest song on Earth in the same talk. My life would have been completed if at some point he had drawn a couple of katanas to fight the enemies of freedom.

After lunch, Quim Gil gave the news that Nokia will discard GTK+ and Hildon for the Harmattan release of their Maemo platform, and will be turning to Qt as the main UI toolkit as part of a bigger Nokia movement. He also discussed some of the free technologies that will continue to form the base of the software run by the Internet Tablets.

The lightning talks were quite entertaining, but some of the projects exposed (i.e. Gothenburg, stuff around Nepomuk, etc…) could certainly have deserved a bit more time.

See you tomorrow!

Christmas is here once again, and as always it is a good time to look back. A year ago I was living in Germany, where I was doing an internship at IBM. This Christmas, I’m in Coruña again, working for Igalia in projects of such amazing coolness as Modest and Vagalume. The road has been hard sometimes, but also fun and interesting, and I can only hope that it keeps that way so I can write a similar post next year.
My only advice for the incoming year is to follow the lesson behind Vagalume’s incoming “Recommend to…” button: look for beauty, share it with your friends and have fun. And do a little dancing, too.

Since my workmate Berto released Vagalume, here at Igalia we have adopted the application and giving him suggestions about how to improve it. The N800 makes a nice little Last.fm player, ideal to keep around while you’re working, and therefore the latest version, that displays album covers within a nicer interface, is a great improvement.

Besides creating the icon for the latest release, I have done some other work on the graphical part of Vagalume in my spare time that I would like to share today. First, here we have the (beta) icon shipped with today’s release in a bigger size.

It looks nice when it’s big, but on the device I think that it doesn’t scale well enough, and the three fireflies look more like comets. That’s why I created recently an alternative one, with only one flying light path:

Representing a firefly (that’s what vagalume means) in an icon isn’t an easy thing, for fireflies are small and not quite pretty under daylight, so showing their light paths was the best choice IMHO.

Now that the functionality is there, we’ll devote some time to the UI, to give Vagalume the mojo it deserves. And finally. to end our world domination plan, one of the next things to do should be a desktop applet. Here are my humble suggestions:

I hope that this post motivated you to try Vagalume and take active part in its growing community. Let the ideas come!