Sunday, March 27, 2005

On the Wrong Track

MusicbrainzThat's the title of an interesting article I found on Thursday's edition of The Wall Street Journal Europe. It highlights the increasing issue of music metadata, i.e. information on digital music files, like track name and number, artist(s), album title, genre and so forth. Back in the now old days of physical medium like LPs and CDs, one didn't have to look very far to find those information printed on the disc cover.

Today, music metadata has become as important, if not more, than the actual music they describe, a fact taken early on by grass-roots efforts like CDDB (CD DataBase - acquired since then by GraceNote) doing a great job at giving information on CDs. But while CDs were still providing some structure, the proliferation of "non-physical" music sources such as homemade rips, digital music stores, and above all file-sharing, makes it far more complex to organize gigabyte-sized music libraries.

In that sense, I have been positively surprised after several months of using it by the quality of MusicBrainz, a self-moderating metadatabase that helps identify not only CDs but also mp3 and other digital formats through some clever fingerprinting as well as community-powered updating and editing - in short a Wiki for music information. In my view, the next logical step is to have such smart recognition directly integrated into media players like iTunes: as Suw underlines, the comings and goings between MusicBrainz and iTunes are tedious, to say the least. Even better would be coupling such a music recognition tool with a recommendation service like Last.FM: information not only about the track I'm listening to, but on the ones I may also like as well.

Once again, bringing back relevant context to the digital media experience.

Sunday, January 30, 2005

The Long Tail and active media consumption

TailI am a strong believer in the Long Tail theory and observations, as first written in the best Wired article in 2004 and followed up in an eponymous blog, where Wired editor in chief Chris Anderson brillantly shows how the Long Tail - this inverted exponential curve, is driving the digital media industry.

In a nutshell (but I do encourage you to (re-)read Anderson's article and posts), the red part of the curve is the "mass target" content: the few TV channels broadcast by the big national networks, the Top 50 music and book hits available pretty much anywhere. Now with the low-cost entry barriers of distributing, streaming, selling or plainly putting content online, much more focused and targeted content adds to the total media available. Think of this yellow part as the power of the niches.  Services like Amazon for books, independent online labels for music, local or special-interest TV channels through ADSL TV all bring us a longer tail, a richer choice, a potentially more fulfilling media package.

Such choice and possibilities is fatally followed by an increased complexity that still today resides on the user side. Of course, PVR (Personal Video Recorders) like TiVo in the USA have attracted hundreds of thousands, but now show their limitation in thin-slicing the offer and smart-filtering content (not to mention the regulatory and legal "gray area" services like that are being pushed to).
Once again, the solution could well come from the Web side: Google recently unveiled its Video Search engine, along with Yahoo's own version, which are not hard to image being soon extended to handle mode media types (think streamed radios and podcastings), should help to make aggregation eventually easier.

It is especially important for TV as it is by essence a passive activity, despite all the attempts to make it glitter with interactivity and active use: feeding a household with the whole chunk of 1000s of channels will not make it spontaneously feel better served, except for isolated groups of very early adopters and technophiles. The key once again is to use those "glue" tools and services to  immediately provide a thinly-sliced and carefully filtered choice of channels and content. In short, generalize the iPod Shuffle paradigm, with stripped-down interfaces and smart randomness.

Monday, October 04, 2004

Parking, 3G style

Felica_parkingAs reported by Akihabara News, Japan is once again on top of the technological curve and ICT diffusion in everyday's life. Honestly, what could be more commonplace and anecdotic than the immutable ritual of parking you car? That is, opening the window, taking a piece of printed cardboard you're likely to lose in your pockets; and then, after queuing for paying (if the machine is not down), rushing again to your car before the "exit" timer has expired....

Well this painstaking reality could soon become a thing of the past, thanks to FeliCa, a technology originally developed by Sony. Basically, it is a contactless chip, that holds more information than on the credit cards' chips, with the added advantage of being contactless, as the names implies: just swipe it in front of a wireless reader and that's it (similar to RFID tags).

Now, the neat thing that those chips are integrated into NTT DoCoMo's latest third generation (FOMA) phones - such as the brand new F900iC. It didn't take long for Park24, a large parking company in Japan, to take benefit of that and launch a FeliCa-powered parking area next to the NTT DoCoMo Yoyogi Tower (how appropriate!) in Shibuya, starting October 1st.

In a nutshell, when entering the parking lot, you swipe your 3G phone at the entrance beacon (cf. picture at the top), that way you can pay for the exact time you have been parked. But wait, that's not all! You also get a printed 2D barcode on you receipt.Barcode_parking. Those little pixelated squares are (of course) understood by your camera-phone, and upon taking a picture of it, you are redirected to a surveillance camera pointing at the direction of your parked car.Camera_parking_1
Well, considering the average crime rate in Japan, I don't know if it's really worth the technological paraphernalia, but it's at least a very convincing showcasing of Japan's mastery in mobile, user-pushed and market-driven services.

I now remember how I did find Mint's parking payment in Sweden through SMS already way ahead of its time (compared to lagging France, that is). Different cultures, different adoption curves...

Thursday, September 30, 2004

Opening up the last mile in Sweden

FiberI attended yesterday an inspiring talk given by Lars Hedberg, Secretary General of the Swedish Urban Network Association (SSNf - Svenska StadsNätsföreningen): a non-profit organization, founded in 1998 and headquartered in Eskiltuna, a small city about 100km west of Stockholm. The association was founded by companies involved in fiber optics installation in urban areas as well as metropolitan networks in the greater Stockholm. Today, it is composed of more than 300 members - companies, organization (KTH is a member along with other Swedish universities), more than 150 metropolitan networks represented connected and lots of infrastructure providers (such as Ericsson, Alcatel and Cisco).

The goal is to create a whole new digital infrastructure all over Sweden, based on the principle that healthy competition is necessary and promotes growth through ICT development: it is shifting the paradigm from the traditionally state-owned or former state-owned "last mile" (i.e. the final run of cable that connects the end-user to the first active gateway). France Telecom, Deutsche Telekom, British Telecom are good example of painful and slow transition to unbundle the local loop. Sweden, with the former state monopoly of TeliaSonera, has somewhat better managed to open itself up to the competition. However, with 4.5 million households ultimately freed to choose their providers, it becomes critical to create a new technical and business model and framework, not necessarily controlled by the state, but monitored and regulated by the state.

Of course, Sweden itself brings specific challenges to this vision, most notably the geographical consideration (In northern Sweden, there are fewer people living there than in the entire Sahara desert!). To counterbalance the economic considerations with such a scattered country, the 290 municipalities in Sweden have to cooperate and bring in common their expertise and existing municipal network (as they represent a lot of spare capacity that can be offered to companies, residents and public agencies). As a result, the mission of SSNf is to encourage it through the following actions:
- Give the city and society an open infrastructure, controlled by the city and open for everyone (same terms and price).
- Stimulate the market to use the infrastructure, to share it, instead of having each player re-building everything again from the ground up.
- Make possible a fair pay-back of the invested money (self cost).
- Reduce digging in the street (e.g. in Stockholm there are 70 operators: just imagine if everyone of them was allowed to dig... ).
- Document and rationalize the process: SSNf works closely with with documentation projects and organization, to leverage and standardize on common standards and best practices.

And it actually seems that Sweden has the power to concretize this ambitious vision. Stockholm, through Stokab has one of the largest dark fiber network in the world. It is also quickly becoming a major Internet eXchange point (IX), especially towards the Nordic and Baltic countries (cf. the BalticOpen.net project with a fully functional fiber connection between Sweden and Latvia), Eastern Europe, and even Asia (projected connection of China to the western backbone through Russia, Riga and then Stockholm).

Lars then presented the actual topology for deploying this Open network in Stockholm, as an interlaced double ring:
- A red ring, as the network running in the community of Stockholm (180 schools, libraries, offices, City Hall...). This network powers 46000 employers and 80000+ students.
- A blue ring, as the network for healthcare (30 hospitals and healthcare centers amounting to 26000 patients/week).
These two rings are connected to one logical point, which holds the firewalls, communication equipment, and service distribution, meaning that service providers offer services to 100000+ users from a single point, making it quite simple to distribute the services.

The vision is that such model can be successfully deployed all over Sweden, even at a smaller scale: SSNf helps its members connect their respective city networks ‘islets” to each other and collaborate, cooperate to be more effective overall. In 5 years (+/- 2 years), the objective is to have a nation-wide mesh of local and regional nodes, throughout Sweden.

This will in turn empower the next step - the Broadband connected Household (as we can already witness to some extent in Japan): nation-wide Fiber To The Home (FTTH) makes it possible for advanced Triple Play services: High-Definition TV (HDTV) needs at least 20Mbit/s per channel, rich-service (gaming, telemedicine, e-learning) are also bandwidth-guzzling. Of course it means heavy upfront investment, but on a longer term it could save 300000 SEK (~33000€) / person / year!

Great outlooks indeed, and as the price of fiber and related equipment continues to drop, it will soon reach a comparable price to using copper or DSL for the last mile, with the added advantage of offering more than ten times the bandwidth. The real challenge is to craft the proper business models so that broadcasters come in and take advantage of this virtually unlimited digital distribution channel, without having tens of incompatible equipment to buy or competing standards for running services. The nation-wide switch to digital TV in 2008 is therefore a tremendous opportunity for a massive adoption of the faster, bigger and better digital world.