fredag 27 juni 2008

A new law of hardware envelopment

All portable technology attempts to expand until it can play mp3. The hardware which cannot so expand is replaced by hardware which can.

Case in point: Amazon Kindle / Sony Reader

söndag 22 juni 2008

Recently some people have been discussing the so-called darknet, or person to person copying, specifically with regards to media piracy.
So how do we store and use information? Let's define three main variables of personal information storage and use:

H - the amount of information a human is interested in processing in a given timespan.
As new high-resolution movie and lossless multichannel audio formats are introduced, H goes up. H, however, has a hard limit depending on visual and auditory acuity, and can be expected to top out below the gigabit/s level unless touch or smell television are invented.

S - the amount of information a human is interested in storing for safekeeping and future consumption.
S fluctuates wildly with the various technological means. High quality streaming audio services may decrease S, whereas information scarcity - whether artificial by means of draconian copyright laws, or natural by means of relocation to a country with limited internet connectivity, will increase S. One could easily imagine A, the inverse of S, as being the amount of information easily available outside local storage, in other words the sum of streaming services.

I - the amount of information that can be exchanged with other parties in a given timespan.
Information exchange occurs by several different vectors - sight and sound, sneakernet exchange of magnetic or optical media, across the Internet or other networks. The preferred vector of exchange might not necessarily be the fastest one in terms of bits per second, as convenience could favor a lower bandwidth channel requiring less human intervention.

How do these variables affect the way we use and store information? It's interesting to note that the early days of wide scale digital information exchange favored copy-parties and physical storage media.
The Internet won out as the information exchange medium of choice not because it's so incredibly fast - it just became fast enough, and was more convenient from a collaboration perspective.
This happened around 1996, and can be attributed both to the historical and technological fact of private internet connection bandwidth finally exceeding the H of those days, as well as the invention of efficient compression techniques, allowing H to be squeezed down to the meagre channels afforded.
Since then, we have increased our standards to conform to increasing availability of storage space and bandwidth, often trading one for the other. Projecting into the future, there are two possible outcomes:

1. H < I, lim(S) = 0 - Total streaming. As bandwidth increases, people prefer to use online storage solutions rather than expose themselves to possible data loss. This neatly severs the gordian knot of DRM, and allows the media industry their vaunted subscription solutions. Actors in this camp are Google, Apple, Netflix,, Spotify and all the streaming music services. It's interesting to note that streaming is emerging as the dominant market model for audio, the lowest bandwidth medium, as available bandwidth has indeed now far exceeded H_a, the amount of interesting audio information. Streaming as the winner is also indicated by physical storage growing exponentially in storage space but only linearly in access speed, a fact immediately obvious to anyone who has copied a dvd movie to a USB stick.
2. H > I, lim(S) = inf - Digital vaults.
As physical storage increases, people prefer to permanently store thousands of movies conveniently at hand instead of waiting even half an hour for downloads or initial streaming buffering.
In recent years, hard drive capacity, increasing exponentially, has obsoleted optical media like cds and dvds as the preferred storage medium, reducing the physical space needed for media storage to a fraction of its historical amount, and necessitating the introduction of blu-ray to feed the media industry's demands. If this exponential growth continues, there might be no need for streaming, since every person will have their own copy of every work ever produced.
Big actors in this camp include the media industry, which would prefer to sell something - anything - physical to which information can be tied. Their task is made easier by the shrinking of the physical medium - today a high-quality feature film can be stored in a micro-SD card about the size of a postage stamp folded twice.
Digital vaults as the winner is indicated by the exponential growth of available storage and guaranteed portability of storage to low-bandwidth environments.
People climbing Mount Everest or flying long distance will surely want their music to be in their pocket rather than at the end of a million mile pipeline.

In the end, the winning paradigm will depend on the limits of H. If media companies can come up with new ways to grow H, there is a future for the digital vault.
When, however, the limits of human senses are far exceeded by internet bandwidth, when connectivity is universal and if all interesting works are made available at a flatrate subscription model, there will be no need for local storage.
Until then, the race is on.