Information Infrastructure Projects

Today Obama is inaugurated and steering his economic recovery plan through Congress will be a key target for his first 100 days. Governments throughout the developed world are looking for significant infrastructure projects to stimulate employment and to stimulate demand. That is the good old Keynesian solution to a recession which has come right back into fashion. Why has there been so little attention given to public infrastructure investments in the information field? Universal broad band seems to be the only plausible candidate so far ($6 Billion is the projected cost of this in the US plan, less than %1 of the total package). Most of the ideas that we read about from Obama, or from think tanks, are that employment will be created by investing in mass transit, or in efficient power transmission, social housing, or neglected infrastructure projects (old bridges and roads etc). Surely we should also be thinking about infrastructure projects for the information sector? But whether the projects are physical or digital, it has to be recognised that there are at least two problems with pump-priming investments. In the blogged comment of the Nobel Laureate Gary Becker:

The activities stimulated by the [Obama] package to a large extent would draw labor and capital away from other productive activities. In addition, the government programs were unlikely to be as well planned as the displaced private uses of these resources. (Becker -- Infrastructure in a Stimulus Package)

So large-scale publicly funded information infrastructure investments that are to be part of a stimulus package must not suck up highly skilled labour which is in short supply, and if they are to be part of a timely intervention they must not involve substantial leadup and planning time. Its a no-no to suggest putting billions into the creation of vast libraries of Open Source software, after all most competent programmers are gainfully employed. As for the requirement of effective planning, it will be much easier to identify a promising project that can be scaled up or replicated than to start something completey de novo. There are such projects in the information field, and rapid scaling up looks most promising for initiatives which involve large amounts of data collection. Data collection is often relatively unskilled, or with skills that can be easily taught. Of course, it is also important that the information being collected is of real use. Here are three suggestions:

• Open Street Map -- has been making great progress has global ambitions and is run and led by enthusiastic amateurs who appear to be capable of organising relatively large scale social endeavours (mapping parties/conferences). Governments could fund volunteer groups who would generate large amounts of geo-data, perhaps with particular emphasis on ecologically relevant mapping data, and culturally relevant historical data which is neglected by commercial cartographers.
• Google Book Search has made a very substantial step towards solving the problem of 'Orphan Copyrights' in the USA, but there are nevertheless plenty of gaps in the US data, and the Google collaboration with university libraries is much less advanced in other countries. Although using this data to full-effect reguires the Google engine (or something like it from Open Source or another company), the accumulation of the raw scans is a relatively low-tech business. It needs little more than a careful and well planned logistic pathway and an investment in scanning machines. If governments in Europe and the rest of the developed world were now to invest in large-scale scanning of the published literature, they could reach effective agreements for Open Access or appropriately regulated Approved Access to enormous bodies of printed literature. This would have the additional benefit of putting in the public domain, material which ought to be of general benefit; the Google effort would also be, and seem to be, much less monopolistic if there was a large body of government funded scanning as a counterweight. Much of the technical expertise necessary for scanning the world's out of print literature already exists in libraries and, with the example of Google Book Search in front of them, the world's libraries could be geared up to scan, within a few years, several times the 7 million books already scanned by Google.
  
• Ecological audits. As we tackle global climate change it will become increasingly necessary to have better information about animal and plant species and the environmental impacts of climate change on living systems. Amateur efforts like the British Bird Survey are pioneers in this work. If Governments invest now to create better and ongoing records of natural diversity these records will be very useful in the global challenge of moderating and putting a break on climate change.
  

Going Local

Yesterday's FT had a piece about mapping as an interface to the web. This is one view on why this change is important:

Erik Jorgensen, a senior executive in Microsoft’s online operations, says the software company is building a “digital representation of the globe to a high degree of accuracy” that will bring about “a change in how you think about the internet”. He adds: “We’re very much betting on a paradigm shift. We believe it will be a way that people can socialise, shop and share information.” 'Way to Go? Mapping to be the Web's next Big Thing', Financial Times, 21 May 08

Google, Nokia and others are investing in parallel projects. The article speculates that controlling the geo-interface may put one company in a dominant position. But perhaps that will not happen, in part because their is an open source foundation under construction in OpenStreetMap, Its coverage is improving in Wikipedian fashion (getting better all the time). The current view of Florence is good on the railways and autostrada, but lacking in detail.

As it happens we have started adding geo-tags to our data this week (so we can now render as live links, post codes mentioned in text or advertisements). We will blog about this shortly. As a side note: one guesses that geo-coding will become important to us all for one reason not mentioned in the FT's article yesterday. But headline news on the front page. Oil goes to $135 a barrel. It is not really a paradox to suggest that we may care more about exactly where we are, as we learn to travel less.

Digital Editions and Greedy Networks

Digital editions avoid most of the wasteful practices which bedivil the print magazine industry (paper not needed at all, physical distribution costs gone, wrappers and collateral offers virtual or abolished, landfill for used magazines spared, no returns from sale-or-return etc.). But there is a worry that the environmental savings will be wiped out by the unsustainable and resource-greedy growth of the internet. Steve commented on this yesterday and he cited a fascinating discussion of the issues -- concentrating on Second Life.

Larry and Sergey, the idealistic Google founders, are anxious about the growth in power consumption for their data centres and users' network devices. When one starts to worry about the ecological impact of digital publishing, power ends up being the only real concern. On all other concerns the business is much, much cleaner than print publishing. If Larry and Sergey think that there is a problem with excessive power consumption, I am not going to contradict them, but it is notable that the profile of the problem has changed. What matters now is not, how much power and paper publishers use and waste. The key question now is what resources will users expend and will their pattern of digital consumption be very greedy? With traditional print publishing the resource use and the ecological impact stems from the printing process. There is a shadow, a trail, of ecological impact from the printing plant outwards. The only way users add to this ecological impact is by switching on the lights so that they can read the magazine that has been so laboriously shipped to them. But with a digital information system, the ecological impact and the power consumption is much more a matter of how and how much the system is used and what devices are operating at the periphery. The publishing function uses relatively trivial power and cash resources. In the next generation, the consumers, and their network devices, not the printing plants and paper factories are where the ecological impact is to be located. Brewster Kahle gave a fascinating lecture at Rice University two weeks ago, in which he pointed out that the Library of Congress, all 20 million volumes of it, could be stored in a tower memory box no larger than the podium from which he spoke. A large memory disk system of that capacity can be had for $60,000 (as he noted that is the cost of a garage or a good party in Silicon Valley) and the power used in running that box for a year is comparable to the central heating costs of a large house. Having and powering a very large digital library is cheap and getting cheaper. All the magazines and newspapers ever published could easily be accommodated in such a box. But it is the way consumers use digital resources that will have the environmental impact. If consumers care about ecological impacts then ecologically efficient systems will be built and used.

I find this an oddly hopeful thought: however the ecological challenge of global warming is to be met, it is only by a broad shift in consumer behaviour and expectations that we will see any effective action. Governments and business groups are getting concerned about climate change and ecological issues; but we can be sceptical of all treaties, targets and policies unless and until they have widespread popular support. Consumer behaviour is the key. Which will be my contribution to today's conference Green and Lean: how will publishing survive?.