How significant are the linkages between g-governance (GIS based public service delivery) and SDI?
G-governance applications are a good test of whether your SDI is adequate or not. If SDI is inadequate, g-governance will fail. It might fail for other reasons too but this is one reason that will make it certain to fail.

How important is 'user needs assess- ment' while planning to develop and implement SDI?
User needs assessment is very important and it is the most frequently ignored aspect. The situation is tricky as it is difficult to define who the user is. Is it the intermediate user, is it the person developing and using the data which a mapping agency might think is a user or is it the end user because your aspiration is that the data and the system should be transparent to the end user. On the other hand, the end user experience is what makes the application work or not work. But I think you can't ask end user what he would like to have because they cannot imagine it. What you can do is through focus groups, pilot tests or real life tests, you can find where the application will work with the finding what end user has anticipated.

My experience of using geo-enabled applications on websites which local authorities now swear by is that almost none of them have been properly tested with end users. If after being 30 years in the industry, I find it difficult to use some supposedly geo-enabled websites, I can't imagine how an ordinary person will deal with them.

With regard to development and maintenance of SDIs, how different are the challenges in developed and developing countries?
The enormous advantage developing countries have is that they do not have too much history. The cleaner the slate you start with, the easier it is. If you look at the countries that have made fastest progress, both in terms of SDIs and g-governance, they had very weak systems before. Countries like India, UK or Malaysia, that have good SDIs have 200 years of history of national mapping agencies, of government security and of custom practise which one should overcome before starting to build modern SDI. I think developing countries have an advantage, particularly those being helped by international organisations and World Bank to build modern SDIs from scratch.

What about the technological issues?
GI technologies are a bit like buses, there is always another one coming along. It is a mistake to wait for technology to stabilise. Probably it will never do. But I think people who developed SDIs using technologies of 1980s and 90s find it hard to adapt to the world of distributed computing, of open standards, of cloud based computing, of computing that has global display platforms like Google Earth. If you invested millions in proprietary systems, then your willingness to dismantle those in order to do something else is quite small.

What are your views on security issues associated with sharing of geospatial data?
I think security issues are absolutely critical and the one issue that it is most likely to impede is widespread access of data. We can cite the trouble Google had with Street View as an example. People felt that it is an intrusion into their privacy. My own view is that privacy is passé. What we have to have now is guaranteed confidentiality, not guaranteed privacy. I think the genie is out of the bottle. What you need is strict laws on who could use spatial data and by whom and how the data subject could monitor/audit the usage. And then I think it is up to the industry to explain that most people are much more to gain from having the information in the system than not having it. Those who are weak in the society, the poorest, are the most least to lose. They actually lose by not being recognised within the system. It is the rich, privileged people who can afford the luxury of being anonymous.

What are your views on the sustain- ability of SDIs?
The bottom line is SDIs are sustainable only in two circumstances. One, they have made their case that they will benefit the citizen. So, the old problems of GI community like cost-benefit analysis wouldn't go away. Except that the cost-benefit analysis is always not financial. Sometimes it may be critical. Particularly, security issues can override cost considerations. But SDI is very easy to be cast aside. I think it is notable in the US that President Clinton signed the presidential order for an NSDI after Mississippi floods but by the time of 9/11, the data was not there for New York. By the time of New Orleans floods, the data wasn't there. Disaster is a big issue in the Press. It is a three-day wonder. It doesn't guarantee sustainable funding. To make it sustainable, we need to build aspects of every day life into SDI so that the consequences of removing it would be unthinkable to the politicians. A good example is the GPS. We are so dependent on GPS in so many ways that it is probably impossible now to remove GPS. So you become sustainable by becoming totally necessary. Another issue for sustainability is funding. I get quite cross because quite a few people always present the funding issue as a dichotomy. They say either tax payer pays from the tax funds, the next time there is no money in treasury it will get cut or you sell the data. I don't believe that's true. My view is that the most sustainable way of funding an SDI is by having a series of laws so that if you change the data in the SDI because you built something or knocked something down, or you acquired some land or you built a road, then you pay. No one complains if they have to register a marriage or register birth and death. In some countries, people have to register their cows and dogs. On each occasion, you pay as you changed the data and everyone who needs it gets access to data for free. I think that is the only sustainable mechanism for a lot of funding. There are bits you cannot fund that way but you would fund most of the dynamic data in SDI that way. For data that changes less often, there are other means of funding it. For example, an environment agency can fund the water model because it can't do without it. If you identify a government organisation that cannot carryout its function without this infrastructure, it can be funded.

Majority of national SDI initiatives all over the world have not been suc- cessful. Do you think local SDI is more practical than an SDI at national or global level?
I think it makes most sense for data collected be as close to its point of origin as possible. So I think local SDI is good in that sense. What we don't want is local SDIs being incompatible to upload their data to NSDI. For example, we have a war going on in UK between Royal Mail, which does national addressing and a company working with local government taking addresses from 400 local authorities. The two have come to a compromise according to which we now have top down standards which apply nationally but local data collection, local responsibility and integration with international datasets. National agencies deal with quality assurance and while data is collected at local level. But then you have to allow locals to charge for planning applications and updating data. You can't have a situation where locals give to national for nothing and then has to buy data back. That will never work. I can't see any local authority accepting it in the long term.