need to compromise between being "exactly wrong or approximately right".

3. The time and technical skills required for selecting indicators might make it difficult for decision makers and stakeholders to participate fully in the selection of indicators. At the same time, experts carrying out the assessment have the responsibility of ensuring that the selection of indicators and the assessment as a whole are technically and scientifically sound.

     Hence, in the area of indicators, a way must be found to maximize both the technical excellence of the assessment and the commitment of participants from government, civil society, and business.

     The focus of this assessment on poverty, sustainable livelihoods and sustainable ecosystems marks a clear trend that future agricultural development is moving away from the exclusive production focus of the past. However, indicators available today can support assessment of these broadened goals of agricultural development only partially: more efforts are needed to develop sufficiently appropriate indicators.

Units of analysis and reporting. The IAASTD uses indicators which measure at several scales, from individual to farm, nation, region and global levels. Numeric indicators use metric units while qualitative indicators are descriptive. Information from smaller units will be aggregated up to sub-global and global assessment levels. The results will thus be generic but presented in such a way that it makes sense to other units of analysis.

Dealing with systems

The IAASTD basically deals with two different sets of systems, a biophysical and a socioeconomic set. On the one hand, there is the biophysical set with the underlying ecosystem in which the agricultural system and the unit-based production system is established. Primary ecosystems have been altered to a greater or lesser extent by agricultural production systems that define themselves according to economic criteria of efficiency as opposed to the multifunctional character of ecosystems. Usually, forest ecosystems are converted into grassland for livestock rearing, or a system with bare soils for cultivation. Depending on the capacity and suitability of this new agricultural land, production takes place over shorter or longer periods of time, from a single or a few years to decades and even centuries on the most suitable land. Assessing the future of these production systems requires taking into account their current suitability, including the degradation of ecosystems or parts thereof which has taken place, and the potential of these land areas to support agricultural production of goods. In addition, the multifunctional character of ecosystems has to be considered as a crucial aspect important to societies and the global community.

     On the other hand, political, economic, social and cultural sets of systems shape human livelihoods and agricultural production systems in the different contexts in which the latter operate. A large disparity exists between these contexts. A majority of agricultural workers are poor smallscale farmers in developing countries, with a high degree of dependence on subsistence systems, i.e., production by

households for their own consumption, and a high degree of dependence on both the biophysical and socioeconomic systems. A minority of agricultural workers live on larger production units and in industrialized nations, profiting from wealthy economies and a variety of subsidies to maintain their production and/or production systems. Assessing the future of agricultural systems will require thorough analysis and evaluation of these different contexts and the livelihoods derived from them through agricultural activities.

      Many of these contexts and systems are evolutionary; shifts in parameters must be expected, and the state of natural and human environments will continuously change, be it through factors such as opportunity (e.g., new business options or access to new resources) or constraints (such as further decapitalization of small-scale farmers). The degrees of uncertainty are rather great and difficult to foresee.

Dealing with scales (spatial and temporal)

Assessments need to be conducted at spatial and temporal scales appropriate to the process or phenomenon being examined. Analysis of issues must take place across several spatial scales simultaneously because an analysis at a single scale will miss important interactions. For example, national policies embedded in a global system have an impact on local decisions regarding AKST. Moreover, vulnerabilities are related to various scales. A comparison of a larger scale poultry production system with a decentralized backyard poultry system reveals different scales. While an infection of the former system is relatively easy to prevent, a possible outbreak would be catastrophic. In the latter system an infection of the flock is harder to prevent while an outbreak would affect a smaller number of poultry. Most of the analysis in the IAASTD is carried out at national and regional levels, but informed by experience from ground realities.

     The IAASTD is structured as a multiscale assessment in order to enable its findings to be of greater use at the many levels of decision-making. A global assessment cannot meet the needs of local farmers, nor can a local assessment meet the collective needs of parties to a global convention. A multiscale assessment can also help remedy the biases that are inevitably introduced when an evaluation is done at a single geographic scale. For example, while a national AKST assessment might identify substantial national benefits from a particular policy change, a local assessment would be more likely to identify whether that particular community might be a winner or loser as a result of the policy change. For example, in contrast to privately funded research, where the donor derives benefits, benefits derived from public goods research does not go to the funding agency itself, rather to other members of society, and there is no direct incentive to do more (CGIAR Science Council, 2005).

Dealing with values and valuation

The IAASTD deals with two valuation paradigms at the same time. The utilitarian paradigm is based on the principle of human preference for satisfaction (welfare). AKST systems provide value to human societies because people derive utility from their use, either directly or indirectly. Within this utilitarian concept of value, people also give value to AKST aspects that they are not currently using (non-use values), for example people value education systems even