done with base data analysis plus an informal exploratory
 (sondeo) survey and a one-shot formal survey? Or is a detailed
 twice weekly formal survey required for-a period of one year?
 Can modelling techniques help improve understanding-or does
 this come at too high a cost? In the testing stage, should
 farmers be selected that are the better farmers, most cooperative
 farmers, or simply representative farmers? Representative
 farmers may not, for example, be so cooperative, thereby reducing
 the efficiency and effectiveness of dialogue and the timely
 conclusion of the testing stage. Considerable controversy still
 exists concerning the way in which these and other questions
 should be resolved in the interests of minimizing costs and time.
 (b) Finding ways to maximize the return from the location specific
 nature of "downstream" FSR by determining the transferability of
 the results to other similar "total" environments. Introducing
 some flexibility into the improved practices increases the
 potential of transferability, but this may come at some cost in
 terms of the potential level of return. Is this,or is this not,
 desirable? Controversy exists with respect to this.
 (c) Seeking best of readily available solutions-that is "better, but
 not necessarily best" or "non-perfectabilitarian." How much
 fine-tuning should there be, thereby extending the testing stage?

 6.6 In terms of developing improved practices (technologies), should
 emphasis be placed on single trait innovations,which may preclude the
 exploitation of possible complementary or synergistic effects between
 the various components in packages of improved practices? In theory
 the former would be desirable, but in practice the latter are much
 more common. A possible compromise is to design and develop packages
 of improved practices that permit, in an explicit manner, a stepwise
 approach to the adoption of the various components of the package.

*7. Function ns

 7.1 The location and linkages of the FSAR program will determine whether:

 (a) FSR, or helping to generate relevant improved technology, is the
 primary focus.
 (b) FSP, or helping to develop relevant agricultural policies/support
 systems, is the primary focus.
 (c) If linkages are good between research and planning/implementation
 institutions, some mix between the two may be possible, i.e., FSR
 plus FSP = FSAR.

 At the moment, the major thrust is on developing relevant
 improved technologies for the crop sub-system. This is, in part,
 because most FSR-type programs are located in agricultural research
 institutes whose mandates revolve around crops.

 7.2 FSAR-type programs help generate and communicate information in the
 following ways:

 (a) Through providing a means for farmers to communicate their needs
 to researchers (FSR) and planning/implementation agencies (FSP)