Discussion Directory UMM :Data Elmu:jurnal:A:Aquacultural Engineering:Vol23.Issue1-3.Sept2000:

8. Discussion

During the last 5 years Fish-Vet has been deployed worldwide in various settings. It has proven itself as an aid to diagnosis, a valuable companion for the field veterinarian and as an educational aid in veterinary schools. The main difficulties with the program can be divided into external and inter- nal. External problems include the fact that, unlike terrestrial animal diseases, the whole fish disease field is in a state of constant flux. New diseases are discovered continuously as new species are commercially developed and old spe- cies are raised in new geographic areas. These problems can be overcome by ongoing updating of the information in the program. Another problem arises from the fact that, especially in aquaculture settings, we are dealing with a large number of animals. Individual animals will probably exhibit different stages of one or more diseases at the same time. Therefore, sampling is essential to reaching a diagnosis. Internal problems are a result of deficiencies in the diagnostic engine itself. It is clear that we cannot hope to easily achieve best-of-breed results across the board. Any program limiting itself to a subset of problems can conceivably reach better coverage and accuracy. This has to be weighted against the advan- tage of having a single program to use for any situation. An added problem with partial implementations is that they require some pre-diagnosis processing in order to decide whether they are suitable for the problem at hand. This precludes their use in any non-expert setting. The development of Fish-Vet has shown us that we can reach our aims by using a ‘mix-and-match’ approach, foregoing commitment to any one technique. We believe that in the future we will see expanding use of hybrid systems, a trend that has already started. This allows the creation of programs with the advantages of all methods and almost none of the deficiencies Molnar et al., 1993; Ikeda, 1996; Lopez and Plaza, 1997; Reategui et al., 1997. Within the Fish-Vet program there are two other modules not discussed here. One is a bacterial identification module using biochemical profiles and the other a parasite identification module. Both should be used to confirm and support the diagnostic process, especially because at times this may be the only way to reach a diagnosis at all.

9. Conclusion