This report has examined the approaches taken in pilot trials undertaken to test the feasibility of and problems likely to be encountered in setting up a computerized data bank of animal genetic resources. It has highlighted two different methods of data storage viz. 1) by a fixed, predefined set of descriptors, each of which must have an entry, or 2) by the use of codes whereby only data are implied by the value of the code associated with them. The consultant recommends the first approach but with a refinement whereby missing data occupy little or no disc space. This can be done by a number of software packages that recognize null fields. A further refinement available is that only data entered are stored i.e. trailing blanks are stripped off all free-format textual information. These two options together will considerably reduce disc space required. Cost of such software could be as little as US$ 650 per site and as high as US$ 5 000 per site. The consultant is confident that a package somewhere in the middle, around US$ 1 100–US$ 2 000, will meet all requirements.
Such software is able to run on a variety of microcomputers. These computers have memories from as low as 128 KB to as high as 4 MB, and can handle hard discs with capacities ranging from 5 MB to over 100 MB. Cost of these systems run from around US$ 6 000 to US$ 35 000.
The consultant feels that a satisfactory startup hardware solution could be bought for around US$ 7 000.
Several man-months of work remain to be done once software and hardware have been selected. This covers the definition of data for all species entry of textual descriptions, specification of reports and generally becoming familiar with the operation of the software.
It is possible with some of the more expensive software packages, to have all data definitions done before the package is purchased. This saves purchasing the software that actually sets up the definitions and screen formats etc. (i.e. you do not purchase the software that actually writes further software for you; you only buy the finished product) . The consultant advises against this, however, on the grounds that requirements inevitably change. It is more convenient and often cheaper to have local expertise available to do any redevelopment rather than to have to pay suppliers at consultant rates to came in and make changes (assuming you can get them again).
The other issue that has been examined is the question of a global site versus several regional sites. In spite of the higher setup costs for several sites, the consultant leans strongly towards a regional emphasis purely from a management point of view. The project is much more likely to succeed when users have a feeling of active involvement. It then becomes the project of the region rather than a nebulous world project. Results assume more importance. Data are local and relevant. Retrieval of those data is more efficient and faster. Locals become more aware of what the computer can do for them. It becomes a tool, not some machine thousands of miles away that eats up their data and never gives anything in return.
An option not really considered in this report was the extent and use of existing computing facilites. The brief of the consultant was to provide recommendations on the equipment to be used (among other things). Recommendations have not been specific in terms of manufacturers at this stage as it is not advantageous to do so. By the time the decision is made as to whether to proceed or not, many more products will be available. These will undoubtedly be cheaper than current products and may well be superior. Apart from the facilities at FAO in Bangkok and the Universidad Central de Venezuela (neither of which appeared suitable as a base for this project) the consultant has not determined what other facilities would be available. There would undoubtedly be several. It would be imperative that each of these sites ran the same software. The important information, though, is that many of the modern breed of microcomputers could do the job. More specifically they would need at least 256 Kb of memory, floppy disc or magnetic tape for backup, hard disc with at least 10 MB memory and potential growth to up to 80 MB. Any of the operating systems CP/M-86, MS-DOS, PC-DOS or UNIX and its derivatives would be suitable.
