Report of the Workshop on implementation of the Common Open Source Tool (COST)
31 participants from 15 countries attended the meeting, bringing 15 case studies to be analysed by COST. It is to be noticed that not all countries brought case studies, be-cause of problems in exporting their datasets or inconsistencies with the information required for the importation to the COST data structure.
The agenda of the meeting was divided into periods of presentation immediately fol-lowed by practical sessions.
The Bayesian approach and the simulation package were presented. The simulation package is a tool for validating and optimising the sampling strategies. As such, the tool is developed and available, but the considerable work of validation and optimi-sation remain to be done, and this is not reserved to the COST team.
During the discussion on the tool, its use and future developments, a lot of recom-mendations were made, among which:
• Improve the documentation of the methods in the help files and in the user manual;
• Improve and create error messages, everywhere they are too generic;
• Continue debugging the code;
• Extend the Bayesian package to other sampling strategies;
• Maintain the mailing list;
• Make the tool user friendly;
• Create a FAQ section;
• Validate thoroughly the proposed precision methods;
• Create templates for processing data;
• Include survey data and calculation of abundance indexes/
The majority of these recommendations argue in favour of a COST 2 project.
More pragmatic and small changes were suggested and noted for an implementation in 2010. These changes will be listed in the WKCOST report, and a message will be given to all COST users when new releases of the packages will be available.
It is recalled that the procedure to subscribe to the COST users mailing list is to fill the form on this website https://lists.berlios.de/mailman/listinfo/cost-project-users
In conclusion, out of the 15 case studies analysed during the session, all of them suc-ceeded in the importation, exploration, validation and consolidation of the datasets, 12 succeeded in estimating a length and age structures, 7 succeeded in estimating biological parameters and 6 succeeded in estimating discards volumes. The decreas-ing figures are mainly due to the unavailability of the relevant data for the estimation.
