Hi there, I'm just new to the list, so hello to everyone :-) I'm currently trying to find out how to adapt BATs in MonetDB for a scientific project. What I want to do is to add an encoding to data as some kind of a native DB feature. I.e. that a BAT stores data in some encoded fashion to detect certain kinds of errors that alter data in memory. By that, talking e.g. about an int BAT (column), the data width would increase e.g. from 4 to 5 Bytes (just as an example) where the additional byte stores the coding information. As I see it there are for a first shot two possibilities to do that: 1) Take one atom (e.g. int) and adapt all BAT / atom access code which touches int BATs. 2) Add some kind of new atom which is handled in a specific way. Since I have just started looking into MonetDB I now hope for a few hints to decide whether one of these approaches is sensible and whether there is maybe a simpler way. Thanks a lot in advance for any helpful comments :-) Best Till Kolditz -- Dipl.-Inf. Till Kolditz wiss. Mitarbeiter Technische Universität Dresden Fakultät Informatik Lehrstuhl Datenbanken 01062 Dresden Tel.: +49 (351) 463-38077 Fax: +49 (351) 463-38259 E-Mail: Till.Kolditz@tu-dresden.de jExam Group www: http://www.jexam.de Mail: till.kolditz@jexam.de