John Skoyles's (1992) recent proposal that psychological raw data be maintained on file servers as data archives is a good one and a bellwether for future use and provision of data. I wish to comment only on one aspect, that of data security: the data must be protected against modification by accident or intent, whether by its owner or by other users. Proposals of the kind Skoyles is making must from the beginning take the security and integrity issues into account.
1.2 Owner: The potential for modification after the fact must be eliminated so that the research community can have confidence in the archiving process. The malleability of electronic data is such that special precautions are needed that would not be necessary in an environment where only print data was involved. It is not that we expect data modification to take place; but confidence in the system will require that users be assured that it cannot take place.
1.3 Others: Accidental modification or destruction is a familiar possibility to all of us. Using the data files as primary source data must not be allowed; only copies. Well organized backups will be essential. Fraudulent modification must be protected against for reasons similar to those given above.
1.4 What I want to emphasize is not so much that there are immediate and complete solutions. There aren't (are we should proceed anyway): But proposals of the kind Skoyles is making must from the beginning take the security and integrity issues into account. If the research record is to be maintained, we must, as we see more and more such proposals, continue to ask that they give sufficient thought to preserving the data intact and protecting it from modification.
Skoyles, John R. (1992) Ftp internet data archiving: A cousin for PSYCOLOQUY PSYCOLOQUY 3(29) data-archive.1