From: Neil J P. <ee...@el...> - 2002-03-03 21:41:16
|
Arne Pallentin wrote: > I can see the advantage of using a unified datastructure for all library > formats, what I can't see is the advantage of another text based > database (as I understand xml would be). Would be the change to a "real" > database structure be to complecated? (I'm obviuously no programmer) As > an example Jan Muellers Kaspaliste (http://site.voila.fr/janmueller/) is > not to bad. This is at least true for the table structure. The program > uses PostgreSQL as backend, but an even more open aproach would be > probably desireable. I don't think that a DB approach to storage should be the *only* option; could pyblio not use a generic API like gnome-db or unixODBC? (I don't know much about them, I must add). While a DB for storage and searching might be smaller and faster, I'd prefer the option to store in plain-text if just since this allows 'hand-hacking' of the data, eg. for extraction of remaining data if the 'db' is corrupted. With a general 'reference storage' API, this could even be extended to deal with import from remote archives - there seems to be a parallel between (for example) merging references from a local file (eg in bibtex format), and merging those from a remote 'file' from a ref-search server. Do people agree? -- Neil |