Dear SRS users/maintainers/programmers,
as you all know Thure Etzold's SRS allows you to search in an
increasing amount of databases. A couple of weeks ago the problem arises
that we installed a new (sub)database and I thought it would suffice
to copy the corresponding SDL/odd definition, compile this stuff and
build the indices finally. Unfortunately `hyperlink.sdl' and the
`prodom.sdl' didn't work with all the other sdl's used here (version
numbers didn't match also, the old problem of synchronizing
It seems to me, that with the increasing interest in SRS and more
programmers writing their own sdl definitions, it becomes really
difficult to incorporate new definitions in a "running" installation.
Thure also confirmed that the situation tends to develop in an
"chaotic" direction. Obviously there are no separately maintained
srs-source-xxx.tar and srs-odd.xxx.tar archives available.
What do you think about the idea to maintain such an archive to
prevent the diverging of the definition files?
One could collect all sdl/odd definitions in addition with an
(SHORT) documentation file and kind of a changelog (author,
version, like all these nice gnu programs)? The local
dependencies, like file paths, which are defined in srsdb.sdl
could go into a srsdb.new.sdl to prevent the accidental
overwriting of an existing one. Once in a while one could release
a complete set of definitions and make them available per ftp and
W3 (isn't this U6?), and SRS form itself.
Has anybody have an idea how many scientist were using and how many
sites were maintaining SRS in the world?
Waiting for your comments, suggestions, preferably no flames,
Dr. Jorg Sprengel
Institute of Genetics sprengel at scan.genetik.uni-koeln.de
Weyertal 121 Tel.: x49 221 470 3405
50931 Cologne FAX : x49 221 470 5163