Archive Maintainer Guidelines
The Unified
Computer Science Technical Report Index (UCSTRI) catalogs
resources available by anonymous FTP from many various FTP sites.
Creating and maintaining this service has proven a bit of a headache,
because each site has a different format for its local index file
cataloging resources. This also presents some significant limitations
on the service:
- The format is somewhat ugly for some sites.
- It is not possible to allow for sophisticated markup (e.g. use the
<CITE> tag of HTML and the <AUTHOR>, <PUB>,
<EDITOR>, etc. tags of HTML+.
- It is not possible to allow more sophisticated queries, such as
searches by author or by title, because it is not possible in general
to determine which part of the record contains these data.
- It is not possible to sort the reports into chronological order
for the same reasons; it's not always possible to figure out where the
date is (if it's even present.) Currently, the file modification time
is used as a heuristic of publication date.
- There are some things that are simply wrong; for example, many
index files also contain a brief tutorial on how to use FTP which
contains names of actual TRs, creating spurious entries.
Go on to a typical index
file...
Marc V