Frequently Asked Questions

Where can I send complaints, questions or compliments?

You can contact me by email at Tim Brody, or further information about eprints can be found at the EPrints project.

What is this listing?

We are promoting open access to the research literature pre- and post-peer-review through author self-archiving in institutional eprint archives. Open access to research maximises research access and thereby also research impact, making research more productive and effective.

This registry has two functions: (1) to monitor overall growth in the number of eprint archives and (2) to maintain a list of GNU EPrints sites (the software Southampton University has designed to facilitate self-archiving).

How do I submit my site? (which sites are appropriate?)

To suggest a site follow the instructions at Suggest an Archive. Please submit sites that are (1) running GNU EPrints (because it helps us to know who's using the software) and/or (2) sites that archive open access research articles (because we'd like to monitor how well Open Access is doing overall). If the site is not an institutional or central archive but (3) the archive of a specific journal or journals, please specify this explicitly (so we can compare the growth of open access via the self-archiving of conventional journal articles versus publishing in open access journals). (Other categories are thesis/dissertation archives, database archives, demonstration archives (not yet operational) and "other").We ask for an administrator's email address in the event we need to get a clarification or a site is moved or removed. Your email address will not be published.

Why hasn't my site appeared?

Suggested sites go into a buffer and are added at the editor's discretion (to see 'dead' sites, which include rejected ones, see the dead sites listing). Some are duplicates, some are nonfunctional or inappropriate sites, some are web spam.

What does full-text/PDF/MS-Word/Research Papers mean?

The Full-text percentage is an estimate figure provided by the archive administrator or an editor. The estimate shouldn't be treated as anything other than an indicator (it's either unknown, 0, 25, 50, 75, or 100).

The PDF/MS-Word and Research Papers estimates are based on an automated process that has checked 1000 samples (or all records if the archive is smaller) for an attached PDF or MS-Word format document, and performed a test on those documents to determine whether it looks like a research paper. The test consists of looking for a sequence of years in the document (i.e. something akin to a reference list). There are many factors that will cause this automated sampling to fail:

• Documents hosted on a server with a domain name different from that of the OAI interface are ignored (this is to avoid counting journal links).
• HTML documents aren't counted due to issues of telling the difference between an HTML document, and metadata-only pages. Postscript-format documents aren't counted.
• PDF and Word docs that can not be converted to plain-text can't be identified as research papers.

How do I get a parseable list of sites?

You can download the records in Dublin Core format from the OAI-PMH interface. Just the OAI base URLs can be accessed via a ListFriends listing.

To get the full record (except for email addresses and internal data) use the plain text listing.

Plain-text listings are in RFC header-style format: multi-line values are prefixed by a single space, and records are separated by a single, empty line.

If you require more detailed data please contact me specifying what you need ("how do I download ROAR" type emails will be ignored).

How do I use the record graphs on my site?

Click a mini-graph to get a larger version (well, actually the size is specified in the URL). The graphs are generated on-the-fly from Celestial's holdings, which are updated nightly UTC. If you use these graphs on your site please set up a 'Cron' job to periodically create a local copy, rather than linking directly to this site. n.b. the graphs are in 'PNG' format.

Arguments to the Datestamps Script

format	This should always be 'graph'
baseURL	OAI BaseURL as listed in Celestial
width	Width in Pixels
height	Height in Pixels
mindate	Earliest date to show in yyyymm format, or 'auto' for first record
max	Maximum y-axis value, or 'auto' for the maximum record count
title	Title String (optional)

How do I search this listing?

A basic search function has been added, although given the paucity of data that this listing contains if you want to make a general search you will probably be better off using a Web search engine (e.g. Google).

What does Not registered in Celestial mean?

This means the archive has not been listed/harvested by Celestial yet. This may be because the archive doesn't have a functioning OAI-PMH interface or - for new entries - hasn't been added yet by an editor.

What does OAI Interface Unknown mean?

Either the archive doesn't have a functioning OAI interface, or we couldn't track down where it is. Site admins should say on their 'about' or 'help' page where their OAI interface is and use a common URL for it (e.g. /perl/oai or /cgi-bin/oai). Please always include a functioning OAI interface when registering your Archive and contact Tim Brody to fix a missing OAI interface.

What does Error in Records mean?

There may be no records in the archive, no OAI-DC records, or the datestamp may not have been exported in the record headers.

The registry uses OAI-DC records harvested by Celestial. The datestamp as given by the archive is stored in the 'provenance' part of the record, and it is this datestamp that is used by the registry to plot the number of records over time.

You can use the Celestial status page to identify potential problems.

What does Thumbnail Unavailable mean?

Most likely the site has not yet been visited by the automated thumbnail tool. While the tool does its upmost to create a thumbnail, missing links, redirects, and slow connections will cause it to fail (resulting in either no thumbnail or a grey expanse).

Legal Things

This listing in no way endorses or recommends the content of the sites listed. You are free in whole or part to link to, view, copy, or reproduce this listing without restriction. If you wish to attribute this listing please use "Registry of Open Access Repositories at the University of Southampton". If you have space it would also be nice to credit me (Tim Brody).

When linking to this site please use http://archives.eprints.org/ - no guarantee is made that links to pages further in will persist! The code behind this listing has been written by Tim Brody, based on the GNU EPrints listing by Chris Gutteridge, with input from the rest of the crew (Stevan Harnad, Les Carr, Steve Hitchcock). Running on the usual Apache/PHP/Perl goodness.