This document is the Basel document. Its initial version was written by Thomas Krichel on 2006-01-03.
This paper is the requirement document for phase 2 of ACIS. It is not the first one. There is an earlier piece, the Helsinki document. The Helsinki document deals exclusively with citations. It does so quite competently and therefore is still valid as a part of the requirements of stage 2.
In the meantime, new requirements have arisen.
Recent work by Thomas Krichel on the Konz project has demonstrated that it is possible to used search engine technologies to find freely available versions of online academic documents.
More generally, we expect that academic documents will become more and more dispersed.
An initial stage of this dispersion trend is that there is one authoritative version, with a number of draft versions. Often the authoritative version is not in public access but the draft versions may be.
We assume that a document, for which we have metadata in an ACIS based installation, is available in the form one one or more full-text papers. It is assumed that a paper is only contained in one file. Therefore papers can be idenified by their URLs.
An ACIS installation may already hold authoritative metadata for different versions of a very similar document. However, these will all be treated as different documents by the ACIS installation. But as a consequence, the same paper may appear as a version of different documents. It may also be accepted several times. If two documents have the same paper as versions to them, the two documents are related. ACIS does not further process the relationships between documents.
In the normal running of an ACIS installation, the papers will be assumed to be in open access. The system will present them through their URLs. Checking those URLs will not be part of the ACIS system.
When entering a version recognition interface, the user will be presented with papers that may be full-text version of a document described in the metadata. The full-text version is represented by its URL, which can be followed to (hopefully) view the candidate full-text to for examination. The user can then check if the paper is a full-text of the document. Rejected papers will no longer be presented to the user.
While there may be a requirement for other user choices, besides "is a version" and "is not a version", they will not be tackled because they are too difficult.
The recognition of the full-text of a file is quite similar to the identification stage described in the Helsinki document. The interface for both can therefore be quite similar.
When an ACIS service offers version recognition, each registered author will have a document location profile.
In terms of representational issues, the version recognition
profile does not concern personal data, but
document metadata. The input is of the form
<text id="
id ">
<hasversion>
<text>
<file>
<url>url_1</url>
</file>
</text>
</hasversion>
<hasversion>
<text>
<file>
<url>url_2</url>
</file>
</text>
</hasversion>
<hasversion>
<text>
<file>
<url>url_3</url>
</file>
</text>
</hasversion>
</text>
The output is of the form
<text id="
id ">
<hasversion>
<text>
<acis:authoraccepted date="yyyy-mm-dd"/>
<file>
<url>url_1</url>
</file>
</text>
</hasversion>
<hasversion>
<text>
<acis:authorrejected date="yyyy-mm-dd"/>
<file>
<url>url_2</url>
</file>
</text>
</hasversion>
<hasversion>
<text>
<file>
<url>url_3</url>
</file>
</text>
</hasversion>
</text>