User:Cplot/CitationProposal

Below is a working draft for a proposal to clarify the WikiMedia software and Style policies surrounding notes, citations, sources and references. I welcome comments on this draft on the associated talk page. Please refrain from adding comments to the proposal itself. I welcome minor edits to fix errors, but please begin substantive changes through comments on the talk page.

Possible solutions for notes, citations, and sources (and references)
First let me appoligze for the long and complicated post that follows. This is my own attempt to think systematically about these problems that have burdened me and clearly others for a while. I just read through the thread "Inline references suck!" and from that and my own experience there's certainly many issues that need resolution. Inherently, I think this will need to be solved by the software developers, but they'll need and want input from article editors. So it's useful to have these discussions here, but only if we come to some agreement and make feature requests of the developers. So there are two things I think we should try to accomplish here:


 * Specify prudent style guidelines for the software as it exists now
 * Decide how the software could better meet the needs of the article editors and article readers

Differntiating the elements
Both of these methods (citebook and &lt;ref&gt;/Cite.php) have advantages and disadvantages. Both methods try to separate presentation from content. But they also both confuse citations, notes and sources. I understand notes as text incidental to the main discussion that would normally be presented parenthetically or as footnotes or endnotes. By sources I think we mean a book, periodical, artwork, web page or some other medium that provides the authority for the statements made in the main text and gives credit for the ideas contained there. These are typically represented in a list at the end of an article. Semantically each source can be reprsented by as little as a URI or some other unique identifier like an ISBN or ISSN. A citation links the text (whether the main body or the incidental note) to the source in a specific way (with e.g., a page number or id/anchor reference in a web page). And finally, a reference relates the overall article to the source in a particular way (as in an annotated source reference).

I see both methods, bookcite and Cite.php, as primarily designed for "note" elements as opposed to citations or sources. They may be used for citations and sources, but they don't meet the needs of those elements very well. Here are some examples that begin to separate these elements in order to think about proposed changes to the MOS and the underlying software:


 * a note may be just a citation: e.g., (Ptolemy 1998, p43).
 * a note may contain more than a citation: (many astronomers catalogued the position of stars. See Ptolemy 1998, book vii-viii).
 * a citation may contain a note: (Ptolemy 1998; Note in particular Ptolemy's listing of the constellation Ophicius in the catalog).
 * source references may also contain notes
 * &mdash; Ptolemy, Claudius (1998). The Almagest City:Publisher (This work is considered one of the earliest comprehsnive catalogs of naked eye observations of stars).


 * Citations inherently contain sources: (Ptolemy 1998 p43) contains a source: The Almagest ISBN-0852291639.

It is even more complicated because: These issues are best handled through the restraint of the article editors. One level of nesting may be acceptable, but beyond that the article loses coherence: the notes subtract rather than add clarity.
 * Notes may contain lists of citations
 * Notes may (though probably should not) contain nested notes (e.g., footnotes to footnotes)

However, the difficulty arises in balancing the need to cite to provide authoratative support and give credit with the difficulty of maintaining and reading so many citations. Through our collective efforts, sometimes the citation of one source offends another editor who wants to see the source they're familiar with cited: and so may add another citation to a growing list of citations. This growing list of citations may even be useful to some readers, but it doesn't help to have it so prominently displayed for all readers.

So from this analysis, let me make two sets of proposals for group discussion. One for revisions to MOS guidelines. And the second for feature requests for the software.

Suggestions for MOS guideline revisions

 * 1) should mainly encourage both parenthetical notes and parenthetical citations unless they become too unwieldy
 * 2) should use a hybrid of Harvard style parenthetical citations and bookcite/Cite.php citations (depending on the method selected for the article; especially if and when the list of citations grows too long for a parenthetical)
 * 3) should use a either a parenthetical notations (for simple notations) or bookcite/Cite.php notations
 * 4) should list both bookcite and Cite.php as the two recommended methods for citing sources.
 * 5) should discourage changing an article from one method to the other without cause and without first establishing agreement on the talk page.
 * 6) should recommend editors only cite two or three sources Harvard style and then place any additional citations in notes (using the article's selected method: bookcite or Cite.php).
 * 7) should discourage multiple anchors to the same citation/note since neither system currently handles this very well  (at least not from a reader's perspective who may have trouble jumping back to the main body of text where they left off)
 * 8) should encourage manual creation of source lists ("references" or "books cited") since neither system does a good job of pairwise matching of sources to citations
 * 9) should highlight the advantages and disadvangates of each system: bookcite and Cite.php

Suggestions for sofware feature requests

 * 1) The &lt;ref&gt; tag, though poorly named, is probably a move in the right direction for notes. A new element name such as NOTE or NOTATION with a corresponding tag (e.g., &lt;NOTE&gt; or &lt;NOTATION&gt; or whatever isn't already in use here in wiki or elsewhere) should be implemented. The element should allow some nesting for flexibility (though we may want to discourage it beyond some agreed upon level).
 * 2) Another element for citations that can contain notations or be placed within notations should be created.
 * 3) An attribute (e.g., source="ISBN:0852291639") should be included in the notation element to link citations to sources.
 * 4) For sources that cannot be automatically recoved through lookup and must be entered manually in the article, a wiki specific value could be used (e.g., source="wiki:001)
 * 5) Wiki software should automatically sort the reference/source list upon building the page.
 * 6) Wiki software should automatically insert three m-dashes (&mdash;&mdash;&mdash;) before sequential sources by the same author(s)
 * 7) This same attribute on a notation could link the notation to a source. This could be presented as notations in the list of references/sources to create an annotated bibliography.
 * 8) Wiki should incorporate a lookup mechanism to match ISBNs, ISSNs, URIs, etc and automatically generate the reference lists: caching sources in case the lookup later breaks.
 * 9) Other sources that do not have standardized method of identification like the above would still need to be added manually.
 * 10) For other sources that don't have easy lookup mechanisms (other than ISBN, ISSN some URLs) Wikipedia could offer a simple HTML form for generating cataloging sources (e.g., newspaper articles). Editors, would thus be removed from the cumbersome and error prone task of source editing cite templates. The form could generate a unique catalog number for the reference (checking for duplicates upon entry) to cite in any Wiki artile (a bit ambitious I admit).

I think these feature requests would go a long way in separating presentation from content, create much more manageable articles and allow greater flexibility to readers in how they prefer to view these elements. For example, eventually the reader could control the breakpoint between parenthetical citations or notations and footnote citations or notations. A printed version of wikipedia could do likewise.