Rules for decoding author & title fields?

[tfmorris]

Apologies for questions in the form of an issue, but I'm not sure where else to ask them. The title and author fields appear to have add special markup and/or transformations applied to them. What are the rules to recover the authors' names and the titles of the works?

Authors

• Charles I, King of England, 1600-1649. - appears to be almost usable as is, except for the dates
• Uncle Tim, 1782-1865. - single name with dates added
• Coleridge, Samuel Taylor, 1772-1834 - last,first with dates, no trailing period
• CURTIS, Walter. - last, first, with no dates, surname uppercased

Titles

• Lines on the Conflagration of Moscow. [electronic resource] - removing piece in [square brackets] appears to leave a valid title
• Book of Word & Music of Humorous Songs, Burlesque Sketches, & Recitals. [With a portrait.] [electronic resource] - ditto, although this example has two [bracketed] texts
• [Britannia rediviva: a poem on the birth of the Prince.] [electronic resource] - removing bracketed texts here leaves nothing

Presumably there are rules/code for applying the transformations in the first place and hopefully they're reversible so the actual information can be extracted again.

[Orangeaurochs]

The rules for this are AACR2 (Anglo-American Cataloguing Rules 2nd ed.) now in the process of being superseded by RDA (Resource Description and Access), both of which are sadly behind paywalls. AACR2 bases its punctuation closely on something called ISBD (International Standard Bibliographic Description), a version of which is available here: http://www.ifla.org/files/assets/cataloguing/isbd/isbd-cons_20110321.pdf See p. 68-69 for some punctuation patterns for titles.

AARC2 and RDA records are encoded in MARC21 (MAchine Readable Cataloging 21), so the publicly available MARC21 manual is a very good source of examples if not a rigorous statement of principles, although it does address many of them: http://www.loc.gov/marc/bibliographic/. In particular, see the 245 field (title): http://www.loc.gov/marc/bibliographic/bd245.html and the x00 (various personal author fields): http://www.loc.gov/marc/bibliographic/bdx00.html

With your examples above, there are a number of complications:

• No trailing period. Could be bad cataloguing or a system stripping it off. In some slightly different contexts (authority records) it is correct not to have one.
• Surname uppercased. Not standard AARC2, but tolerated in RDA.
• [GMD] in brackets, e.g. [electronic resource], should come after the title proper, and not after the period as in your example. You are right that if you remove it, it normally should leave everything else intact.
• [other things in square brackets]. Square brackets are heavily used in cataloguing to indicate information added by the cataloguer that wasn't on the book (or whatever). In the last example, there was either no title at all, or it was taken from somewhere else (a bibliography for example).

There will be many more such complications I expect, especially if you're dealing with non-book or rare book things. It's a minefield! 8-)