| [vheissu] | [criticism] | [biblio] | [bio] | [misc] |
The Persons table contains 4 fields: first, middle, last name and generation. This last field is also used for a remark like [sic], a nick name like [Skip] and a fake 'title' like Professor (as in Professor Irwin Corey). There are 6 different entries in the generation field; one needs an extra exception. When the first name is not known, a dot is added in the form [this is to circumvent a British habit for not always registering the first name]. The middle name neither has one, nor the generation field [which is the name of the database field and cannot be changed anymore, but it is rather ineptly chosen].
The basic form for showing the first person's name is last, first, abbreviated middle and generation. Here, however, we see the first of a series of language and/or country issues; [1] in Spanish or Portuguese, there are no middle names. Hence, these items can contain rather long names; [2] in most European countries no middle names are used in public life -this implies that it is better not to register these, unless the person is from the Netherlands where only initials are used.
Option 1: An article has only one author
| Possibility | Format | Example |
| 1. No middle name and no generation | last, comma, first | Hollander, Charles |
| 2. A middle name and no generation | last, comma, first, initial middle name | Krafft, John M |
| 3. A middle name and a generation | last, comma, first, initial middle, dot, generation | Smith, Mack L., Jr |
| 4. No middle name and a generation | last, comma, first, comma, generation | Piela, Albert, III |
Option 2: Multiple Authors
There are no articles on Pynchon with more than 3 authors. Hence, we first have to find out that, if there is more than 1 author, then how many? Depending on this search result, we first have to generate a string with the value " and " (for 2 authors) or a comma and a space [", "] for 3 authors. The process is somehow repeated for a third author. In that case, the query returns the string [" and "] , otherwise it returns nothing.
| Possibility | Format | Example |
| 1. No middle name and no generation | first, comma, last | Anne Battesti |
| 2. A middle name and no generation | first, comma, initial middle name, dot, last | Martin E. Fitzpatrick |
| 3. A middle name and a generation | first, comma, initial middle name, dot, last, comma, generation | [no ex.] |
| 4. No middle name and a generation | first, comma, last, comma, generation | [no ex.] |
The author tag is alwas closed with a dot and a space, unless the author is not known --not uncommon in the popular press;. another issue that needs to be solved is that this dot is not allowed to be generated when the last name already ends with one.
Persons and Their Functions
When an item is an article in a book, the query determines if this book has one [or more] authors or editors. This relates to a table Functions which contains a field with a short description of the function. The query thus retrieves in the latter case one or more names, following the rules described above this field and generates another comma and a space, followed by the short description [Ed.] on condition that the editor is known.
Four other situations need a description: [1] An article is a translation, [2] an interview, [3] a dissertation of the third cycle or [4] a journal has an issue with one or more guest editors:
[1] For a translation the same rules apply as for an editor - the Functions table however gives a string [Tr. ] and the query generates also "by " and, if necessary, " and ".
[2] In the case of an interview the interviewee is registered as the author. We have seen items where the inerviewee is bracketed, others where this is not the case. Furthermore, the interviewer has to be mentioned. This affects only a few items, but this issue still has to call for a resolution.
[3] Specific for a dissertation is that it has only one author, and each author has only one dissertation. Because of this situation, and the fact that dissertations are shown separately, [a list reflecting the status in the Summer of 2006 can be downloaded here] we have opted for following a first name - last name order, sorted, of course, by last name, first name.
[4] Journal's guest editors need to be shown [a guest editor is linked to an issue, which is a specific temporal instance of a publication]. Another issue that needs to be resolved.
The Titles table contains references to 11 other tables. This is necessary in order to qualify an item as detailed as possible; apart from that it has its own content. The main content is of course the title and the subtitle of an item. We face interesting questions with this most exciting table on several levels. Basically, a complete title consists of a main title and a subtitle and can be as long as twice 256 characters. It has to be opened and closed by a double bracket; there is also a sign between title and subtitle, depending on the language. Finally, there is the use of capitals, quotes and references to books and journals. What seemed at first sight pretty straightforward is in fact rather, uuuh, ... challenging. Here becomes clear what any database developer or programmer knows so well: by solving one issue, one always creates another one; in other words: there is a price to pay, and that price can be high. An example: imagine a study with articles about (and including) The Meritorious Price of Our Redemption, and the 3 editors called it The Meritorious Price of Our Redemption, by William Pynchon [1650] : Facsimile edition, this complicates heavily the smart sorting order problem: this is the only place in the database where is inserted an html tag in order to be sure that the title of the tract is shown in italics (there is no other solution for this but it breaks the basis rule of distinction form/content.
An item is not located in time. It is linked to an issue (or better, the other way round: an issue can contain one or more items). The issue table contains following fields: year; period (ranging from a day, over a month to a period spanning several months), volume, and number. The main problem here is that there are several combinations visible (ex. below are real):
| Year | Period | Volume | Number |
|---|---|---|---|
| 1979 | |||
| 1979 | November | ||
| 1997 | September | 16 | |
| 1990 | March, 15 | 37 | 4 |
| 1979 | October | 1 | |
| 1979 | 16 | 1 | |
| 1973 | 40 | ||
| 1979 | 3 |
3 remarks on these issues:
[1] The period field contains a day: the database raw data is in the format month, comma, day. The query renders this as day, month.
[2] Dissertations listed in the DAI: the format in the query is : DAI 63 [2002]: 951A.
[3] When there is no date available, the query does not return 'n.d.' (no datum) but 's.d.' (sine dato) in order to cover all languages.
No author: sort by title, subtitle
The first character(s) of title causes the generation of new strings
ASCII value = 39: ignore 1 character, unless there's a blank space in position 5, 6, or 7
ASCII value = 60: ignore 3 characters
IF there's a blank space in position 2, 3, or 4: ignore first 2, 3, or 4
Item is a book and title is empty (i.e. no chapter inserted): sort by author(s)/editor(s), title publication, subtitle publication
Title starts with one of these words - develop select ...case ...with... ends
| English | A, An, The |
| French | Le, La, Les, L', De, Un, Une |
| Portuguese | A |
| Dutch | De, Een, Het |
| German | De, Der. Die, Ein |
| Spanish | La |
| [^] | [about] |