Jump to content

Disambiguating parentheses

From Meta, a Wikimedia project coordination wiki

Friday, June 15, 3:40 PM -- Soon, we will be able to disambiguate topics by using parentheses. It's worth starting to think about some general principles about when to use parentheses, and what to put in the parentheses.

Let me work with some examples, and draw some principles from how I would decide them. en:Apple can mean either the fruit or a kind of computer. But probably, it would be wisest to let en:apple mean the fruit and include a link on the en:apple page to en:Apple Computer.

Here's a similar example, using the recently-edited article en:icon. Originally, some programmer type, in all seriousness (and doubtless without meaning any harm to anybody), innocently went about writing the article as though there were no such things as religious icons and that "icon" unambiguously meant a programming language. In actuality, en:icon in its religious sense is probably the very most important sense, and nobody but programmers gives a rat's patoot about Icon-the-programming-language (with all due respect to the programmers!). So we might just let the en:icon page discuss the religious sense--rather than putting it on en:icon (religion)--and on the en:icon page include links to en:Icon (programming language) and en:icon (computers) or en:icon (computer jargon) (or something like that).

I believe that consideration of these cases supports the following principle:

When there is one central, best-known, most common meaning to the term "xyz" ("xyz" here is just a variable), and in that sense "xyz" names a topic about which we'll want an encyclopedia article, then we should put discussion of that topic on en:xyz with no parentheses, and from en:xyz we should include links to other articles under the heading "xyz" but qualified with parentheses, such as en:xyz (pdq) and en:xyz (abc).

Here's another issue that needs adjudication: given that we've decided we're going to use parentheses to disambiguate topics, what should we put in the parentheses? Take the en:icon example again. Suppose we've decided (for whatever reason) to make en:icon just a pointer page, and we needed parenthetical qualifications for our three senses of "icon." How do we decide what to put in parentheses?

I might choose en:icon (religion), en:icon (programming language), and en:icon (computer jargon). Why?

In the first sense, I wouldn't choose en:icon (religious symbolism), because "religion" is probably going to be useful in disambiguating many other titles. Another possibility is en:icon (art), I suppose. Again, though, my favorite for the religious sense is just en:icon, because the religious sense is after all the most basic sense.

I might choose en:Icon (programming language) (much like the already-existing en:Icon programming language), rather than en:icon (computers), because the latter is too broad: there are at least two things that "icon" (or "Icon") means when discussing computers (the language, and the little clickable image). I also wouldn't use en:icon (language), because "language" here is insufficiently precise: when presented with the word "language," what most people immediately think of is an ordinary, natural language like English or French.

Finally, I might choose en:icon (computer jargon), because that's something the nonspecialist could understand. I imagine there is some piece of computer science jargon to describe the general sort of thing an icon is, and you could title the article en:icon (that-piece-of-jargon), but that wouldn't be clear to the person who, after all, needs to know what "icon" means. Besides, "computer jargon" is probably going to end up being a very useful way to qualify various titles; think "disk," "processor," "processing speed," "mouse," "keyboard," etc., etc.

So there are a few principles/remarks that seem reasonable to me:

  • In general, put stuff in parentheses that is both easy to guess (to expedite "accidental linking") and that successfully disambiguates titles.
  • Use qualifying words that are used to qualify other article titles as well (to expedite "accidental linking").
  • If at all possible, keep the qualifying words simple and clear, so that nonspecialists can understand them.
  • So, the qualifying phrases are probably going to name rather general categories that an article's subject falls under, but not so general as not to be able to distinguish it from the other subjects named by the title.

Well, that's my first stab at these issues. I'm sure others will have lots to say too.

--Larry_Sanger