Wikipedia talk:Manual of Style/Dates and numbers/Archive 83

Recent edit concerning the point about consistent decimal places

I'm transferring a discussion between Coren and me from his/her talk page.

BEFORE:

The number of decimal places should be consistent within a list or context (“The response rates were 41.0 and 47.4 percent, respectively”, not “The response rates were 41 and 47.4 percent, respectively”).

AFTER:

Outside of scientific contexts where numerical error may be significant, the number of decimal places should be consistent within a list or context (“The response rates were 41.0 and 47.4 percent, respectively”, not “The response rates were 41 and 47.4 percent, respectively”).

I don't get it. Why is scientific context at issue here? Tony 04:29, 1 September 2007 (UTC)

Because significant digits and precision count. 41 is different from 41.0— the former means "some value in [40.5 41.5[" while the latter means "some value in [40.95 41.05[. In a scientific text, changing the number of significant digits changes the meaning. — Coren ^(talk) 05:32, 1 September 2007 (UTC)

But in a list of values, we expect that they'll be of equal decimal-place precision, yes? If the list is 36.8, 51.1 and 53 and 54.5, not only does it look odd, it raises the question as to whether the writer intended 53.0 or, as you put it, somewhere in the range 52.5 ≤ x < 53.5 (in most cases, writers do mean 53.0, and even scientists have been known to be sloppy about this. I don't understand why your text exempts scientific contexts where numerical error "counts" (and where does the boundary between counting/not counting lie?). The wording needs tweaking, in any case. Tony 01:15, 2 September 2007 (UTC)

I suspect that the wording should note that scientific contexts sometimes have requirements for explicit precision that will prevent consistency, rather than implying it is the norm. SamBC(talk) 01:34, 2 September 2007 (UTC)

The previous (and changed) wording concerns only the same list or context. I'm unsure why you'd want to change the level of precision for only one item. Can we have an example of where this might be the case; otherwise, I think the change should be reverted. Tony 04:49, 2 September 2007 (UTC)

Items that happen to fall into the same list are not inherently measured to the same precision. If I measure something with a yardstick marked in inches with no finer markings, and you measure some other things with a more precise ruler, our measurements have different precision. If our measurements were listed together, in some contexts it would be undesirable to bludgeon them into the same precision. Scientists tend to be more careful about significant digits than other people are. So yes, there are edge cases where this is a legitimate issue. In general, within the context of Wikipedia, I am not sure how often this would come up. — Aluvus t/c 05:05, 2 September 2007 (UTC)

• So can we solve the issue by inserting "generally"?

The number of decimal places should generally be consistent within a list or context (“The response rates were 41.0 and 47.4 percent, respectively”, not “The response rates were 41 and 47.4 percent, respectively”).

Tony 06:30, 2 September 2007 (UTC)

• It may be worth specifying this one category of potential exceptions to help prevent wikilawyering. It may not. I've just noticed that "generally" doesn't seem to hold much weight with the wikilawyers. SamBC(talk) 08:10, 2 September 2007 (UTC)

I have trouble getting my brain around this "one category", and why one item in a list would warrant different treatment. Generally is possibly a useful way of allowing editors a little latitude (i.e., do it unless you present a good reason not to). Perhaps someon can come up with an example of this elusive 41 and 47.4. Tony 10:00, 2 September 2007 (UTC)

Can we not make the exception explicit:

Unless they were measured with unequal precision, the number of decimal places should be consistent within a list or context (“The response rates were 41.0 and 47.4 percent, respectively”, not “The response rates were 41 and 47.4 percent, respectively”).

−Woodstone 10:41, 2 September 2007 (UTC)

• Or may I rearrange the text?

The number of decimal places should be consistent within a list or context (“The response rates were 41.0 and 47.4 percent, respectively”, not “The response rates were 41 and 47.4 percent, respectively”), except in the unusual instances where the items were measured with unequal precision.

Tony 13:55, 2 September 2007 (UTC)

Split into Wikipedia:Manual of Style (dates) and Wikipedia:Manual of Style (numbers)

Resolved

 – No consensus for split.

I’ve never really understood why dates and numbers are handled on the same MoS subpage. Glancing over the discussions, it seems different people are looking after these parts and much of the archives already has been divided. I’d like to split them up into Wikipedia:Manual of Style (Dates) and Wikipedia:Manual of Style (Numbers). Has anyone serious objections? Christoph Päper 09:31, 2 September 2007 (UTC)

Or perhaps Wikipedia:Manual of Style (dates) and Wikipedia:Manual of Style (numbers), in order to more closely follow the Manual of Style in other regards? ~ Jeff Q (talk) 09:39, 2 September 2007 (UTC)

Of course, sorry. Christoph Päper 16:13, 2 September 2007 (UTC)

I have very serious objections to such a major, sudden change without consensus. Remerge them now, please, and raise it in the usual way. Tony 09:50, 2 September 2007 (UTC) Ah, I spoke too quickly. My view is that there are too many MOS submanuals and that the more there are, the harder they are to coordinate. I see expertise here nicely combined in this coverage of dates and numbers. Can you provide an example of how the division would improve things? Tony 10:04, 2 September 2007 (UTC)

I hope the maintainability would increase. This talk page tends to get rather large, but the groups discussing dates and numbers have a rather small intersection. Christoph Päper 16:13, 2 September 2007 (UTC)

Why is this page called (dates and numbers)? It is also about units of measurement, not just plain old numbers. Maybe it needs to be split into three parts instead? —Preceding unsigned comment added by 81.179.80.238 (talk) 00:12, 3 September 2007 (UTC)

And it's about a lot more than dates, which is why the section is now called "Chronological items". Tony 01:24, 3 September 2007 (UTC)

If the majority prefers so, we could of course settle on Wikipedia:Manual of Style (chronology) and Wikipedia:Manual of Style (metrology) or some such. Numbers and numbers with units are close enough, in my opinion, to be handled in one page, but dates and times only connect to these through durations (e.g. “30 s”). Christoph Päper 16:53, 3 September 2007 (UTC)

Don't like metrology (too similar to meteorology, and sounds ugly). Unconvinced that the current structure is faulty; I like the way a critical mass of experts contribute here. Tony 22:21, 3 September 2007 (UTC)

Why? Because dates are numerical, perhaps? — SMcCandlish talk cont 09:51, 19 September 2007 (UTC)

It is also about currency, which is more than mere "numbers". 81.178.90.168 (talk) 2007-09-20 00:33 UTC.

That's right, it's the "Numbers in several guises" submanual. Tony (talk) 00:50, 21 September 2007 (UTC)

Removal of metric units

Please can people comment on whether metric units are permitted in Vessel monitoring system, Fisheries management and similar circumstances? Lightmouse 10:30, 2 September 2007 (UTC)

If the editor is not directly quoting the actual treaty itself then I see no problem including the metric conversions. It helps promote understanding of the subject matter. BTW, that editor changed nautical miles to just miles, which also should be changed back but I'll let you handle that. —MJCdetroit 13:35, 2 September 2007 (UTC)

Since, as noted above, nm and knots are acceptable in SI, why convert in the first place? Knots and nautical miles are standard in shipping and aviation, because of their direct relationship to the curvature of the Earth, even to countries that use the metric system. Secondly, even if the conversion is important enough to be made, from line 198, there's no need to convert 5 nm to km twice in the span of a dozen words; the conversion hasn't changed, and it's not like anyone would forget that quickly. Sacxpert 20:00, 2 September 2007 (UTC)

I wrote the original article for vessel monitoring system, and used nautical miles when they were clearly the usage of the source material, but miles when that was the specific language of a treaty. With nautical documents, it is usually a fair assumption that nautical miles are meant. Assumptions, of course, are dangerous, and adding a conversion to kilometers, to me, adds further ambiguity to situations where a legal document is not specific.

There were a few places where a speed was mentioned, such as the 5 knot limit above which a vessel is considered unable to operate a scallop dredge. Fisheries managers monitoring this would typically either see it on a marine radar, or based on calculations from successive GPS-based reports. Nautical navigation tools customarily report in knots, and that is the convention used by mariners from thoroughly metric countries. I really believe that the metrication is gratuitous in these context.

To turn to an aeronautical example, English is the official language of air traffic control worldwide, even if it were a German aircraft talking to a Brazilian controller. Quite a number of airlines require all cockpit personnel to speak to one another in English, based on an assumption that if someone were switching languages between the cockpit and air traffic control (or other aircraft), they might make a slip and use a language that the receiver didn't understand. In other words, there are strong safety and navigational reasons for using certain traditional units, accepted by the SI oversight organization. Saying the metric conversion is needed "to make it more accessible", I believe, can further confuse a reader who otherwise will not understand the discipline's conventions. In like manner, when I write or edit pharmacology articles, I use only metric units. Howard C. Berkowitz 20:43, 2 September 2007 (UTC)

A pedantic but important point: You will hear all sorts of languages being used in aviation. English is the common language rather than the official language, although the distinction is sometimes lost on people. ICAO merely says English be made available whenever an aircraft station is unable to make use of the language spoken by the controllers on the ground. Your German pilot and Brazilian controller example is true. However, there is no ICAO regulation that forbids Spanish controllers speaking Spanish to Spanish pilots in Spain. Just thought I would point that out because we should not reinforce the already popular misconception that English communication is mandatory in the air.

Saying miles when the source says "miles" is a good thing. Attempting to be more precise than our data is a bad thing; we should not convert when it is unclear whether miles or nautical miles is meant. (It does no great harm to say "200 miles" when either could be meant, but do we convert to (320 km) or (360 km)?) Septentrionalis PMAnderson 21:03, 2 September 2007 (UTC)

As of the 8th (2005) edition of SI, the table containing both the nautical mile and the knot is now entitled "other non-SI units", replacing the titles "units temporarily accepted for use with the SI" (6th) and "other non-SI units currently accepted for use with the International System" (7th) formerly used for that table, thus they are no longer accepted for use with the SI. Instead, the 8th edition states that authors should have the freedom to use those units, but users "should always give the definition of the units they use in terms of SI units." I assume this means a statement like "1 nautical mile = 1852 m" should be included, not that the units be converted. — Joe Kress 05:45, 3 September 2007 (UTC)

You forgot to mention that they use the symbol "M" for nautical mile and "kn" for knot. That doesn't mean that we should follow their lead, IMHO. Gene Nygaard 16:44, 9 October 2007 (UTC)

Let me make a suggestion here, which perhaps will need to be linked to a Wiki article that covers both "units temporarily associated with the SI" and a category that, for want of a better term (suggestions welcome), might be called "Wiki units for ambiguous situations." The first part will cover, among other things, the nautical/aeronautical units and why they are preserved (i.e., relationship of nmi/nm to latitude, not just tradition).

A section, however, might include the term (capitalization significant) "Miles", and indicate that this is used when the term "mile", without metric equivalent and without disambiguation as to Imperial or nautical, appears in a legal document. In the latter case, the point would be made explicitly clear that since the type of miles is not known, introducing a metric equivalent could give the sense of greater accuracy than is known. A technical question here: in making a Wiki link, would something on the order of "Non-SI Units Temporarily Accepted#Miles" take the reader directly to the subsection on Miles within a larger article?

Are there any other examples besides Miles where such ambiguity exists? Tons, perhaps? —Preceding unsigned comment added by Hcberkowitz (talk • contribs) 12:17, 3 September 2007 (UTC)

I can think of a number of (commonly used) units for which ambiguity exists. Examples include calorie, gallon, gill, horsepower, kilobyte, mil, ounce, ton and quart. Thunderbird2 20:06, 9 October 2007 (UTC)

And mole. Too bad the CGPM disn't have enough sense to use the kilogram mole in the first place (for consistency with the kilogram as a base unit) and to call it a loschmidt or something named after a person. Gene Nygaard 21:17, 9 October 2007 (UTC)

The gill (Thunderbird2's link above is to wrong article) is a "commonly used unit"? Commonly used with more than one meaning? Though I have seen it a zillion times in lists of units, but I've never actually heard anybody use it to express a measurement (and most Americans, at least, would likely mispronounce it if they did), and I've seen it used once or twice in a book, maybe by Henry Fielding (1707-1754) or someone like that. Gene Nygaard 07:00, 10 October 2007 (UTC)

Well, perhaps not any more. During my mis-spent youth it was a standard measure of volume in English pubs (for spirits, not beer). As I no longer frequent these I don't know if that is still the case, so I withdraw it from the list :). Thunderbird2 08:29, 10 October 2007 (UTC)

British pubs do not use the gill anymore. It was replaced by the ml in the 1990s I think. "Make sure there are no 1/6 or 1/3 gill measures still in use in the bar as they are illegal." Lightmouse 10:58, 11 October 2007 (UTC)

There's also the cable, carat, pound and rayl, some of which are admittedly less common than others. Oh, and (though these hardly need separate mention) megabyte, gigabyte & co. Thunderbird2 11:55, 11 October 2007 (UTC)