Quantity datatype (tracking)


The quantity datatype allows to enter quantities. It has quantity value, lower, upper, and a unit.

NOTE: Quantities are assumed to be uncertain per default, exact quantities are considered to be an exception. The idea is that quantities are usually measures. Since this leads to confusion and discussion, here's the rationale for reference: When doing arithmetics with quantities, such as unit conversion, defaulting to "exact" values would lead to "false precision" in the output.

From https://en.wikipedia.org/wiki/False_precision:

"In science and engineering, convention dictates that unless a margin of error is explicitly stated, the number of significant figures used in the presentation of data should be limited to what is warranted by the precision of those data."

And from https://en.wikipedia.org/wiki/Significance_arithmetic:

"If a calculation is done without analysis of the uncertainty involved, a result that is written with too many significant figures can be taken to imply a higher precision than is known".

bzimport added a subscriber: Unknown Object (MLST).
bzimport set Reference to bz54318.
Denny created this task.Via LegacySep 19 2013, 11:56 AM
Event added a comment.Via ConduitOct 16 2013, 3:29 PM

I wonder if the unit should be controlled by a predefined code list. Otherwise we will have a multitude of units for the same quantity type, e.g. W, kW, MW etc. for power.

Alternatively, in stead of unity type, allowing selection of a quantity type like "time", "length", "area", "power", "currency" etc. where the unit is predefined to a non-prefixed unit, correspondingly to "s", "m", "m2", "W", "unspecified".

Lydia_Pintscher added a comment.Via ConduitNov 27 2013, 4:15 PM
  • Bug 52916 has been marked as a duplicate of this bug. ***
bzimport added a comment.Via ConduitJan 29 2014, 8:36 AM

physik wrote:

I saw a RFC. Is someone working on that?
Semantic annotation of formulae that contain units is part of my PhD topic... therefore I'm very interested in that topic and open for potential contributions.

Lydia_Pintscher added a comment.Via ConduitJan 31 2014, 12:40 AM

Status update: We have just released the first version of the quantities data type on Wikidata.org. It does not yet support units.

@Physikerwelt: feel free to get in touch with me via email :)

Liuxinyu970226 added a subscriber: Liuxinyu970226.Via WebNov 27 2014, 7:30 AM
Snipre added a subscriber: Snipre.Via WebNov 29 2014, 12:39 PM
Lydia_Pintscher added a project: Wikidata.Via WebDec 1 2014, 2:23 PM
Lydia_Pintscher removed a subscriber: Unknown Object (MLST).
Stryn added a subscriber: Stryn.Via WebDec 8 2014, 10:32 PM
Lydia_Pintscher moved this task to backlog on the Wikidata workboard.Via WebMar 12 2015, 1:06 PM
Ricordisamoa added a subscriber: Ricordisamoa.Via WebApr 2 2015, 11:33 PM
Kelson added a subscriber: Kelson.Via WebApr 24 2015, 2:05 PM
daniel edited the task description. (Show Details)Via WebMay 5 2015, 8:42 AM
daniel set Security to None.
MSGJ added a subscriber: MSGJ.Via WebJun 4 2015, 8:42 AM
Lydia_Pintscher added a project: Epic.Via WebJun 23 2015, 3:30 PM
Wolfvoll added a subscriber: Wolfvoll.Via WebJun 30 2015, 11:12 AM
Darkdadaah added a subscriber: Darkdadaah.Via WebJul 29 2015, 4:28 PM
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptVia HeraldJul 29 2015, 4:28 PM
jeblad added a subscriber: jeblad.Via WebTue, Aug 25, 12:56 PM

This is actually wrong "Quantities are assumed to be uncertain per default, exact quantities are considered to be an exception. The idea is that quantities are usually measures. " A quantity is exact, a measurement is inexact. You count the number of coins, and the number is exact. You measure the weight of the coins, and the weight is an inexact number with an error. That error isn't bound by limits but by an error function, which in the limit is an standard deviation.

daniel added a comment.Via WebTue, Aug 25, 3:17 PM

@jeblad: so we agree about everything except the names. Perhaps I should have used "Measurement" instead of "Quantity"? But a measurement is a specific event, that would also be misleading. "MeasuredQuantity" would be more exact, but it's a bit long.

I say the bike shed should be yellow!

Darkdadaah added a comment.Via WebWed, Aug 26, 1:35 PM

For what it is worth, "quantity" seems just fine to me. Both a measurement and a count can give a value for a given quantity. Also, a measurement is not only an event, but also its result, so there is no need for a "MeasuredQuality".

To be even more pedantic, there is also another kind of quantity: estimations. Those are not directly measured or counted; they are calculated using a model with known quantities (from which they inherit their precision), e.g. the age of the Universe is estimated at 13.798±0.037 Gy. In that case the error may come both from the measurement and from the model.

Add Comment