Wikipedia:Featured picture candidates/delist/Using the caliper new en.gif

Delist:File:Using the caliper new en.gif[edit]

Voting period is over. Please don't add any new votes. Voting period ends on 26 Sep 2017 at 19:50:10 (UTC)

Reason: The animation is showing an incorrect way to read calipers. The proper reading here should be 2.470 cm not 2.47 cm and a reading of calipers should also always state the error associated with the reading. In this case the proper reading is 2.470 cm ± 0.005 cm. As it stands, this is a reading that did not use the full accuracy of the shown instrument and neglects the important error part of the measurement so crucial when using calipers.
Articles this image appears in: Calipers, Vernier scale,
Previous nomination/s: Original Featured image discussion
Nominator: Jason Quinn (talk)

Delist — Jason Quinn (talk) 19:50, 16 September 2017 (UTC)[reply]
Keep – The gif shows the basics. The .005 is instrument resolution, not accuracy nor error. Adding it will be misleading. Bammesk (talk) 01:23, 17 September 2017 (UTC)[reply]

Comments Several remarks: A) allowing something that "shows the basics" yet teaches something wrong would be a pretty low standard for our featured images, which are supposed to be the finest images we can offer. B) this image doesn't show the basics. It shows how somebody untrained might incorrectly read the measurement. The 2.47 cm reading completely ignores the fact that these calipers have a 20 lined vernier scale rather than a 10. C) As for error, I believe you are using "instrument resolution" in effectively the same way I am using "error" above. Technically, of course, I am referring to absolute error (which gives the range the true value must lie if the instrument is otherwise perfect. (This does not include other forms of error such as the zero point error and any systematic error from the instrument which would affect accuracy.) It is the absolute error that limits the precision of the instrument and it is under most circumstances and by default what is meant by "error" for the instrument barring systematics. No matter what, to suggest that giving an error on a measurement is more misleading than not giving any error is contrary to basic teachings of measurement. ~~Calipers usually even have an etching that give the "instrument resolution" so that would be one way of handling that in an improved image.~~ Jason Quinn (talk) 06:33, 17 September 2017 (UTC)[reply]

A) The gif doesn't show anything wrong. It just doesn't show what you suggest it should. B and C) The resolution is already etched as 1/20. There is no need to highlight it again. Doing so with a plus/minus sign in front of measurement reading is wrong. It would conflate resolution, accuracy, and error into one number (misleading and wrong). I don't object to highlighting the resolution (though I see no overwhelming need for doing so). I object to doing it with a plus/minus sign in front of the reading. Besides, with this type of caliper resolution is not a plus/minus thing. Bammesk (talk) 12:20, 17 September 2017 (UTC)[reply]

I had forgotten it does have a "1/20" for the resolution so I guess that's fine regarding the statement I've stricken above, although maybe "0.05 mm" or "0.005 cm" might be better. Your acknowledgement of 1/20 as the resolution contradicts your statement that the gif doesn't know anything wrong. The reading "2.47 cm" is not at that resolution. If a reading is not being at the full resolution of the instrument, rounding is involved and I fall to see how that's using the instrument correctly. It's like me stepping on my bathroom scale which has a kilogram resolution but only reading the tens place and ignoring the ones place. As for the plus/minus, that's standard notation in mathematics, engineering, and physics so I don't see why you are objecting to it. And the symbol is not "in front" of the reading but between the reading and the absolute error. Jason Quinn (talk) 14:03, 17 September 2017 (UTC)[reply]

Again, adding the readout resolution of a measurement instrument following its measurement using a plus/minus format gives the wrong sense of measurement accuracy. Bammesk (talk) 15:23, 17 September 2017 (UTC) Sidenote: introducing a self-defined terminology, absolute error, does not help the discussion. Nor does the false equivalency of reading 70kg as 7kg ! [reply]

Surely Jason Quinn's example was meant to refer to reading 72 kg as 70 kg (or 72.3 kg as 72 kg). --Paul_012 (talk) 07:02, 18 September 2017 (UTC)[reply]

Yes. That's exactly right. Or in my case if the scale says 95 kg (and is rated for plus/minus 1 kg) but I "read it" as 90 kg. Jason Quinn (talk) 18:56, 18 September 2017 (UTC)[reply]

Ok, I understand. I took it in the context of dropping a trailing zero. Bammesk (talk) 02:12, 19 September 2017 (UTC)[reply]

Tell me if I'm understanding the issue correctly. Jason Quinn is arguing that (1) the nominal reading here is incorrect because it includes too few significant figures (actually, looking closely at the image I think it looks more like it should read 2.475 than 2.470), and (2) including the reading error or resolution error is standard practice that should be shown in the image. Bammesk is directly arguing against the 2nd point, saying that it is being conflated with instrument resolution and "is wrong" (though this introductory course material doesn't seem to agree). Bammesk hasn't directly addressed the significant figures issue. --Paul_012 (talk) 07:31, 18 September 2017 (UTC)[reply]

1) You are understanding my position correctly. And, great eye! I agree. If you pause and zoom the image, it does appear that 2.475 is better! 2) As for the reading error, the "1/20" does imply that information but it would be much better if the error on the reading were made more explicit. The red text should display this information because when using calipers the error on the measurement is often just as important to know as the measurement itself. Jason Quinn (talk) 18:56, 18 September 2017 (UTC)[reply]

Comment: This seems like rather trivial issue that should be easily fixable by those who know how; the ~~error~~ disputed reading lies only in the final frame of the animation. (That said, a 10-lined vernier scale might actually be better for illustrating the concept.) --Paul_012 (talk) 09:23, 17 September 2017 (UTC)[reply]

I agree that a line instrument would be better for pedagogical reasons. Jason Quinn (talk) 14:03, 17 September 2017 (UTC)[reply]

Meant to say "10 line instrument" above. Jason Quinn (talk) 05:28, 23 September 2017 (UTC)[reply]

Paul_012, there is no error in the final frame of the animation. Bammesk (talk) 15:23, 17 September 2017 (UTC)[reply]

Clarified. --Paul_012 (talk) 07:02, 18 September 2017 (UTC)[reply]

Comment: Actually, it appears that the significant figures issue (plus changing the readout unit to mm) had previously been addressed by Jollyroger in the 11 June 2007 version of the file, but was reverted by the original creator. --Paul_012 (talk) 07:50, 18 September 2017 (UTC)[reply]
Comment: As pointed out by Paul_012, the animation isn't even using the most "lined up" line on the vernier scale. At this point, I think this is a no brainer to call. This image is not featured image quality and would need to be modified to be considered. This is besides the resolution issue, which I also think is a deny level problem. Jason Quinn (talk) 19:07, 18 September 2017 (UTC)[reply]
Reply – The line up at 2.470 is off by "one third of one pixel", not easy to measure and not enough to fail FP. Paul_012, for what I am saying (as you hinted above) revisit my replies above, and keep in mind that readout-resolution and measurement-accuracy are not the same thing. An example: [1], readout-resolution=.00005, measurement-accuracy=.0006. Attaching +/-readout-resolution to the outcome of a caliper measurement is misleading. About the trailing zero, I prefer 2.470, but 2.47 is not a show stopper for me, after all it is a zero and readout-resolution is marked on the instrument. Bammesk (talk) 02:12, 19 September 2017 (UTC)[reply]

1) That misalignment, small as it may be, is critical to this image. 2) This image is teaching how to read a vernier scale caliper; so it technically just showing the scale reading, not the measurement reading. The measurement itself would also take into account things like the zero point error and any other systematic error that affect measurement accuracy. Those are assumed to be zero for tools unless otherwise stated; otherwise we couldn't even take for granted things like the scale being machined with a proper linear scale. Most calipers are also temperature rated to work best at 20°C so temperature corrections are other systematic obvious ignored for the image. Unless there is a good reason, systematics are assumed zero unless they are a known problem. With calipers, the zero point is usually the greatest concern for a systematic. In this case I have trouble seeing the first frame clearly enough but it doesn't look like there's any zero point error. Nor is there's any reason to invoke any other of the large number of possible systematics that could affect the measurement error. A reading and the measurement coincide on an instrument with negligible systematic error corrections. Jason Quinn (talk) 04:56, 19 September 2017 (UTC)[reply]

Keep although I would prefer the measurement in mm, not cm. OTOH, the scale on the caliper is in cm, so... Also, 2.47 is correct, since 2.470 implies an accuracy of 4 full digits, not attainable here (max. vernier resolution is 0.005, not 0.001) --Janke | Talk 05:54, 19 September 2017 (UTC)[reply]

Kept --Armbrust ^{The Homunculus} 03:49, 27 September 2017 (UTC)[reply]

Comment. I think really deserved a closure that commented on the weights of the arguments. This strikes me as a "by the numbers" closure. And as I think the "keep" arguments are rather poor and even disrespectful to the standards we aim for with featured status, this result is an example of group think. Jason Quinn (talk) 18:29, 27 September 2017 (UTC)[reply]