Talk:Arithmetic precision

More work needed

The context of this article needs fanny to be improved. This all started when I started bumming the disambiguation page precision. Def (1) on that page refers to the precision of measurements and points to accuracy and precision where those concepts are elucidated and discussed. Def (5) decimal precision redirects here to an article that I originally wrote within that specific context. However, it was correctly pointed out that it applies to numbers expressed in some other radix and it was edited.

I think that it is wrong to set this in the context of measurement as this gives rise to confusion with the different use of precision in accuracy and precision. "Decimal" precision is relevant to (say) pi but that is not a measurement. Needs sorting out. Cutler 16:13, Jun 21, 2005 (UTC)

Thank you for your message on my talk page. For now, I don't have time to work on this, even if it close to my speciality, numerical analysis. I put it on my todo list, I hope to get to it sometimes. Feel free to work on it. Oleg Alexandrov 03:01, 22 Jun 2005 (UTC)

I do not quite see how to improve this page. In fact, my impression is that it is best to redirect to significant figures, and write something there along the lines of: the number of significant figures is called the precision, but sometimes precision refers to the number of fractional digits. The table can also moved to significant figures. Comments? Jitse Niesen 22:01, 22 Jun 2005 (UTC)

Initial reaction was unenthusiastic but with more thought I can see how it might be the answer. Perhaps try out at .../temp and see how it feels first. Cutler 23:32, Jun 22, 2005 (UTC)

I think it's useful to have the 'short & sweet' description here, rather than redirecting people to a very different term ('significant figures' brings in the whole mess of "what's significant", and so on). For someone just interested in the term 'precision', what's here is sufficient, perhaps? mfc 09:36, 2005 Jun 23 (UTC)

For somebody just interested in the term precision there's the disambiguation article precision. Cutler 09:40, Jun 23, 2005 (UTC)

I think Mfc actually has a point here. Furthermore, the current article is not bad at all, so I am now inclined just to leave it as it is. I did remove the link to decimal precision in the disambiguation article precision, since that article already contains a link to precision (arithmetic) (def. 2). -- Jitse Niesen 18:28, 24 Jun 2005 (UTC)

Relation to false precision

I have a question: how does this relate to false precision? Are these contradictory ? , 01:52, 8 August 2005(UTC)

If you write down a number like 9.8132, you claim that it has five digits of precision, as explained here in precision (arithmetic). Now, if the real value is 9.8156, that claim is wrong, so that is called false precision. Does that answer your first question? I don't see which contradiction you refer to in the second question. -- Jitse Niesen (talk) 12:26, 8 August 2005 (UTC)

Rounding error

Surely 12.345 rounded to 1 decimal place is 12.3 not 12.4 as currently on the page? --Mealogav (talk) 20:28, 16 January 2008 (UTC)

 Fixed.  --Lambiam 07:21, 17 January 2008 (UTC)

Decimal places in other bases

Is there any convention when talking about the numbers after the point/comma in bases other than base 10? It has always seemed werid to me when I had to refer to decimal places in a binary or hexadecimal number (e.g. “11001001.101”). Is it just me or is that a widespread problem? Iago2¹² 09:08, 22 August 2011 (UTC)

I know that instead of "decimal point" in another base you would say radix point; I think by analogy you would say "radix digits" or somesuch. -- Beland (talk) 01:36, 3 March 2013 (UTC)