File talk:Table isotopes.svg
I noticed that tungsten, whose atomic number is 74, is stated to have 4 stable isotopes in the respective Wikipedia article (http://en.wikipedia.org/wiki/Tungsten). However, in this graph, those particular isotopes appear to undergo decay of various types. Is that perhaps reason to review the entire graph for its accuracy? Jimzoun (talk) 13:10, 31 October 2011 (UTC)
- Yes, I just noticed the same. The picture (which I like a lot by the way) correctly shows that Technetium (43 protons) and Promethium (61 protons) have no stable isotopes but wrongly suggests so for Tungsten as well. According to the Promethium Wikipedia page: "Promethium is one of only two radioactive elements that are followed in the periodic table by elements with stable forms, the other being technetium."
- I see in the image's metadata that it was likely created by https://fr.wikipedia.org/wiki/Utilisateur:Napy1kenobi using some unknown data table and unknown svg tool - maybe Napy1kenobi (or somebody else) could help us locating the data source and then we could either fix this isolated error or re-generate the whole image? Vaclav.hanzl (talk) 14:00, 3 January 2022 (UTC)
- Just to add a correction if anyone has the capacity to do it: I noticed Nickel-63 is coloured as stable, but actually has a beta-minus decay mode with a hundred-year half life. I noticed it was the only case of two neighbouring equal-mass-number isotopes both of which were stable (copper-63 and nickel-63) -- turns out that can't happen! Smowton (talk) 11:10, 28 November 2023 (UTC)
Method to build table.
[edit]
Unfortunately for anyone asking how this table was made (to correct errors found): handmade from a raster and a lot of patience. So errors must be corrected by hand. Fortunately since then, the data are much more accessible (eg NEA's JANIS toolbox). 2A01:E0A:999:F230:5873:5CA1:BC2D:5B09 16:52, 6 January 2022 (UTC)
- Thanks for the answer Napy1kenobi and for all the patience involved. The nice image now exists in 7 language versions, each with its own history of corrections - like making svg more compact, which is useful, and corrections related to magic numbers. But I did not see correction of data itself (as discussed above). Correcting data by hand for each of the languages really looks daunting and error prone, I'd be inclined to try some programmatic solution. I considered a dirty script working on the svg source code and exploiting regular structure of the small rectangles (fixing color of some based on the x and y coordinates) but the structure unfortunately diverged in individual language mutations. So I guess that two sort-of real options are:
- carefully check picture correctness and hand-fix all the 7 languages
- get data table (from JANIS?) and generate all 7 language variants of the svg by a common (say Python?) script (using strings from current versions). Vaclav.hanzl (talk) 18:29, 7 January 2022 (UTC)
Corrections discussed at 3 places
[edit]
Apart from this page, fixing the image (which has 7 language mutations) is also discussed here:
