Include uncertainty in measured quantities?

[Sleort]

Measured quantities, like melting points an boiling points, always come (and should be presented) with an uncertainty. What about including these, e.g. by using Measurements.jl?

[rahulkp220]

@Sleort That's an interesting suggestion.Would you mind adding this functionality through a PR?

[Sleort]

I had a quick look at it, and although introducing the functionality itself shouldn't be too hard, updating all the (automatically generated) data would be a major project I unfortunately don't have time for... (Besides, if one starts being more serious, it's no longer sufficient to say just "melting/boiling point", as such quantities also depends on the specific allotrope/isotope etc.)

I fixed a few grave errors, though (due to the automatic table generation, I presume)

[giordano]

Where is the data? Perhaps I can have a look at this.

[rahulkp220]

@giordano sure. It's here

[giordano]

@rahulkp220 Yeah, I saw the source code of this package, I was asking for the data of the uncertainty of the melting points. I can't find them anywhere.

Perhaps @Sleort knows a source where this information can be retrieved.

[rahulkp220]

@giordano Oh okay.

[Sleort]

@giordano I tried to track down some uncertainty data, but it seems to be somewhat tricky (as also mentioned in the discussion of this more comprehensive list of melting points at Wikipedia.) I guess one would have to look into the original research papers to find such information...

Also, as I mentioned above (after having given my proposal a second thought), melting/boiling points (especially accurate ones) really depend on details (allotrope, isotrope, pressure...) beyond merely "element", so giving an accurate "value with uncertainty for element X" could be somewhat misleading...

[carstenbauer]

@Sleort I think you're right that uncertainties aren't that simple and thus would be misleading. Should we close this?

[Sleort]

Yeah... For now, at least.