stat.interval syntax code should not be S

[martina-vicinanza]

I'm trying to define the UCD for a numerical interval of a photon flux in hard gamma ray (i.e. from 0 to an upper limit).

Being both em.gamma.hard and stat.interval defined as S words, I came up with
phot.flux;em.gamma.hard;stat.interval or phot.flux;stat.interval;em.gamma.hard.

Since I'm describing a numerical interval first and then characterizing its physical content,
stat.interval should be the first word of this UCD.

This conflict with the current syntax code and maybe it needs to be discussed wheter it should be a Q or a P word.

[BaptisteCecconi]

Hi @martina-vicinanza, I tend to agree with you (as discussed this afternoon).

If used as a primary word stat.interval indicates that 2 values are expected in the value container (e.g., a cell of a table), and then the UCD describing the quantity should follow.

If used as a secondary word, then it qualifies the primary word.

Any thoughts? @loumi @SebastienDerriere

Note that this interpretation is in line with the discussion that occurred during the face to face Semantics meeting in Dec 1st 2025, in Paris.

[msdemlei]

On Wed, Dec 03, 2025 at 08:33:08AM -0800, martina-vicinanza wrote: martina-vicinanza created an issue (ivoa-std/UCDList#111) I'm trying to define the UCD for a numerical interval of a photon flux in hard gamma ray (i.e. from 0 to an upper limit). Being both _em.gamma.hard_ and _stat.interval_ defined as S words, I came up with _phot.flux;em.gamma.hard;stat.interval_ or _phot.flux;stat.interval;em.gamma.hard_. Since I'm describing a numerical interval first and then characterizing its physical content, _stat.interval_ should be the first word of this UCD. This conflict with the current syntax code and maybe it needs to be discussed wheter it should be a Q or a P word.

I'd argue against Q. For the purposes of UCDs (or at least the stat.interval atom), something *is* not an interval of something, it is *something* (which happens to be represented as an interval). Perhaps there is a case for an interval as such, but let's not overload stat.interval for that, and this is not that case. What you have is fluxes, and the concrete form of them is secondary. The concrete sequence needs to be phot.flux;em.gamma.hard;stat.interval We have an interval of gamma fluxes, not an interval in the gamma region. You can also try this with translators of UCDs to plain text (e.g., https://svn.ari.uni-heidelberg.de/svn/gavo/hdinputs/ucds/res/ucdexplainer.py, but that's currently broken due to U Heidelberg <expletive>). That thing yields Generic interval between two limits (defined as a pair of values) of photon flux or irradiance in the hard gamma ray (>500 keV) for the right UCD and Generic interval between two limits (defined as a pair of values) of photon flux or irradiance hard gamma ray (>500 keV) for the wrong phot.flux;stat.interval;em.gamma.hard. As a human, you can tell which one works and which doesn't. Do we need to amend the UCD grammar to let a machine work that out, too? I don't think so. If we wanted to do that properly, we'd probably have to define a dependency structure for each atom, and given that it's rather rare for people to form new UCDs that would seem an unproportionate effort. My advice is to first try and see what others have assigned anyway: https://dc.g-vo.org/ucds/ui/ui/form (but that wouldn't have yielded anything useful in this case, I think).