Add example to help explain MIN/MAX clarification

[tomdonaldson]

30 Adds clarifying text on the intent of MIN and MAX for array-valued PARAMs and FIELDs.

An example illustrating MIN and MAX for each of an array, non-array and xtype value may make the text easier to digest.

[Bonnarel]

Sorry how should we understand the new VOTable text in the case of SODA issue #2 (https://github.com/ivoa-std/SODA/issues/2) which I recently tried to solve with SODA Erratum 3 (https://wiki.ivoa.net/twiki/bin/view/IVOA/SODA10Err3). Should we understand that here there is an xtype (interval) for the BAND PARAM and that in that case interval rules should apply to the meanning of the MIN and MAX array values ?

[msdemlei]

On Thu, Jan 26, 2023 at 08:06:38AM -0800, François Bonnarel wrote:

Sorry how should we understand the text in the case of SODA issue

2 (https://github.com/ivoa-std/SODA/issues/2) which I recently

tried to solve with SODA Erratum 3 (https://wiki.ivoa.net/twiki/bin/view/IVOA/SODA10Err3). Should we understand that here there is an xtype (interval) for the BAND PARAM and that in that case interval rules should apply to the meanning of the MIN and MAX array values ?

For the interval xtype, I'd suggest to keep the standard VOTable interpretation; that's good enough here, since the limits on both the upper and the lower limit of the interval are identical.

In general, we'll have to define the MIN/MAX interpretation where the xtype is defined (and yes, I wish we hadn't gone down that road).

[Bonnarel]

Le 27/01/2023 à 09:17, msdemlei a écrit :

On Thu, Jan 26, 2023 at 08:06:38AM -0800, François Bonnarel wrote:

Sorry how should we understand the text in the case of SODA issue

2 (https://github.com/ivoa-std/SODA/issues/2) which I recently

tried to solve with SODA Erratum 3 (https://wiki.ivoa.net/twiki/bin/view/IVOA/SODA10Err3). Should we understand that here there is an xtype (interval) for the BAND PARAM and that in that case interval rules should apply to the meanning of the MIN and MAX array values ?

For the interval xtype, I'd suggest to keep the standard VOTable interpretation;

so MIN[0] will be the minimum of the lower limit and MIN[1] the minumum of the higher limit

And MAX[0] the maximum of the higher limit and MAX[1] the maximum of the higher limit ?

that's good enough here, since the limits on both the upper and the lower limit of the interval are identical.

Humm. I don't understand this statement. Why should they be identical ?

Another option would be MIN contains the smaller length interval and MAX the higher length. I find it more immediately useful (and the interpretation is different)

In general, we'll have to define the MIN/MAX interpretation where the xtype is defined (and yes, I wish we hadn't gone down that road).

—

OK. this discussion is for the SODA erratum, not really for the VOTable (solved) issue, anyway.

Reply to this email directly, view it on GitHub https://github.com/ivoa-std/VOTable/issues/32#issuecomment-1406163005, or unsubscribe https://github.com/notifications/unsubscribe-auth/AMP5LTBYQAJMGJF7JXTTUBTWUOACRANCNFSM6AAAAAAUHTYSUU. You are receiving this because you commented.Message ID: @.***>

[msdemlei]

On Fri, Jan 27, 2023 at 08:58:11AM -0800, François Bonnarel wrote:

Le 27/01/2023 à 09:17, msdemlei a écrit :

For the interval xtype, I'd suggest to keep the standard VOTable interpretation;

so MIN[0] will be the minimum of the lower limit and MIN[1] the minumum of the higher limit

And MAX[0] the maximum of the higher limit and MAX[1] the maximum of the higher limit ?

No, as laid down in the PR, MIN and MAX are as for the datatype (i.e., atomic). @. is the minimal value for both the lower and the upper limit, and @. is the maximal value for both the lower and the upper limit.

that's good enough here, since the limits on both the upper and the lower limit of the interval are identical.

Humm. I don't understand this statement. Why should they be identical ?

I thought there was a bit too much "limit" in this sentence, but your following remark makes me believe we're on the same page, really. First, if you do a cutout between 1e-14 and 2e-14 Joule, any of the following intervals probably be good, no?

1e-14 2e-14 1e-14 1.00001e-14 1.99999e-14 2e-14

That observation made me write the statement above. Calling the limits of the interval "bounds", perhaps it can be written less confusingly as: "Minimal and maximal values of either bound of the interval are the same."

      -- Markus

[Bonnarel]

Le 31/01/2023 à 08:20, msdemlei a écrit :

On Fri, Jan 27, 2023 at 08:58:11AM -0800, François Bonnarel wrote:

Le 27/01/2023 à 09:17, msdemlei a écrit :

For the interval xtype, I'd suggest to keep the standard VOTable interpretation;

so MIN[0] will be the minimum of the lower limit and MIN[1] the minumum of the higher limit

And MAX[0] the maximum of the higher limit and MAX[1] the maximum of the higher limit ?

No, as laid down in the PR, MIN and MAX are as for the datatype (i.e., atomic). @. is the minimal value for both the lower and the upper limit, and @. is the maximal value for both the lower and the upper limit.

Ah OK. So in these conditions the "wrong syntax" identified originally by Alberto Micol in the BAND example was not wrong. Was it ?

And SODA erratum 3 has to be rejected.

I comment my own SODA erratum 3 email with this option.

that's good enough here, since the limits on both the upper and the lower limit of the interval are identical.

Humm. I don't understand this statement. Why should they be identical ?

I thought there was a bit too much "limit" in this sentence, but your following remark makes me believe we're on the same page, really. First, if you do a cutout between 1e-14 and 2e-14 Joule, any of the following intervals probably be good, no?

1e-14 2e-14 1e-14 1.00001e-14 1.99999e-14 2e-14

That observation made me write the statement above. Calling the limits of the interval "bounds", perhaps it can be written less confusingly as: "Minimal and maximal values of either bound of the interval are the same."

-- Markus

— Reply to this email directly, view it on GitHub https://github.com/ivoa-std/VOTable/issues/32#issuecomment-1409883575, or unsubscribe https://github.com/notifications/unsubscribe-auth/AMP5LTGUORPTKRAMCHOA6DLWVC4KVANCNFSM6AAAAAAUHTYSUU. You are receiving this because you commented.Message ID: @.***>