search

ivoa-std / DataLink

DataLink standard (DAL)
3 stars 6 forks source link

UCD par.min/max?? #60

Closed mbtaylor closed 2 years ago

mbtaylor commented 2 years ago

The "Custom Access Data Service" example in section 4.5 makes use of UCDs par.min and par.max. No such values exist in UCD1+ (there is no top-level par tree, and I don't think there ever has been). Should this be stat.min/max? Needs to be changed to something legal in any case.

This is a legacy issue from DataLink 1.0.

msdemlei commented 2 years ago

On Wed, Oct 20, 2021 at 09:47:16AM -0700, Mark Taylor wrote:

The "Custom Access Data Service" example in section 4.5 makes use of UCDs par.min and par.max. No such values exist in UCD1+ (there is no top-level par tree, and I don't think there ever has been). Should this be stat.min/max? Needs to be changed to something legal in any case.

This is a legacy issue from DataLink 1.0.

Oh, right, my bad. That's from the time when I had hopes we could establish the min/max thing as the recommended pattern for "PQL" arguments.

That's over now, and so I'd say the LAMBDA_MIN/LAMBDA_MAX pair should be replaced by a DALI-style single parameter:

<PARAM arraysize="2" datatype="double" name="BAND" ucd="em.wl" unit="m" va
 xtype="interval">
  <DESCRIPTION>Spectral cutout interval</DESCRIPTION>
  <VALUES>
    <MIN value="3.52631986e-07"/>
    <MAX value="9.21500998e-07"/>
  </VALUES>
</PARAM>

(which is what DaCHS actually does for this kind of service these days, e.g., http://dc.zah.uni-heidelberg.de/feros/q/sdl/dlmeta?ID=ivo%3A%2F%2Forg.gavo.dc%2F~%3Fferos%2Fdata%2Ff08751.fits).

Dissent? Should I just write a PR?

Bonnarel commented 2 years ago

On Wed, Oct 20, 2021 at 09:47:16AM -0700, Mark Taylor wrote: The "Custom Access Data Service" example in section 4.5 makes use of UCDs par.min and par.max. No such values exist in UCD1+ (there is no top-level par tree, and I don't think there ever has been). Should this be stat.min/max? Needs to be changed to something legal in any case. This is a legacy issue from DataLink 1.0. Oh, right, my bad. That's from the time when I had hopes we could establish the min/max thing as the recommended pattern for "PQL" arguments. That's over now, and so I'd say the LAMBDA_MIN/LAMBDA_MAX pair should be replaced by a DALI-style single parameter: <PARAM arraysize="2" datatype="double" name="BAND" ucd="em.wl" unit="m" va xtype="interval"> <DESCRIPTION>Spectral cutout interval</DESCRIPTION> <VALUES> <MIN value="3.52631986e-07"/> <MAX value="9.21500998e-07"/> </VALUES> </PARAM> (which is what DaCHS actually does for this kind of service these days, e.g., http://dc.zah.uni-heidelberg.de/feros/q/sdl/dlmeta?ID=ivo%3A%2F%2Forg.gavo.dc%2F~%3Fferos%2Fdata%2Ff08751.fits). Dissent?

No, no. Agreed !

Should I just write a PR?

Yes please