meta.ref.ivorn : deprecated term still in list.

[loumir]

meta.ref.ivorn has been deprecated in previous version (UCDList1.3).

For consistency it should be removed from the current list .

cds tools allow to check deprecated terms and propose replacement .
question : should the deprecated list maintained at cds , uploaded to this git repo?
http://cdsweb.u-strasbg.fr/UCD/ucd1p-deprecated.txt

[msdemlei]

On Thu, Jun 17, 2021 at 01:19:21AM -0700, Mireille LOUYS wrote: meta.ref.ivorn has been deprecated in previous version (UCDList1.3). For consistency it should be removed from the current list . cds tools allow to check deprecated terms and propose replacement . question : should the deprecated list maintained at cds , uploaded to this git repo? [http://cdsweb.u-strasbg.fr/UCD/ucd1p-deprecated.txt](http://cdsweb.u-strasbg.fr/UCD/ucd1p-deprecated.txt)

This is an interesting question -- if we ever move the UCD atoms to RDF, I'd say the deprecated atoms should stay in, marked as deprecated, as per our policy there that terms don't vanish once they're no longer preliminary. But that's of course because of the validity rules we have for those vocabularies, which may or may not apply to UCD atoms at this point. Given that, my take would be to keep only the atoms people should actually use in the repo and rely on VCS history and the change log of this UCD list EN to document deprecated terms. Having two UCD lists in the repo is probably going to do more harm than good.

[gmantele]

I have been suggested to update a small library helping to check UCDs : Ucidy.

In this library/validator I use 2 files provided by this UCDList document: the PSV file with all IVOA recommended words and another one mapping the deprecated words and their suggested replacement. This latter is used by my library to give advice to users by suggesting him/her to use instead the recommended replacement word.

So what you say is that now I have to maintain this list by myself instead of relying of a file provided along the standard (anyway, was it of the standard?) ?

As far as I can see there is no new deprecated words in this new version. Is it correct?

[gmantele]

Anyway, using the updated list (UCD1+ 1.4) and the former deprecated mapping, my tests failed because of 2 UCD1+ words:

• meta.ref.ivorn (as we just spoke about above)
• pos.resolution

pos.resolution is apparently deprecated since UCD1+List 1.10. Is it normal it is back in? I possibly mis-interpreted all of that as I just scanned the document...I have not read it yet carefully.

[loumir]

Hi Gregory, +1 for ucidy t seems good to maintain 2 lists , so the assign tools will import the endorsed note 's list as well as other validation tools could do.

[BaptisteCecconi]

Hi @gmantele, I confirm that pos.resolution has been put back in the UCD list during UCDlist 1.3 RFM.

The new definition of pos.resolution is probably different from the initial one.

[BaptisteCecconi]

Shall we maintain the second list (deprecated terms and suggested replacements) in this repo, too ?

[gmantele]

I understand the reasons why @msdemlei would prefer not to do so. But as a consumer of these lists, it would be more safe if the authors of the UCD1+List maintain both lists, for the following reasons:

• less risk of error when an implementation has to build itself the mapping,
• and so less risk of divergence between implementations (though I think there are not many implementations for this second list),
• it avoids ambiguity in case a deprecated word becomes no longer deprecated (as it happened here with pos.resolution).

This is why it would be more comfortable for me if both lists are officially maintained by the authors.

Now, if it is too risky/difficult for you, I can still adapt and build this list of deprecated words by myself every time a new version is released (which mean that I may ask a confirmation from time to time for some words...which is ok for me if it's ok for you).

However, please, still keep maintaining the list of official UCD1+ in a separated file as done currently (in this format or another)....this is really the most important for me.

[BaptisteCecconi]

ok, we can discuss this while preparing the next version.

About the ucd-list.txt file: the plan is to maintain it as it is now (as a separate file).

[BaptisteCecconi]

The term has been removed from ucd-list.txt and has been included in ucd-list-deprecated.txt