>> * How do we deal with obsolete items in a registry?
>>
>> Sometimes, an item becomes obsolete in a repository,
>> but it is very rare that it is acually deleted.
>> Therefore, the status can still be 'active' while the
>> item is obsolete.
>
>This is an interesting idea. Possible ways to deal with this (in the
>long term):
>
> 1) add a status type, "obsolete" (I like this the most.)
>
> 2) change nothing; just for the purposes of the IVOA registries, mark
> these as deleted.
>
> 3) add some other mark up to tag this characteristic. (I like this the
> least.)
In the HST archive we (CADC and STECF) have adopted the following mechanism to keep track of the various versions of a given dataset that might get re-archived for various reasons. It consists of four fields:
dataset_name:
in the registry context should be the ResourceID
current:
can be "Y" or "N". There can be only one (or zero) version of a given dataset that has current="Y" all the other versions must be "N"
creation_date:
unique date for each new version of a dataset
validation_date:
a date that gets updated every time we check whether the current
version is still current.
Since this field is updated only for current datasets, as soon
as such dataset becomes obsolete (current="N") the validation_date
is never touched again and hence becomes the date this dataset
"died"
With this system we can always reconstruct the history of changes, and keep track of the current datasets. The same can be done for Resources.
It is particularly important, as Sebastian was saying, when there are publications that refer to resources that might have become obsolete. The VO should keep track, and not delete those resources, isn't it ?
Alberto Received on 2003-11-24Z23:03:52