Mistakes in registry

[Aurelien Stebe]

Hi Ray,

I would just like to inform you of similar efforts done here at ESAC
concerning the curation issue, as you were not present at the last 
interop in Madrid.
Since we started harvesting and filling our Registry, we grew more 
concerned about
this issue and began working on a curation utility that would latter 
plug into our Registry.

This utility was presented at the last interop. It is for the moment 
only in terminal mode
and checking SIAP resources, but it's very modular and checks the 
Registry resource entry aswell
as the actual service itself, following the IVOA standard recomendation 
for that type of resource.

We hope to have it soon checking all current types of services and 
resources, and be
enough modular to plug into the registry, work as a standalone 
application or a
librairie that can be used in other applications.

Cheers,
Aurelien STEBE


Ray Plante wrote:

>Hi J-Christophe,
>
>  
>
>>I have remarked that in the registry there are some mistakes. For 
>>example, sometimes the "?" is missed when the serviceURL has not ARGS 
>>and sometimes the "&" is missed too when ARGS are indicated.
>>    
>>
>
>  
>
>>What do you think about a validation service ? 
>>    
>>
>
>You've definitely touched on an real problem right now in the VO.  At the 
>moment, addressing the quality of registry information is a top issues in 
>the NVO at the moment, and we're plugging away at it.  Just to give you a 
>quick summary of how things are going:
>
>  *  this last year, the IVOA Registry WG adopted a mechanism that allows 
>     registries to rate the compliance level and quality of a registry 
>     records.  This is to be coupled with a practive of feeding this 
>     information back to publishers so that they can be corrected.  
>
>  *  the NVO has been under going the arduous, manual process of reviewing 
>     the records, correcting mistakes where we can or contacting 
>     publishers where it is not appropriate.  
>
>  *  We've been developing validators.  The first completed one for 
>     Conesearch services (beta) can be found at 
>
>        http://nvo.ncsa.uiuc.edu/VO/services/csvalidate.html
>
>     This validator has been run on all know ConeSearch services; the 
>     results can be viewed at:
>
>        http://nvo.ncsa.uiuc.edu/VO/services/csinventory/summary.html
>
>     A SIA validator is nearing completion; this will be followed by a 
>     registry service validator.  
>
>Your feedback on any particular records would be very helpful.  You should 
>contact either the contact person listed in the record, or the 
>contact for the registry that you received the record from.
> 
>  
>
>>Of course, I can check myself if the serviceURL is valid and add the '?' 
>>or '&' missing in my program. But I consider that if we have a norm 
>>describing well syntax, it is pitty to not use it and to have spent so 
>>time to build a norm which is not respected.
>>    
>>
>
>Many clients and services have had to resort to fixing these URLs on the 
>fly.  
>
>  
>
>>So, I think it is very easy to check some fields like serviceURL. The 
>>other ones, I don't know because I miss some documentation. For example, 
>>I don't know where the different values of ResourceType fields are 
>>defined ? if it is possible to add others ones or not by a user ? I 
>>don't know what is the difference between SIAP, SIAP/ARCHIVE and 
>>SimpleImageAccess ?
>>    
>>
>
>Sounds like you are using the STScI registry.  If you use the interactive 
>web interface, you will see things like SIAP and SIAP/ARCHIVE.  This is 
>some what of a historical artifact and not part of the standard interface, 
>as it were.  SIAP simply labels a SimpleImageAccess service; SIAP/ARCHIVE 
>refers to a a SimpleImageAccess service whose ImageServiceType metadatum 
>is set to "Pointed" (I believe).  It's probably not documented, and we 
>should probably take it out.  
>
>The definitions of all the *core* concepts described in the resource 
>record are defined in the standard document, Resource Metadata for the VO
>(http://www.ivoa.net/Documents/PR/ResMetadata/RM-20051115.html).  Another 
>document that is specificly defines the ResourceType is in the works; 
>however, documentation for all of the (emerging) standard 
>schemas--including definitions of the terms--can be found inside the 
>schema files, all of which are available at 
>http://www.ivoa.net/xml/.  Metadata that are specific to services like SIA 
>can/should be found in the service's respective standard document (e.g. 
>http://www.ivoa.net/Documents/WD/SIA/sia-20040524.html).  
>
>hope this helps,
>Ray
>
>
>
>  
>