applications software metadata

Tony Linde Tony.Linde at leicester.ac.uk
Fri Jan 27 02:31:34 PST 2006


Sorry, missed one point:
 
<<If the latter, what is the formal content of the registry record that the
machine can parse and use?>>
 
That is what the Registry-DM tiger team is working on right now.
 
T.


  _____  

From: Tony Linde [mailto:Tony.Linde at leicester.ac.uk] 
Sent: 27 January 2006 10:26
To: 'registry at ivoa.net'
Subject: RE: applications software metadata


Thanks, Roy. I'll try to clarify your points.
 
<<Registration of applications for the NVO apps registry is through a web
form, as defined below. It is not XML>>
 
That was what I meant, whether you'd defined a schema for registering apps.
AstroGrid did do so as part of the CEA (now being turned into UWS by IVOA).
 
<<What is the purpose and intended usage of the applications registry?>>
 
There isn't an applications registry, just the registry, but I know what you
mean. The purpose is to allow the discovery of applications: apps which
might be added into a workflow, apps which might be downloaded and run on
client machines. I would exclude stuff like Excel but there won't be
anything stopping someone adding it if they believe it is a worthwhile VObs
app.
 
<<What exactly is meant by "application">>
 
I guess we mean any app which is useful to the VObs community in doing
astronomical science. The prime distinction will need to be between apps
requiring intermediate user intervention (GUI-type apps being the obvious
extreme) and those which can be executed automatically (eg as part of a
workflow) with all user intervention completed before the app is submitted.
 
Another way of looking at the apps registration is that it is for any app
which does not have its own separate registration schema, such as those apps
which give access to data, images, spectra etc and for which we have defined
a more complete set of metadata.
 
<<How about these: ...>>
 
Any and all of those I guess. <<Excel? Apache webserver?>>: I would expect
people to be sensible. <<STC schema?>>: is not an app afaik. <<My Pipeline
for My Data? A Java class to translate sexagesimal to decimal? An IDL
script?>>: any and all of those probably, providing there is an appropriate
means of providing the parameters and invoking them. <<An IDL tutorial?>>:
probably, unless we choose to create separate schema for educational
materials. The question to ask is 'would people/apps use the registry to
discover such a resource?'. This might be one case where Google would
provide more info. <<IDL itself?>>: no, IDL license precludes this. You
could register it but since you cannot execute it remotely there seems
little point. Though, I guess if it is a deployment of IDL which allows
authorised people to submit scripts and get the results then, yes, it could
be registered. In this case it would be more like registering a CEA server
which allows other apps to be executed.
 
<<Would the registry be primarily for humans or for machines to access? If
the former, surely we rely primarily on natural language description and
images? If the latter, what is the formal content of the registry record
that the machine can parse and use?>>
 
The registry is *always* for machines to access. No person can interact
directly with the registry, all interaction is mediated through software. It
is up to the search software to provide a human-friendly interface to the
registry; to the workflow software to provide an interface relevant to
people trying to construct a workflow; to client software to provide an
interface relevant to people looking for a downloadable app relevant to
users who want, say, a data display app; etc. The registry is solely the
repository of metadata; how that metadata is presented to users or other
applications is up to mediating software and how it is designed and built.
 
<<Why would somebody go to the IVOA applications registry rather than
sourceforge/google to find an application?>>
 
Because the registry contains the information in one single hit which
describes how to invoke that application in a manner appropriate to use of
the VObs. No-one is trying to stop people using Google. I use it many times
a day for looking for information. But if I want information about
applications (esp open source ones), I go to Sourceforge or Freshmeat
because they contain all the metadata I need to know whether those apps are
suitable for the task I have in mind. But if I want information about apps
which will run within an AstroGrid workflow, I go to the registry because it
contains the metadata about the apps and app servers which will allow me to
plug an app into a workflow.
 
<<which of these 22 should/could/may/cannot be registered>>
 
A bit difficult as all the links seem to be down just now. But I hope the
above would allow you to answer the questions yourself.
 
Cheers,
Tony.


  _____  

From: owner-registry at eso.org [mailto:owner-registry at eso.org] On Behalf Of
Roy Williams
Sent: 26 January 2006 18:50
To: registry at ivoa.net
Cc: Roy Williams
Subject: Re: applications software metadata



On Jan 26, 2006, at 9:56 AM, Tony Linde wrote:


How does this get registered? Can you supply a sample registration (ie, the
xml)?



Registration of applications for the NVO apps registry is through a web
form, as defined below. It is not XML, it is relational -- with the schema
defined below. You can see samples by clicking on any of the links in
http://www.us-vo.org/projects/tools.cfm

Before embarking on the joys of XML, the registry group might wish to answer
some of these questions:

-- What is the purpose and intended usage of the applications registry?
-- What exactly is meant by "application" How about these: Excel? Apache
webserver? STC schema? My Pipeline for My Data? A Java class to translate
sexagesimal to decimal? An IDL script? An IDL tutorial? IDL itself?
-- Would the registry be primarily for humans or for machines to access? If
the former, surely we rely primarily on natural language description and
images? If the latter, what is the formal content of the registry record
that the machine can parse and use?
-- Why would somebody go to the IVOA applications registry rather than
sourceforge/google to find an application?

Some food for thought is in the NVO Summer School proceedings(*). Scroll
down to the "Software Index" and ask which of these 22
should/could/may/cannot be registered in the proposed applications registry.
Roy

(*) http://us-vo.org/summer-school/2005/proceedings/index.html




From: owner-registry at eso.org [mailto:owner-registry at eso.org] On Behalf Of
Roy Williams


Sent: 26 January 2006 16:53


To: registry at ivoa.net


Cc: Roy Williams


Subject: Re: applications software metadata



Here is the list of applications at the us-vo.org website:
http://www.us-vo.org/projects/tools.cfm

If you would like to register your application, please click on "log in" at
the lower left 0f this page, using the name/password Charles/Messier.

Here is the metadata that is used for the application description:
Project Title
Headline (One sentence description of the project.)
Category (Please select appropriate category)
Project Description
Project Contact Information (First Name, Last Name, Email)
Project URL
Image (Is there an image or logo to accompany this project? )



California Institute of Technology
626 395 3670



California Institute of Technology
626 395 3670


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/registry/attachments/20060127/edcdc512/attachment-0001.html>


More information about the registry mailing list