I nternational
V irtual
O bservatory
A lliance
The goal of this IVOA Note is to introduce and explain practices followed and requirements found while creating and deploying astrophysical resources dedicated to educational purposes within the standard VO framework. Issues, proposed solutions and desirables are here reported to be possibly taken into account in future modifications of relevant standards.
In detail, we discuss: the curation of educational resources inside or along with standard registries; use cases and techniques for registering and locating documents, tutorials and similar within the registries; dealing with multiple languages.
(updated automatically)
This is an IVOA Note generated by discussion between Education IG and Registry WG members mainly.
A list of current IVOA Recommendations and other technical documents can be found at http://www.ivoa.net/Documents/.
blah
Advances in technology and communications are creating new and exciting opportunities for teachers to bring astronomy into their classrooms. As the VO makes science-grade data publicly available and classroom sets of (suitably) networked PCs are now standard in schools, exciting projects come within reach of teachers. In order to make things happen , it is important to disseminate material to help teachers tap into these resources. These include documented step-by-step tutorials, use cases explaining how to perform basic astrophysical research using VO tools and resources, and similar exist in various formats and have been translated in different languages.
New opportunities also come on the observational side. There is a growing availability of remotely controlled telescopes dedicated to education in many countries world-wide, from the Bradford Robotic Telescope on MountTeide, Tenerife (http://www.telescope.org) to the radio telescopes of the Radio Physics Lab, IUCAA, Pune (http://www.ncra.tifr.res.in/rpl). In some cases, educational telescopes are linked into a network with the aim of guaranteeing the best observing conditions, including deep sky observations during regular daytime school hours, and the best instrument for the particular program of interest. Examples of these networks are iTelescope.net (http://www.itelescope.net) and EuHOU-MW (http://euhou.obspm.fr/public).
As telescopes enter classrooms more frequently, interest is growing for a public archive of observations and hence for publishing and curation tools, together with the basic applications needed to retrieve, display and analyze data. The VO already includes most of the technology needed to satisfy the requests of educational observatories. In fact, since several years, VO, and in particular the European project EuroVO, is devoting part of its resources to education (http://wwwas.oats.inaf.it/aidawp5). It is therefore a natural decision for VO to tackle the problem of publishing educational data in VO archives.
Resource registration for both educational data services and documents is the most appropriate approach toward making educational resources available within the VO. While technically this may seem trivial, keeping too technical research services out of the the resources devoted to education will require some effort, that will also be needed in order to avoid contaminating VO professional research with obviously inadequate material.
In the next section we discuss the idea of educational resources curation, then Registering Texts we work out the use cases and needs for registration of tutorials and documents. Finally, we discuss the idea of introducing language internationalization in the resources.
From a technical point of view the registration of educational services does not require extensions to the existing for VOResource standard (std:VOR). The only real need for investigating changes to what already exists is due to a use case's distinction between resources to be used in teaching and dissemination versus all the research driven resources that exist in the VO.
For simplicity here we will distinguish these two groups of resources as educational and professional but without any intent of putting them on different levels of importance.
On the one side, teachers and educators may find it difficult to filter out from all VO resources those that are suitable for their tutorials and examples. On the other side, educational resources should not be retrieved by a standard professional query. Given that it is not a matter of data quality, but only a distinction upon the resources' scope, nevertheless this duality leads to an issue about the proper way to tag resources for educational usage.
In the next subsection we propose a possible tagging solution, based upon the existing ContentLevel element of VOResource, but requiring a small change to it. The subsequent subsection describes the idea of a curated registry for educational resources and the reasons for it to exist.
std:VOR already has the ContentLevel element allowing data publishers to optionally identify their resources as being suitable for one or more of the following audiences:
This element turns out to be misused by many publishers, presumably because it is not really clear what the subtle differences between the available possibilities are; also, to require a fairly substantial enumeration to convey "for school use" seems, in retrospect, not likely to promote widespread adoption. We hence propose to simplify the content model to:
We expect this to reach two goals:
This change in the already existing standard will require only a small effort to update already registered resources because nearly 97% of them currently have ContentLevel set to research, about 2% of them have no ContentLevel defined at all and only the remaining have a different value (or set of values) set for this element (Appendix A details better these figures).
Until the change in VOResource can be performed, it can work as a "best practice" recommendation, possibly even at a registry level, where registries can map existing ContentLevels values of University to Research and everything else except Amateur to General.
Even in the case of the simplified ContentLevel tagging system a curated registry for educational VO resources will be useful for educators in order to let their students work with a registry without having to worry about confusing material or overwhelming data sizes. A good example for this is the educational version of the Aladin sky atlas that has a built in, curated set of resources suitable for educational level tutorials.
Curation will require some effort in managing and keeping up to date such a registry but, most important, it is subjected to some restrictions coming from the IVOA resource registry architecture.
If such a registry were a standard publishing registry (std:RI1), its resources would be harvested by the full registries: this means that any dedicated educational resource would end up in the full VO set of resources. For reasons mentioned above, this is not desirable.
If it were to be a full registry, it will harvest itself all the existing resources, and not all of them will fit, or be suitable for, the educational scope the registry has to be preserved for.
We need a resource (the curated, in std:RI1 parlance, local, registry) capable of :
Educational material is not only about services – text-like material like tutorials, worked-out use cases, or textbook-like material are at least as important. Within the VO community, there is a large body of educational material for a wide variety of audiences ranging from pre-school to researchers:
To date, such material has been collected informally by the various projects on plain web pages. It is, in consequence, hard to find, with knowledge of it often passed on antecdotically. In order to improve upon this situation, we propose to keep record of educational material in the registry.
The VO already has a registry extension for standards, which of course are also text-like (std:STDREGEXT). This extension, however, focuses on metadata important for standards – e.g., vocabularies and status – that is not pertinent for educational material. Conversely, it is not concerned with document language (which can safely be assumed to be English for standards), and it disregards the issue of locating formatted and source version, which for educational material is important. We therefore propose a simple registry extension covering text-like material, dubbed DocRegExt.
The design of DocRegExt has been guided by the desire to fulfill the following use cases:
On the use cases of locating editable forms of such texts – which has been found to be necessary fairly regularly – we note in passing that representing source-product relationships is in principle in the domain of provenance and thus not in scope for the registry. However, in the case discussed here the relation is so simple and its representation so useful that we propose to include it in a DocRegExt.
To satisfy our use cases, we have designed a registry extension with
a single definition, extending the basic vr:Resource
element with three concepts to make the doc:Document
resource type.
The full schema is given in Appendix app:schema
We considered having language as an attribute of accessURL to allow language-specific document discovery. We decided against this mainly for reasons of maintainability; the same reason is behind the recommendation to have both access and source URLs as landing pages.
Since the access URL is supposed to point to a "landing page" anyway, it is tempting to just unify it with the standard VOResource reference URL. As a "human-readable document describing this resource" std:VOR quite conceivably could very well work as a landing page. For many documents having identical access and reference, urls will certainly work fine. However, we want to cover the cases in which different projects offer different translations or even versions of a document at different sites, which is not possible with just a single reference URL. A similar reasoning is behind including source URIs, except that in this case we wanted to allow URIs with non-HTTP schemes like svn or git.
Document-typed resource records should define relations to other
general resources (e.g. applications, services, ...)
they use; extending the vocabulary of allowed types in VOResource,
we suggest the relationshipType for these relations should be
uses
.
TBD: do we want i18n-ed titles?
In the relational registry std:REGTAP, DocRegExt is
entirely represented in the res_details
, with details
/language
-- A language the document is available in./accessURL
-- A URL allowing access to one or more
renderings of the document./sourceURI
-- A URL allowing access to an editable version of the document.Here is a (slightly abridged) example record:
TBD: should we use ISO-3166-1 two letter
country codes or ISO-639-2 two letter language codes?
MMo: I vote for 639-2. This shouldn't require changes now,
but it's better to clarify it in advance IMHO.
Registering text document as VO resources allows search for tutorials and other
materials through standard registry interfaces, but keeping
tutorials up to date, in their master form and also in their translated
versions, is another important issue to allow proficient use.
A versioned repository (using subversion as the version control system)
has been set up at GAVO data center (
http://svn.ari.uni-heidelberg.de/svn/edu/) and collects part of the
already existing VO tutorials with the goal of preserve them and let users
update and translate them in favour of the whole community.
The repository has an internal structure that takes care for:
The EURO-VO AIDA project WP5 produced multilingual tutorials, meeting the needs for high schools and lower level educational degrees in countries where English is not the native language. Clearly, it would be an added value to be able to register an educational or document resource not only in english (as it is done currently with standard VO resources) but also in other national languages.
Here we propose discuss whether this means:
This appendix reports some statistics on the usage of the ContentLevel element in std:VOR as of 2014-01-30, taken from the GAVO RegTAP endpoint http://dc.g-vo.org/tap . There are 14392 useful resources (excluding authorities, standards and similar) that expose 26 different values as their ContentLevel. In the following table these values are reported in order of count.
count | content_level string |
---|---|
13937 | research |
290 | |
41 | university#research |
40 | general#university#research#amateur |
24 | university |
14 | university#research#amateur |
7 | general |
5 | research#general |
4 | general#research |
3 | secondary education#community college#university#research#amateur |
3 | research#university#community college |
3 | elementary education#middle school education#secondary education |
3 | general#university#research |
2 | research#university |
2 | research#amateur#university#community college |
2 | general#informal education |
2 | general#elementary education#middle school education#secondary education#community college#university#research#amateur#informal education |
1 | university#community college#research |
1 | general#university#research#amateur#informal education |
1 | elementary education#middle school education#secondary education#community college#university#research |
1 | general#secondary education#university#research |
1 | university#research#general#informal education |
1 | research#university#amateur |
1 | elementary education#middle school education#secondary education#community college#university#research#amateur |
1 | elementary education#middle school education#secondary education#community college#university#research#amateur#informal education |
1 | university#research#amateur#informal education |
1 | general#university#research#informal education |
This table can be easily updated from the same endpoint (or an analogue one) using the following ADQL query:
SELECT count(*) as cnt, content_level FROM rr.resource WHERE res_type != 'vstd:servicestandard' and res_type != 'vg:authority' and res_type != 'vstd:standard' and res_type != 'va:application' and res_type != 'vr:organization' GROUP BY content_level ORDER BY cnt DESCThe table shows that only about 1% of the ContentLevel values use something different and more complex than research, when the element is not empty. Morever, of this 1% (165 resources), 61 include the general value (roughly 37% of them), 29 (17%) state that are devoted to some education level only, while 148 (90%) state that are also devoted to some education level (up to university).