IVOA Interop Meeting, Northern Fall 2020 (virtual)

Abstracts

This page contains abstracts for (some of) the presentations at the Interop meeting. They are in approximate order of submission, i.e. the order is not significant. If you want to add an abstract for a talk you're giving that doesn't appear here, or edit one that does, please go ahead.

Note: the listed sessions have been entered by hand from the current state of the schedule at 16 Nov. Rescheduling and mistakes are possible, so these times are not definitive - please refer to the main schedule and links from that for the definitive presentation times (and feel free to update this page with corrections).

-- MarkTaylor - 2020-11-16

Simple Cone Search status (Marco Molinaro)

Session: DAL 2 (#11)

Status of development of the ConeSearch-1.1 specification: open issues, priorities, implementation feedback requests.

Applications of Vocabularies 2 (Markus Demleitner)

Session: Reg/Semantics (#14)

With a view to the RFC of Vocabularies 2, since the last interop several applications of the underlying technologies have been implemented. This talk will discuss them, ranging from datalink term expansion in pyvo to stilts’ validation code to a UAT-based registry browser. The talk will also look at experiences with the application of the UAT in the Registry.

MANGO A Model for Source Data (Laurent Michel)

Session: DM 1 (#4)

The MANGO model proposes a flexible way to expose data related to astronomical source objects in an interoperable way.
It takes into account the huge diversity of source data in terms of feature description, format and usage.

ObsCore Extension for Time Series Data Products (Mireille Louys)

Session: DM 2 (#6)

A scenario to describe specific time related metadata in an extra table to be joined with the main Obscore table.

Data Model Posture Review (Markus Demleitner)

Session: DM 2 (#7)

The thread http://mail.ivoa.net/pipermail/dm/2020-September/006096.html on the data models mailing list has brought to light again a fairly fundamental dissent in the community on what the "grand design" of our data model effort ought to be: Small, ideally isolated models tailored to solving specific problems, or a hierarchy of entangled models of increasing complexity. This should ideally be a full session with two introductory talks presenting the two views. I’m adding myself as a speaker of the small and isolated faction. I suspect Mark C-D, François or Laurent might volunteer for the large and complex faction.

pgSphere: MOCs and Maintenance (Markus Demleitner)

Session: Apps 2 (#9)

The pgSphere package currently available in Debian testing now has comprehensive support for MOCs, which, together with TOPCAT’s recent plotting improvements, makes for several interesting use cases. We will discuss some of them, together with the implications for ADQL we have implemented in DaCHS. However, pgSphere maintenance currently is a bit in shambles, which is why the corresponding PR to the mainstream pgSphere package has not been merged so far. After the (brief) contribution, I would like to discuss the options for ensuring the maintenance of the package that, after all, ought to count as critical infrastructure of the VO.

Enriched Query-Server Response: DataLinks and Column Groups (Judith Silverman)

Session: DAL 2 (#11)

The Caltech/IPAC-IRSA TAP service reads database tables at start-up which define VOTable-style column groups and DataLink service descriptors along with the sets of columns associated to them. At run time, the TAP service adds a GROUP element (or a RESOURCE element of type "meta") to VOTable-format responses for any column group (or service descriptor) the set of whose columns is contained in the expanded SELECT clause of the user query. This presentation will indicate how this functionality is supported.

Monitoring VO service performance in NAVO (Tom Donaldson)

Session: Ops (#8)

NAVO has started monitoring the response times of VO service queries to establish baseline performance numbers and to identify areas for improvement. In this talk, I will discuss

• the goals of the monitoring
• the software we’re using and the nature of the queries
• initial results and our plans moving forward

Web SAMP from HTTPS update (Mark Taylor)

Session: Apps 2 (#9)

Review current status in the long story of getting the SAMP Web Profile working from web pages served over HTTPS.

TAP authentication progress (Mark Taylor)

Session: GWS 2 (#12)

I will report on progress in implementing the challenge-based authentication ideas outlined by Pat Dowler at the last virtual Interop. Some prototype client-side authentication code using this approach for working with TAP services has been shown to work with a couple of services, but more work on standardising the server-side implementation needs to be done.

ObsLocTAP status (Jesus Salgado)

Session: DM 1 (#4)

ObsLocTAP is in TCG review status. We will expose the status of the standard, the activities related to this standard and plans of dissemination

Upgrading the largest VO publishing registry (Sébastien Derriere)

Session: Reg/Semantics (#14)

The VizieR publishing registry has been online since 2007, and represents with its 20,000+ astronomical catalogues a significant fraction of all VO resources. The original registry software was written in PERL, and over time a number of issues have surfaced. In order to improve the compliance and maintenance of the service, a new version has been recently developed in Python. We will describe the new architecture, and improvements provided by this new system.

Interoperability Developments in AAS WorldWide Telescope (Peter K G Williams)

Session: Apps 2 (#9)

Over the past few years, the AAS WorldWide Telescope software system has undergone significant enhancements, especially in the project infrastructure such as build tools, deployment automation, and documentation. This work has enabled the development of significant new features for both researchers and more casual users. I will demonstrate some of the developments most relevant to the IVOA community, most notably progress towards rendering of HiPS all-sky datasets and the display of very large FITS images.

The Indra Simulations on the SciServer Science Platform (Bridget Falck)

Session: GWS 1 (#10)

Understanding the observations made by large-scale structure surveys requires input from theoretical predictions of structure formation in the form of numerical simulations, which can be hundreds of terabytes in size. How do we make these simulation suites available, either within collaborations or to the public? How do we connect them to observations available on archives or science platforms? I will discuss my experience developing the tools and architecture to make the peta-scale Indra suite of simulations available to the public and computationally accessible on the SciServer science platform.

Non-browser client authentication with OAuth2 tokens (Brian Major)

Session: GWS 2 (#12)

The CADC recently implemented the authorization server aspects of OpenID Connect. This work allows authentication/login with CADC credentials from external sites using a web browser. During this exercise we kept in mind the need for authentication integration with non-browser clients. Using existing OAuth2/OpenID Connect patterns, and building off the ideas of the working group in the last interop, we will illustrate an approach to end-to-end login for non-browser clients.

Implementation of the IVOA Spectral Data Model at IPAC (Vandana Desai)

Session: DAL 2 (#11)

IPAC is home to several data centers that serve spectroscopic data: The NASA/IPAC InfraRed Science Archive (IRSA), the NASA/IPAC Extragalactic Database (NED) and the Keck Observatory Archive (KOA). Together, these archives curate spectra from many observatories and projects, including Spitzer and SOFIA. We present our experiences in beginning to adapt metadata for these existing data sets into the IVOA Spectral Data Model. We also strongly advocate for the prioritiza- tion of work on the Spectral Data Model in light of large and important spectroscopic data sets to be produced by SPHEREx, Euclid, Roman, JWST, and Gaia.

The EPN-TAP access protocol (Stephane Erard)

Session: DAL 1 (#3)

The EPN-TAP protocol to access Solar System data has recently been submitted as a Working Draft. EPN-TAP is the association of TAP with a specific vocabulary, EPNCore, describing data in a field encompassing Solar System studies, heliophysics, extrasolar planets, and laboratory experiments such as spectroscopy in solid phase. The standard also includes a set of rules to design EPN-TAP services and tables. EPNCore results from design of 50+ data services published in the recent years within the Europlanet H2020 EU programme, and accessible via the VESPA portal (http://vespa.obspm.fr). Version 2.0 has been in use since 2015, and is the first version submitted to IVOA.

VO Education in Covid Times (Priya Hasan)

Session: Edu/Apps/DAL/DM (#15)

The pandemic disrupted the regular flow of classes. Undergraduate and Postgraduate students were losing out on precious time. We were funded by a special call by IAU-OAD for Covid times. We started a project to train students to use astronomy archival data to do targeted research projects. I shall describe the essential VO Tools that were taught and a few of the projects taken up by students. Education through VO is a very powerful tool and can be used very effectively.

Discussion: Protocols for handling file-oriented catalog data (Gregory Dubois-Felsmann)

Session: Edu/Apps/DAL/DM (#15)

Some of our institutions are looking at future services - some beginning in the very near future - in which catalog data are made available to the public in files, as well as through TAP services. There have always been "bulk download" mechanisms available for public catalogs (see for example https://irsa.ipac.caltech.edu/data/download/), but new technologies and use cases are moving us in the direction of wanting to support improved facilities for discovery and metadata service for file-oriented catalog data. The "Parquet" file format, and a rich infrastructure of data science tooling, is allowing for the construction of highly efficient systems for performing analyses across whole catalogs, and cross-matching between catalogs (see for example the AXS project at https://axs.readthedocs.io/en/latest/). The growth of the use of cloud storage facilitates access to large datasets in situ where they are stored, rather than requiring bulk downloads.

We see a number of issues that are worth exploring:

• Discovery of the existence of large file-oriented catalog datasets
• Registration of the equivalence of a file-oriented view of a dataset and a corresponding TAP view
• Metadata service for the content and spatio-temporal coverage of the individual files within such datasets, especially when non-compact spatial sharding is involved, such as in AXS, creating tension with the ObsCore STC-region-based coverage model
• Query capabilities within such metadata services, allowing users to identify the subset of files which cover a region of interest to them (perhaps along the lines of MOC-based queries)
• Means for applying the valuable existing infrastructure of VOTable and DataLink in documenting the schemas of columnar data, while also enabling the use of Parquet or other comparable future formats for the efficient access to the actual data
• Openness of the resulting conventions / standards to the implementation of interoperable APIs in commonly used languages in the community such as Python and R.

MOC 2.0 Status and Discussion (Pierre Fernique)

Session: Apps 1 (#6)

The MOC standard (1.0 and 1.1) was invented to handle spatial coverages. The MOC 2.0, currently being drafted, is proposing to extend these mechanisms to also manipulate temporal and spatio-temporal coverages. We will present the state of progress of this work + discussion

VOTable issue review and planning (Tom Donaldson)

Session: Apps 1 (#6)

We will review the status of the VOTable document in github, take a look at the open issues, and briefly discuss whether and when to move forward with those or other changes.

Implementing Service Descriptors in Firefly (Trey Roby)

Session: Apps 1 (#6)

Firefly, developed by IPAC, has been steadily implementing VO protocols into the web application for the last three years. Last year, we released a partial Datalink support without Service Descriptors. This year we have implemented Service Descriptor discovery support both as part of datalink and standalone. I will discuss implementing Service Descriptors from a UI consumer perspective and summarize what went well and the challenges with some of our solutions. I will also give a brief demo.