Notes on AVO SWG feedback session. Jan 26, 2005, Madrid, SPAIN. ---------------------------------------- scribbled and paraphrased by Mark Allen; slightly revised by Paolo Padovani 1. Interface Discussion ------------------------- S. Ansari - Graphical interfaces of the tools shown at the demo are not standard and are difficult to use for first time users. A uniform style would be beneficial responses: M. Allen. - Having different approaches to tool interfaces is important because it allows for creativity. F. Genova - The interfaces of the tools are designed to try to accommodate the range of users, from first timers to experts. P. Fernique - Compromise is necessary in order to have functionality for all the different types of users - Many of the components of the tools have been invented from scratch, for which there are no standards e.g. the "stack" visualisation component in Aladin. M. Dolensky - What standards are available? N. Winstanley - Style guides are available, but these are not really standards. A. Richards - Documentation is important, even in prototype tools. R. Hanisch - Use the standards (for interfaces) where it makes sense to do so. - We do want it to be as intuitive as possible, especially for the users first exposure to the tools. i.e. we need a simple "buy in", and then users can read documentation for the more sophisticated aspects. - One question is, 'should we have form interfaces for complicated queries'? - The NVO approach is to allow the user to actually see the SQL, but have lots of documentation. The NVO experience is that a few example (SQL) queries can be very illustrative. - A tutorial has been written to help first time users. - Seeing the power of SQL can motivate users to learn it. N. Walton - People will learn SQL when you can do something useful with it, for example to use the SLOAN database SQL experiences: N. Walton and others - It is not so difficult to write SQL when you have examples. P. Padovani - A blank white box in which to write SQL is a put-off A quick poll of the SWG shows that only ~ 30% have any SQL experience. P. Garcia-Lario - The Vizier interface is a good example which has both a form interface as well as sophisticated access mechanisms which can be used for complex queries. 2. Connecting with the Astronomical Community --------------------------------------------- R. Hanisch - How is Euro-VO going to reach a wider community than the SWG? P. Padovani - JENAM/EAS booth, CDS booth, etc. - We have considered going around giving seminars, but it requires time and resources P. Quinn - Our mission is to educate the community. The first part of that is to reach out to the data providers. Euro-VO is an evolving project, and in the end market forces will take effect. F. Genova - Feedback from booth visitors at meetings like the EAS is important N. Walton - The UK VO projects are reaching out to the community with sessions at the National Astronomy Meeting (NAM). W. Voges - The perception of the astronomical community is important, and we need to be connected especially in this transition phase. 3. Combining Data & Quality discussion -------------------------------------- A. Gomez de Castro - Is there a tool that can access ALL observatory logs and catalogs? Something that shows "everything available" ? To construct catalogs, and is there any sort of "IVOA warranty" for catalogs? P.Padovani - Building catalogs is not a job for the VO, rather for the data providers. Also, assigning quality is probably not a job for the VO. However, such a tool is certainly a must for the VO and we need to work on it. The NVO "datascope" is a very good step in that direction. A. Gomez de Castro - How can we combine all the data that are out there? P. Osuna - We are already combining data across archives using the SSA and SIAP standards. - Data quality is not an easy issue, for example Johnson J can mean different things to different people. P. Quinn - The VO is trying to set up the framework in order to enable astronomers to do projects like Ana is describing. A. Richards - We should recognize that different levels of specification are required. Sometimes default or average transformations (e.g from mag to flux) are sufficient for the scientific purpose. M. Guainazzi - VOs cannot impose quality because it is the data providers' responsibility. - Can we actually standardize the meaning of data quality? - No real conventions exist for DQ, e.g. even within a given subfield like X-ray astronomy no real conventions exist. A. Micol - DQ is part of the challenges in the Data Model for describing accuracy using metadata, e.g. having links in data to the transmission curve of the filter etc. - The tutorial meeting (in June/July 05) hopes to motivate data providers to come up with a good description framework for characterizing their data M. Guainazzi - Who should use workflows? P.Padovani - Interactive analysis vs. batch processing N. Walton - It is not just the complexity of a task that may require the use of a workflow, but also the volume. One does not want to interactively do 1000 button clicks. F. Genova - Scripting of some of the functions available in the AVO prototype tool is a kind of workflow M. Guainazzi - Before going public with the workflow tools, it really needs to be smoothed out in terms of usability. 4. Overall Comments ------------------- R. Hanisch - Having been to all VO events, we can note that the nature of the discussion here is different to what it was say 2 years ago. Here we a talking about SQL, Workflow and Usability. - Data Quality keeps coming up, and we will need to take it seriously. We can't police data quality, but we should police meta-data quality. - Some level of trust comes from using data from trademark providers, but privately contributed data sets are open to being misinterpreted. - Assessment of a VO data set/service might be: 1. Does it exist? 2. Does it work? 3. Is it described properly? - This is the hardest one. - Metadata management is required at some level. This will require astronomical expertise in order to specify what needs to be put in the metadata. F. Genova - The CDS has a long experience of data curation. It is a big task that requires astronomy knowledge. NED also has expertise in this.