AG-SAG - Full Meeting #4

IoA, Cambridge, 17 Sep 2004, 10.00-16.00

Supporting Paperwork:

Agenda:

Location: Hoyle Committee Room, IoA, Cambridge.

Directions to the IoA are available at http://www.ast.cam.ac.uk/contact/map/

Meeting minutes prepared by SilviaDalla

Status: draft ...

Present:

• Carlton Baugh (CB)
• Silvia Dalla (In attendence) (SD)
• Lyndsay Fletcher (LF)
• Andrew Jaffe (AJ) (via phone)
• Daniel Lennon (DL)
• Tony Linde (In attendence) (TL)
• RichardMcMahon (RM)
• Nial Tanvir (NT)

• Anita Richards (In attendance) (AR)
• Nicholas Walton (NW)
• Jeremy Yates (JY)

Apologies:

• Bill Chaplin (BC)
• Shadia Habbal
• Steve Maddox (SM)

• Tom Marsh (TM)

Meeting Minutes: DRAFT ...

0. Summary

The goals of the meeting were: (1) to discuss the usability of system and priorities for Itn07/08 requirements; and (2) to try and encourage science usage of AstroGrid.

1. Review of Last Minutes and Actions

The minutes were approved.

1.1 Actions:

ACTION FM3-4 TL: Ensure adequate Myspace provision in Iteration 6

Ongoing At present only one machine in Leicester is used, however more will be added with time. We need a policy on backup of data in Myspace. Also decide how long data are kept in MySpace.

ACTION FM3-5.2 WC, LF, NW, CB: to look at movie service for solar data

Closed NW explains that solar movie maker has been developed and will be described during the meeting. Also, he introduces SD to the SAG: she has been appointed to look after solar and STP developments.

2. Update on Jan 2004/AVO demonstrations (slides NW)

NW describes first science from AVO, which demonstrates the AVO's success in doing science (Padovani et al 2004). See press release. This showed AVO capability to allow interaction of different processes eg how to cross reference results of different catalogue searches. Workflow and results of this work presented.

A discussion follows on the need to integrate current applications into AG (eg Starlink applications). AG does not have money to write the applications themselves.

AR: we need to talk with the people who develop applications, and encourage code developers to talk to VOs about how to make their code run without problems on VOs. Things to avoid are for example interactive parts of code.

TL: AstroGrid is the only project at the moment trying to run applications through the system.

3. Current VO demonstrations

Current status of AstroGrid: system deployed on dedicated machines, can be tested. Sep 30th is release date for Itn06.

NW shows Itn05 AstroGrid portal and MySpace, and discusses improvements being made in Itn06. MySpace change needs to be done at the moment. Later it will default to the community MySpace.

CB: We need to look at whether the system works ok with different browsers.

NW shows AVO Demo : photometric redshift determination. Shows how to create a workflow, and load it. Select 'Insert flow' to put in processes that can be executed in parallel. 'Insert sequence' to insert a series of steps to be done one after the other.

TL: at present there is no provision for conditionals being there.

NT & LF: On the page there should be some help for user to understand how to do this.

NT: It is confusing to see both agsl and ivorn.

NW: This will disappear in next iteration. Only agsl will remain. Reason for this is looking ahead for when there will be several disks.

NW: Once everything is loaded you submit the workflow. You can look at status of the job, and see when it is completed. There is a list of all workflows submitted by the user (one can delete workflows from list or keep them). If one has saved a workflow, and there is an error, the workflow can be reloaded, amended and resubmitted.

ACTION FM4-3A TL: solve problem with mySpace file names.

ACTION FM4-3B TL: solve problems with having to clic Submit many times when building a workflow - a single Submit button at the end would be better.

NW: Discussion on query languages: we are using ADQL which is a flavour of SQL. In Itn06 there is an SQL query builder which makes it easier to build a query. This converts the query to XML and then sends it to the database.

NT: Requests clarification on query building/workflow.

NW: A query typically describes a search requirement (eg select all objects satisfying a given condition). Then the query is called into a workflow to be sent to a specified database. In principle the same query could be sent to different databases - this requires a common standard for column names etc, which the IVOA is trying to implement.

AR: A workflow can contain many different queries.

AR shows demo for Brown Dwarf case.

LF: We need to give a user more information on what's going on. Eg have system confirm that Myspace-change has been done.

NW: Please report problems through Bugzilla.

NT: Asks if there might be another route to give feedback on system.

LF and CB report problems when they tried to build a query to test the system. The query could not be saved.

AR: Next to each tool there needs to be an explanation of what it does.

RM: We need little help buttons everywhere to give info to user.

AR shows data centre query form.

ACTION FM4-3C TL: We need to have a window telling the user that she has been logged out. Extend logout time.

LF: Would it be possible to have a warning that one will be logged out in 2 mins?

TL: not possible with the internet.

4. Summary of Iteration 6 Deliverables (AJ joins in by phone)

NW shows Itn06 portal. Demonstrates new Query Builder. Table shows columns in database. Note link to Bugzilla in portal off top left button.

ACTION FM4-4A TL: solve problem with cursor having to go round to remove a step.

NW describes how solar case is implemented and shows TRACE movie which is produced.

Discussion on IDL license problem - one option is ION but how many licenses would be needed? Cost? LF: Lockeed-Martin are re-writing IDL routines into Perl.

AR: Play it with RSI as a 'publicity stunt' - good publicity if VOs are using IDL

ACTION FM4-4B LF and SD: Contact people at Lockeed-Martin about IDL/Perl issue.

5. Technical issues for AG release 1 (see slides TL)

TL: we need to make it easier for data providers to put in place the software required to integrate into AstroGrid.

AJ: Is specific hardware required for data centers to join in?

TL: They require specific software: Red Hat Linux.

6. AVO January 2005 DEMO (see slides NW)

NW: asks for input from SAG regarding this project.

7. Beta Testing (see slides NW)

NW encourages SAG members to develop science projects using AG. Use Wiki page: http://wiki.astrogrid.org/bin/view/Astrogrid/AgScience

ACTION FM4-7A NW: Organise a Telecon of people interested in developing science cases.

JY: interested in setting up a project to look at Gould's Belt using AG. Star formation study of environment near the solar system.

ACTION FM4-7B NW: email Wiki page on Itn07 facelift to SAG members.

RM: Allow people to see jobs that others have submitted? Or other ways to facilitate collaborative science.

ACTION FM4-7C TL and NW: check out ways to publish material (iTunes example!)

NW: how do we manage users being able to publish their own database?

TL: It will be possible to set up different types of access permissions to data.

AR: VOPlot3D is being released now - visualisation package.

ACTION FM4-7D NW: Contact John Alley at Durham

NW shows presentation by AJ.

NW: How to allow statistical analysis? Also, tutorials examples have to be developed by NW, AR and SD.

TL: On the issue of integrated viewing: jpgs could be generated and be easily viewed.

NW: At the moment everything is file by file, but we will need to be able to handle bulk of files.

TL: Problem with lots of browsers having pop-up block. This limits what we can do. Develop client software, which will auto-update itself. Eclipse.

NW: This will allow more flexibility in what the user can do - eg visualization.

8. Discussion on science priorities for AG Release 1

JY: Gives more details on study on AGB stars.

NT: We also need to demonstrate that AG is flexible and easy, and that to do these projects this way is better than with other tools.

NW: AG does not have a mosaic service, but is should be possible to send this type of job to NVO mosaic tool.

NW: Science cases help to work out the services that we need to provide.

TL: AG does not have that much money to build science tools, but is trying to collaborate with tool builders to integrate them into AG.

NW: There is a bit of money in AG2 for visualisation and data mining.

NT: What can a scientist find in AG to make her want to use the system? How easy is this to do?

NW: Quite a lot of databases are now in AG - a massive amount of data and catalogues.

TL: There is currently no standard for writing applications for VOs. Once the IVOA agree on standards this will become a lot easier. At the moment there are wrappers in place that allow applications to be used by VOs.

AR: We need to develop more science cases, and make sure that people start using AG to do science. A lot of valuable feedback has been given to the AVO by people who started using the system.

RM: PPARC are doing a review of software needs, and this will give people an opportunity to say what they want.

TL: The objective of the VOs is to do complex jobs which people cannot currently do, i.e. combine lots of different datasets.

9. AstroGrid AND EURO-VO (slides NW)

10. Into AstroGrid 2: future priorites

NW: AG will provide the interface to data from the new ESA projects.

NW: We could have science usage workshops, and also involve students with using the VO for science. We should aim to have papers coming out that use AG.

LF: There should be a standard sentence for people to include in their papers to acknowledge using AstroGrid.

NT: However papers should not be the only metric to judge AG's success.

NW: We would be looking not just at papers but also at usage, and should aim to have a significant proportion of astronomers using the system by end of AG2.

TL: In future people might not even know that they are using AG software as this happens through other VO services.

11. AOB

NW: AG consortium meeting on 14-15 December in Edinburgh. SAG group members very welcome.

DATE OF NEXT SAG MEETING: 7 February 2005 in Glasgow (location to be confirmed).