- Events
- Planning to attend/attended:
- Others of interest:
-
- Browsing
- e-Science
- Preparation for June '08 OGC Technical Committee Meeting GLS Presentation:
- Emailed to Peter Schut and uploaded Final Version of GLS Presentation Slides to OGC Portal. I can access them using the following URL:
- The latest version of the GLS specification was posted to pending documents in April. I can access them via the following URL:
- There is no Request For Comment (RFC) on GLS specification version 0.11.0 yet and no specific working group AFAICT:
- Obtained powerpoint template from:
- MoSeS
- Virtual meeting with Belinda to test Skype and plan for next Technical Meeting
- We had to play around with the audio configuration on both our machines so that I could hear Belinda as the default feed coming to me was too quiet:
- I looked for instructions for when Mike did this in another test earlier in the month:
- I suspect that we will both need to readjust our settings for a meeting involving someone else as there was a large hiss on my earphones giving me the impression that they were turned up to the max!
- Agenda items for the next Technical Meeting:
- Belinda is chairing and asked for suggestions after noting that model integration was already a key item.
- I would like Actions to be an agenda item:
- I want a list of Actions including those completed, those on-going and those forgotten about drawn up so we can review progress.
- If I think of anything else, I plan to email and post on the wiki.
- We discussed parallelisation of the dynamic simulation of the demographic model again.
- There is the coarse grained parallelism in that the four modules can be computed independently for each time step iteration.
- The migration model is the most expensive module and parallelising it further is no longer a focus of concern:
- Paul's analysis of the profiling output suggests that it is a data access issue more then anything else and hopefully some optimisation of this will result in a massive speed up.
- Organising and formatting the Individual Level Household and Communal Establishment population data:
- It is more convenient to use Vectors than Arrays and use Zone_Code char[] rather than int values for the keys of a HashMap for the population. For the UK test show this would require:
- 640045064 bytes of memory.
- 122973736 bytes of memory if int[] are used
- This is over 5 times heavier than a more minimal storage, but it is still acceptable.
- Being able to handle all the data in memory means that the processing of all the current output for the UK (which is in bits and still incomplete for a few bits) should be more straightforward than it would be otherwise.
- Specification (emailed to Matt Ford) for developing Geographic Linkage Service (GLS) on NGS
- Security model
- Ports:
- 22 for SSH
- 80, 443, 123, for shibboleth
- 8080, 8443 for tomcat
- For shibboleth, we need the machine to have fully qualified domain name e.g. blue12.ngs.ac.uk
- The only users that will be able to invoke this service will be those with UK e-Science Certificates and which have agreed to the necessary license agreements for the data being accessed and linked.
- Accounts/Individuals/Roles:
- Junaid Arshad
- Andrew Turner (Andy)
- John Watt
- Thomas Doherty (Tom)
- A root account shared between this group would be ideal:
- Junaid is an Honorary Research Fellow at the University of Manchester if that helps entrust him with this.
- Roles:
- Junaid will install the OGSA-DAI and the OGC Extensions and work with John and Tom to put appropraite security for the GLS in place.
- John and Tom will take charge of the Shibboleth configuration and work with Junaid to put appropraite security for the GLS in place.
- Andy is overseeing this work and is in charge of reporting this work to some of the projects it relates to:
- He is not planning to get involved in a hands on manner, he is a back up just in case anything happens to Junaid.
- Wants specific and consise instructions (perferably also a single executable) to build the GLS on another resource.
- To demand appropriate documentation resides in the appropraite places.
- Minimal computational and data storage requirements:
- 1GB RAM/Fast Access Memory.
- 10GB Disc Space.
- (This is likely to change for a fully deployed service (depending on demand), but for development and testing this should do.)
-
- Browsing
- Miscellanea
- SoG
- Responded to the email thread "The future status of CSAP-facing questions on Geog 2970" after Andy Evans posted the out of date questions.
- GEOG2970 is the new level 2, compulsory, 10 credit variant of GEOG3900 (independent study).
- It would be good to challenge the students to think and write about computational geography and the challenges in developing geographical information systems in a variety of contexts.
- I would like to have work that feeds computational geography research at the University of Leeds and beyond because that is what I'm interested in.
- As is the nature of computational geography, some purely applied work, or some mixture of this and more theoretical/philosophical reflection would perhaps be best for an "indendent study".
- Some things to address:
- Using available resources for demographic modelling:
- Contemporary UK
- Population predicition/forecasting
- Modelling the past historical/pre-historical
- There is an ongoing revolution/evolution of society and models of it given Information TEchnology advances that enhance capability for assimilating, using and disseminating geographical information.
- Argue the case for and againt open, collaborative e-Research in addressing the problems facing society like climate change, disaster managment, environmental degradation and population pressure and over exploitation of resources.
- Discuss the geoinformation and communication challenges in harnessing public participation and promoting self governance.
- Describe some challenges of computational geography in 2008.
- I'm keen to see work that addressed the above, but I don't think I'm qualified to do any marking/assessing ;-)
- e-Science
- Completed UK Research Data Service (UKRDS) Service Feasibility Study Survey.
- Virtual Meeting with Matt Ford and Junaid Arshad via Skype
- Aim
- To discuss options for installing OGSA-DAI with OGC geospatial extensions on NGS resources in Manchester and developing a Geographic Linkage Service as part of UK e-Infrastructure for e-Social Science.
- This work overlaps with MoSeS, NeSS and SEE-GEO.
- Introduced the GLS work and clarified what the GLS Server was and the version of OGSA-DAI on which it is based.
- What are the options?
- Use OGSA-DAI shared area
- Manchester supposed to stand up this, but it is not operational as they currently have no users.
- In a way this is attractive because there are no other users to upset.
- Perhaps this is preferable for a more production/stable service once developed under option 2.
- Deployment into a Virtual Machine
- This is good for flexibility and us having root access will be good for when John/Tom need to do shibboleth configuration.
- The Virtual Machine Monitor (VMM) will be Xen RedHat 5 based.
- The NCeSS VMM is VMWare:
- This means simply copying the NCeSS setup across is not a viable option as the migration is non-trivial and it is probably better to setup a new install at NGS.
- That is extra practice for Junaid :-)
- Specification needed
- Details of the Security model
- Ports, IP Addresses and other details of communication (firewalling).
- What accounts are needed?
- A root account shared between Junaid, Andy, John and Tom would be ideal:
- There is a policy barrier in giving out the root password, but no technical difficulty:
- Perhaps Junaid being an Honorary Research Fellow at Manchester helps overcome this as it did with the NCeSS VM.
- Who will be the users and what roles will they have.
- Minimal computational and data storage requirements:
- 1GB RAM/Fast Access Memory.
- 10GB Disc Space.
- (This is likely to change for a fully deployed service (depending on demand), but for development and testing this should do.)
- Agreed to keep a sysadmin style log to help with administration and the development of guidlines for using the VM:
- Details of what software is installed and what configurations are made.
- Actions
- Andy and Junaid to develop specification and send to Matt this week
- Matt to set up the VM after receiving the specification (likely to be ready for us to use within a week :-).
-
- Browsing
- Miscellanea
- Added name to the OSGeo Members list:
- e-Science
- MoSeS
- Organising and formatting the Individual Level Household and Communal Establishment population data:
- Testing different data structures for the Individual Level Household Population (HP) and Communal Establishment Population (CEP) data for the UK:
- For each Output Area (of which there are 223060) we require two arrays of longs, one for HP Household Reference Person (HRP) IDs and another for CEP IDs
- From (http://www.statistics.gov.uk/census2001/profiles/uk.asp), there are:
- 58789194 People in the UK
- 2.36 Average household size
- 24479439 Households with residents
- So:
- 57771476.04 HP
- 1017717.96 CEP
- And therefore on average each Output Area has:
- Less than 5 CEP
- About 110 HPHRPs
- Tests in Java suggest that it is likely that all this information can be held in:
- A HashMap with keys as Output Area IDs as int, and values as an Object containing arrays for HPHRPs and CEP IDs
- 122973736 bytes of memory if int[] are used
- 226786368 bytes of memory if long[] are used
-
- Browsing
- Miscellanea
- Preparation for June '08 OGC Technical Committee Meeting:
- Went through GLS presentation slides and corresponded with Peter Schut
- Asked Chris Higgins to fund this from the JISC OGC Grid Collision.
- Suggested the library get 5 copies of the following book when it is available:
- Geomorphometry, 33: Concepts, Software, Applications
- Handling big rasters question from Steve Taylor
- Steven asked for written details to provide to the GIS officer at the Cairngorms National Park Authority about a problem we experienced reading large asciigrid files into Arc:
- In using ArcGIS we discovered that we could write out some grids to ascii file versions, but were unable to load them back in again.
- I wrote:
- It is hard to provide exact details as I'm not sure I kept records of the errors and so forth. Anyway, here is an attempt at a partial explanation:
- We do not know how ArcGIS loads data becaused the source code is unavailable, also the error messages are often uninformative. However in working with large raster data we have inferred some things about some of the different operations. When loading a large asciigrid data file into a grid more memory can be used than in joining together two grids to make the same. This is down to how the operations work and what occurs as the data is loaded. It is to do with data structures more than anything I suspect. Asciigrid data is a row major ordering and I think the internal representation of a grid is a quadtree, which sort of works on chunks.
- e-Science
- MoSeS
- Retrieved data from the weekends processing:
- Ran checks:
- Process halted prematurely 25565 zones processed, 4436 zones still to process.
- This processing job was halted due to my using all my allocated resource :-(
- SEE-GEO
-
- Browsing
- Miscellanea
- Deegree documentation translation
- e-Science
- Meeting with June Finch
- June joined us in Leeds on a Node visit to learn more about MoSeS and NeSS work and get feedback.
- ...
- MoSeS
- Technical Meeting
- Set NGS Job running for OAs 120000-150000
- Meeting with Jonathan Exon and Mark Birkin to discuss finance with regard MoSeS and NeSS and
-
- Browsing
- e-Science
- MoSeS
- Set NGS Job running for OAs 106900-110000
- Completed and retrieved result.
-
- Browsing
- Miscellanea
- e-Science
- MoSeS
- Population initialisation
- Progress Review
- Transport variable optimisations
-
- Browsing
- Miscellanea
- Completed NGS User Survey
- First post on the Geowanking email list
- SoG
- Meeting with Chris Adams
- He has made contacts at Leeds City Council for data on road crossings.
- Planning a focus on child pedestrian accidents.
-
- Browsing
- Miscellanea
- Completed the survey at:
- Sog
- 2nd Meeting with Chris Adams from this years GEOG5080 Master cohort about Spatial Analysis of Stats 19 and Emergency Services Road Accident Incidence Data
- Chris is looking to obtain data from the Yorkshire Ambulance Service and also considering approaching Local Police Authorities to get some data on accidents involving emergency vehicles.
- We looked at the Stats 19 Data available via ESDS and saw no way to distinguish accidents involving emergency vehicles from this.
- Chris is to have a think about what to do and write down some thoughts about the data that exists and how it might be analysed.
- I provided my opinion and talked about my experience researching Road accident incidence data in terms of assessing road safety. I still feel that we lack the basic data about traffic that is needed to analyse road accident incidence. Other data about the road infrastructure is much better now.
-
-
-
- Browsing
- Miscellanea
- SoG
- Road Accident Data
- Created Usage 32602 with ESDS for Road Accident Data Research and added all Stats 19 Road Accident Data from 1991 to 2006 inclusive.
- e-Science
- Agora (http://agora.lancs.ac.uk) Collaboration
- Agora Test 3
- Belinda and I tested on MoSeS Worksite.
- There is still an issue with Belinda's camera.
- Until Agora is being more actively developed, let's use Skype...
- NeSS
- Briefed Junaid for Thursday's project meeting.
- MoSeS
- Meeting with Paul and Mark where Paul showed progress on MoSeS Web services :-)
- SEE-GEO
- Telecon
- Chris Higgins emailed minutes of this meeting later in the day :-)
- People
- Agenda
- Report back on OGSA-DAI NGS installation
- Planning for security round this installation
- Report back on the Shibb OGC demonstrator
- AHM2008 paper
- Events, eg, Potsdam OGC TC
- AOB
- Notes
- Going through the agenda
- Report back on OGSA-DAI NGS installation
- Andy reported on where we are at with NGS install
- Junaid and Ally have practiced the install together on an NCeSS Virtual Machine
- Still waiting to hear back from Andy Richards about which NGS machine to install on.
- Andy took an action to email Andy Richards, Rob Allan, Chris Higgins to try to get this moving.
- Michael noted that OGSA-DAI 3.0 works best with a new Globus 4.0.5.
- It is very important for this demonstrator to get this installed on NGS hardware.
- Planning for security round this installation.
- Richard went through an email (2008-04-29 GLS Server Install) where he outlined the security model
- Authorization on the portal side is to be done by SCAMP Portlet
- This does the job of saying to the portal - Here is a new portlet to add and these sites can use it (in the UK access management federation).
- Everyone agrees that some security is wanted on the data service ends, but that tying these to work with only the GLS Server on the NGS would be OK.
- Richard is actioning John Watt and Tom to work on this in June.
- Report back on the Shibb OGC demonstrator
- Andreas Matheus, Michael, Richard and Chris met and developed an architecture
- WMS client (UDig) authenticating via Shibboleth
- geoXACML to control/restrict what data is provided from any service depending on geometry etc...
- Michael pointed out that joining Shibboleth and XACML is a great thing :-)
- AHM2008 paper
- Ally maybe has time to feedback before Chris sends it.
- Events, eg, Potsdam OGC TC
- Agile paper by Michael well received
- Potsdam OGC TC
- Andy and Chris going
- Andy planning to present on GLS on behalf of Peter Schutt and GeoLinking IE
- Expecting to present on what was done in the IE and on details of th GLS 0.11.0 specification.
- 4th International e-Social Science
- Nottingham GeoWebSerives
- OSGeo in the monestary
- AOB
- ECGS meeting to talk about how to use available/spare computational resources (scheduling renice style).
- Andy to talk to Paul about improving the presentation of results.
- Chris going to OMII-UK Board Meeting on Friday
- Chris mentions Panos Periorellis (now working with Microsoft?) who editied a book Securing Web Services that I am reading...
-
-
- I did work today, but this is all I've managed as a blog entry!
-
- Browsing
- Miscellanea
- SoG
- Contacted Hazel Blackburn about Human Resources and finding out information about the NCeSS e-Infrastructure for the Social Sciences finances...
- Chronis Seminar...
- EGC RC Meeting
- e-Science
- Agora (http://agora.lancs.ac.uk) Collaboration
- Mike Crabtree, Adrian Fish, Paul Townend, and spent time testing:
- Setup:
- Adrian set up a test worksite on http://portal.lancs.ac.uk which we used.
- Mike and Andy were connecting with direct access PC's from University of Leeds.
- Paul connected via a laptop at University of Leeds.
- Problems
- Firewall?
- Mike explained the fairly liberal policy governing the University of Leeds firewall.
- Adrian tested and adjusted the portal sites firewall.
- Mike's microphone setting needed adjusting as he was too quiet:
- Solution was to increase the microphone volume in Windows XP : Settings : Control Panel : Sounds and Audio Devices : Voice : Recording.
- For some reason Paul's camera driver not picked up:
- NB. His Camera is old and he had to hunt for a driver that would work.
- Results
- All audio channels worked.
- Mike, Andy, Adrian video worked.
- Whiteboard and chatroom only working outgoing for Paul and Mike.
- The diagnosis is that the symptoms indicate issues with TCP.
- Plans
- Adrian to liaise with Xiaobo and plan another test or sequence of tests.
- Needs to know versions of NCeSS Portal Sakai and Agora
- Andy to email and find out availability for a test on 2008-05-13.
- References
-
- Miscellanea
- Events
- SoG
- Road Accident/Safety Planning Research
- Meeting with Chris Adams from this years GEOG5080 Master cohort following email exchanges:
- MASS
- Excellent Tutorial/Practical Session on RePast run by Nick Malleson :-)
- e-Science
- Agora (http://agora.lancs.ac.uk) Collaboration
- Forwarded Adrian's email of yesterday to Paul, Belinda and Mike Crabtree asking for checks on camera driver and firewall settings/configurations...
- References
- Browsing
- Miscellanea
- e-Science
- Agora (http://agora.lancs.ac.uk) Collaboration
- Emailed Adrian Fish about collaborating on
- I am investigating using the tool for MoSeS and GENESIS and have experienced some issues to feedback.
- I want to get some help overcoming problems so that we have the tool working well for virtual project meetings...
- I want to work with Agora as it is open source and a product of e-Science even if it is not the best tool
- The benefits are in supporting e-Science and open source (using to prevent abandonment and promote development.
- I think it is important that the open source research community continue to learn how to develop this technology.
- Agora has some functionality beyond what is offered by Skype or similar, but it does not have all the bells and whistles of a Marratech or Access Grid Client.
- Adrian replied
- Miguel has moved on and that he is looking forward to working with Steven Swinsburg on Agora :-)
- The NCeSS Portal is running a slightly old version of Agora.
- He expects the problems we experienced are to do with camera drivers and firewall settings, both Windows and institution:
- We are to make sure that:
- only camera driver present on the system is the driver for the current camera.
- the Windows firewall is allowing traffic in and out on 4000-4003 UDP and 4004 TCP.
- Browsing
- Miscellanea
- e-Science
- Added Junaid Arshad as an OGC Portal User.
- Confirmed availability for a NeSS, MoSeS, SEE-GEO GLS telecon meeting 2008-05-13 or 2008-05-14.
- References
![[Validate RDF]](http://www.geog.leeds.ac.uk/people/a.turner/images/rdf_w3c_button.gif "W3C RDF Validation")
![[Validate CSS]](http://www.geog.leeds.ac.uk/people/a.turner/images/vcss.gif "W3C CSS Validation")
![[An image of Andy Turner]](http://www.geog.leeds.ac.uk/people/a.turner/a.turner.png)