Technical Meetings
Next telecons: 5 and 19 November 2008: 3:00 PM GMT, 8:00 AM PT, 11:00 ET, 4:00 PM Ireland, 4:00 PM London
Return to ICAN Working Groups Page
Open Tasks/Timeline
- [All] - see specific tasks/deadlines in Liz's spreadsheet
- [Dawn] - Drupal portal update by 1 November 2008
- [All] - discuss and implement vision for prototypes 2 and 3
- [All] - seek to finish prototype 2 by August/September 2009 in advance of ICAN Workshop 4 in Italy
Skype Conference Call 20081022
Chaired by Liz O'Dea; attended also by Tanya Haddad, Yassine Lassoued, Declan Dunne, and Alessandro Sarretta
1. Questions from the last call
Yassine agreed with Luis that the URI should not be constructed on the fly (hard coded in the prototype). The question remains that we still have two choices: i) full URI in the Keyword section, or ii) URI root in Thesaurus section plus concept name in Keyword section. Yassine is still determining which of these is easiest.
Sharing the code on the ICAN web site is premature at this stage - things are changing daily. The code will be shared when it becomes more stable.
Luis also had asked the question about the method that the CSW uses to store the URIs - he circulated an example tag for Yassine to consider.
2. Status of the tasks
Yassine has begun work on the new prototype 2.0 interface in Flex. Our task list includes tasks to improve Prototype 1. Yassine and Declan decided to save time by instead transferring the existing Prototype 1 into a new interface that uses Flex, and implement the prototype 1 changes during the migration. This interim interface is called version 1.1. He will do his best to have a functional version 1.1 By the end of November. He will try to get it working in time for Littoral 2008 conference in Venice, Nov 24-29, where Ned, Kathy and Simon will be giving a presentation and Prototype 1.1 demonstration. Once Prototype v1.1 is available Yassine will let us know and we will schedule a conference call for feedback.
In order for the prototype to work at Littoral, we will work around the FGDC metadata bug by inserting ISO metadata into the OCA GeoNetwork. This is a temporary fix and will give us more time to resolve the bug.
Prototype 2 will definitely include WMS.
Please review our task list and keep in mind deadlines for our tasks. The updated Tech task list is available above ("Liz's spreadsheet").. We will review this in the next conference call.
ICAN Drupal Site: Dawn is working to transfer web content to the new ICAN Drupal site for the Nov 1st deadline. Those involved in contributing to this should be prepared to add content (Dawn sent Drupal author training a couple of months ago to those who are listed contributors).
3. Next calls
Nov. 5th at 8:00am Pacific/16:00 Irish time (Note change of time): Prototype Status, review task list.
Nov. 19th at 8:00am Pacific/16:00 Irish time: Prototype 1.1 feedback call, prior to Littoral conference.
Skype Conference Call 20081007
Chaired by Liz O'Dea; attended also by Tanya Haddad and Luis Bermudez
Outstanding tasks from last call:
- September was a busy month for folks so there are still remaining tasks from the last call to be done by the end of October.
o Tanya’s going to send sample metadata to Ali so he can look into resolving the FGDC-to-ISO metadata bug in GeoNetwork.
o We need to have another call in October with the CMRC contingent to talk about the status of tasks and collaborate. Suggest October 15th ?
Washington Coastal Atlas:
- If we get the go-ahead to join, WA Dept of Ecology will use ESRI ArcGIS Portal Toolkit as the CSW. It will be beneficial to have a COTS product included in the prototype to get the experience. Liz met with ESRI Olympia to discuss if their CSW would have what’s necessary to work with the ICAN prototype (Thanks to Declan for the questions to ask!), and it looks good. They showed great interest in how it would be used in the ICAN prototype. (GIS Portal Toolkit is currently free, but we just learned they are talking about charging for it – apparently the new version is so easy to set up that they aren’t making enough money from support requests! We have ordered it already.)
URIs in the ICAN Prototype:
- Luis raised an issue about how we are discussing the technical approach in the project, after he presented at OGC. In the examples he used, he wasn’t able to see any URIs in the data that was sent. He is wondering if we are still using URIs in the prototype? He wants to be sure that he is selling the prototype correctly.
- The issue is that we agreed originally to use URIs as keywords, and yet our current deployment does not reflect that practice. So he guesses that the URIs are being constructed on the fly. We thought this would be something that Yassine could confirm. Luis requested that we revisit this technical point as it is very important that the full URI be included in the Keywords of the metadata, because it is important to know the ontology version being used, and this information should be provided by the data owner. Everyone tags services with their own vocabulary. Then map it to a central ontology and that’s it.
- Another point raised by Luis was the use of the correct ISO tags for storing the URI. We need to see if GeoNetwork (and ESRI’s portal toolkit) can support the appropriate ISO tags. See Luis’s email to the ICAN_Tech list from Sept. 16, 2008, where he shared this example of what the appropriate tag should look like:
Mis-translation Bug:
- One of the bugs in prototype 1 is that FGDC metadata is not properly translated to ISO on the fly by GeoNetwork. Ali will look into this bug. Hopefully the XSLT can be changed.
- Tanya says that ArcGIS 9.3 has a much-improved Metadata Translator tool. Tanya will play with it and see if it might point to ways to resolve the FGDC bug.
- We need to resolve the GeoNetwork bug in reading FGDC metadata directly and translating it into ISO on-the-fly. It is a harder sell if we are asking all FGDC people to translate all of their metadata into ISO to connect with the Prototype.
MMI:
- Luis’s presentation got a lot of good interest in our prototype.
- MMI plans to release a new suite of ontology tools next month that will be more straightforward and easier to use than existing ontology tools. It will upload ontologies to a registry and make it easy to search across existing ontologies. The tool is currently in alpha testing. Se example at the MMI Semantic web site, http://mmisw.org/regtest/.
Next calls:
Oct. 15th at 8:30 Pacific/11:30 Eastern / 16:30 Irish – October catch-up call
Nov. 5th at 8:30 Pacific/11:30 Eastern / 16:30 Irish – Discuss our vision of prototypes 2 and
Skype Conference Call 20080903
Chaired by Liz and Declan; attended also by Tanya, Dawn, Luis Bermudez, John Helly,
Goal is to discuss the tasklist and timeline that we set out in Copenhagen, as well as the update on the technical work in progress on the current prototype.
Picking up where we left off:
Pinpoint tasks, update deadlines, and make sure that we can accomplish what we set out to do
Do we need to elaborate on any points, set new dates
Let's just go through the list on Liz's spreadsheet
WEB PORTAL
Overview by Dawn
Luis - Drupal would like to help groups communicate better
help someone new to ICAN get engaged more quickly and better - their intro to ICAN
examples of WMSs,
John has been building Drupal sites and written a module to imbed SRB and allow it to be searched from a Drupal page - offers opp to put a federated data network behind a Drupal site
New deadline of Nov 1
Dawn's list of structure - will want to assign tasks to people to populate with content once the basic blocks and menus are modified
- CMS system we'd agree to initially develop using Drupal - Dawn is the
super admin and will be beginning to assemble the various lists of links
we'd talked about adding. She will also initiate the template
modifications discussed in copenhagen. Next step will be interfacing
with Simon's group about possibly connecting to their database of
experts, and other changes that will be enable more collaboration
between the ICAN participants. John Helly offered that he has recently
been working in Drupal and has authored a module to allow the embedding
of a ICAN like search into a Drupal site. Dawn will connect with John on
this.
TECHNOLOGY
- Yassine reported that he has a local solution to the geographic search
bug. He will try to get the solution up onto the sever by October.
Resolve Geographic Search Bug
- The OCA metadata search bug is related to the FGDC metadata
translation to ISO. Geonetwork tries to do an automatic translation that
is not working correctly, which makes the metadata impossible to search
once transformed. Ali from CMRC has been working a lot with GeoNetwork
SLD files and might be able to help track down the source of this bug.
Resolve OCA metadata search bug
Tanya
- Tanya and Kathrin will coordinate on the "ease of metadata insert"
issue. It should be an easy fix.
- same bug related to conversion from FGDC to ISO style sheet
- blunt force method is to switch all metadata from FGDC to ISO before loading
or more elegant is to dive into SLD files and see where we can find out where the translation mistake is happening (this is time well spent because otherwise it will be a problem for all US participants)
having to convert all FGDC metadata will be one more barrier to their participation
- needs time for occasional testing; needs to happen on Ireland end; Tanya can make changes to metadata but needs CMRC to make changes on search function end
- Yassine could only see the bug when connecting to OCA server with CSW query; from interface you can't see it; someone at CMRC is now doing work on GeoNetwork and playing around with GeoNetwork style sheets and perhaps he (Ali) could have a look at it
- Tanya to send sample metadata to CMRC to insert in GeoNetwork and test
- will keep Oct 2008 deadline and see how it progresses (could be difficult bug to fix)
- We set a date of end of October for error handling.
- We will initiate prototype II in November 2008. Yassine is testing
Adobe flex as a interface option.
- we still need to evaluate possible lessons that could be learned from
efforts like InterRisk and ISDE.
- we deferred smart search and WFS discussions to later in the year, as
we need to think more about it, and get other items described below done
first.
- We decided to implement WMS by December 2008. First step will be
thumbnails, second step will be full map access
Manual insertion methods
Tanya
- open-ended issue, prototype 2 at this point with no deadline
- should be a simple issue with a short date -
- a modification of which part of GeoNetwork Kathrin is using - move from manual to bulk import
- need to write a script to point to where bulk insert function is
- November deadline
Improve Error Mgmt
Yassine
- manage basic errors such as if 1 node is not connected it breaks up everything; put in directive to ignore those errors so as not to affect other nodes
- end of October
Next Prototype
Declan
- evaluate architecture
- smart search
- WFS
- start in November on prototype 2
- Yassine testing new technology here (Adobe Flex); Flex is a programming language which is Adobe and it is for client side programming and replaces JavaScript in web app (while JavaScript is executed by browser, Flex is executed by a Flash Payer plugin in your browser; good news is that if you develop any app in Flex it will be executed by any Flash Player in any browser; more dynamic, nicer, more advanced graphics; better for next prototype)
- WMS should be next step, to set these up at each ICAN node; MIDA has this already and OCA can set up quickly
- next, implement thumbnails associated with metadata (search comes back with results including map thumbnail); one larger map to visualize that dataset in detail; can re-use MIDA web GIS engine as a starter
- smart search; not sure how it will work exactly yet, many scenarios to consider; giving more flexibility in typing and selecting keywords; not obliged to select global keyword, can select others which will then be translated to global keyword; or can use MiDA keyword at local level which gets translated to global keyword
- WFS can be discussed further in Nov and Dec; focus on WMS for the time being
INTERFACE
Lots of items that all need to be moved to Oct
- we discussed fixing major functional bugs in prototype I by end of
october, and then defer other major interface improvements (anything not
easy) to prototype II. We hope to have a stable version of prototype II
ready by March 2009 for new ICAN folks to plug into
Nested search - implementing a tree
Yassine
- would prefer to add the new functionalities in Prototype 2 as technology will be changing anyway; save effort by just fixing errors in current version and then adding new functions for Prototype 2
- however if Prototype 1 is done by October, then time frame for these new functions in 2 will be straightforward; differences will be at interface level (HTML/Javascript to Flex programming), while engine on server side will remain the same
- get Prototype 2 to work before ICAN 4 workshop in Trieste, Italy
- at some point Prototype 1 will go to an unpublished state so that people won't see it and get confused;
- Prototype 1 is demonstrating a backend solution; 2 will have more of a focus on the front end user experience
- wrap Prototype 1 by October and move into Phase 2 after that; set up WMS servers in meantime; postpone interface issues until Phase 2
Overall timeline question - what do we tell new ICAN nodes about when it is possible for them to plug in (e.g., David Hart of Wisconsin) - they should plug into Prototype 2 - we need to migrate function of 1 to 2; by end of year
- they need to wait for Prototype 2 because of the way that we define key words and mappings will likely change (SKOS in RDF used in GeoNetwork but we are now using OWL which is not supported in GeoNetwork;
- we are using URI of term that holds tags to metadata record, but the way that we do mappings might change (right now we do mappings in OWL, but it is worth investigating the use of SKOS instead); MMI uses SKOS for most of the mappings, and it is built into the MMI VINE tool
- Luis - new users should be able to just set up WMS server, tag keywords, and then mappings can be done regardless of prototype version
- Yassine - trying to save effort of doing things twice
- key is that GeoNetwork is being used to provide CSW service for all atlas nodes; GeoNetwork supports the use of thesauri so you can define that in RDF or SKOS; current version of GeoNetwork does not do inferencing, maybe next release may support that; but why is reasoning needed? Some of the reasoning is done at the global level, as well as at local level (where we have a hierarchy of concepts); atlas provider must link records to thesaurus in GeoNetwork; language that we use to define our concepts should be same one use in GeoNetwork (problem is that GeoNetwork does not support OWL)
- Luis suggests that to keep things open, if someone has WMS and want to use CSW they can just publish in RDF, SKOS, OWL, or whatever format without worrying about doing reasoning. Can we let everyone use SKOS? All reasoning should be done only at portal side/semantic mediator side.
So if we Prototype 1 and minimum way to participate which already exists, let they continue for users while we learn about new technologies and players. We should not make user try to do reasoning on their side (keyword matching with thesaurus on network). People may also choose in future to stand up CSW without GeoNetwork (e.g., an ESRI version may be forthcoming), so we may not want to wed our architecture too closely to GeoNetwork
- Let's commit to use existing ontologies and try to improve only a few things. Let's incorporate that into documentation for new nodes which will also reside on the Drupal
- Liz is working on a technical work plan to justify her work on ICAN; she has produced a task list; she has cookbook of all steps needed to get Washington Coastal Atlas up to speed as a node; great template
- Bugzilla ready by December; should apply to interface, but also to everything else (ontologies, etc.)
Ontologies
Tanya, Liz, Declan, Yassine
- Declan suggested that we do a comprehensive review of what we have
done with ontologies once we are finished with fixing the bugs from
prototype I. This will help lead us into documentation of the ontology
portion of our work
- we had a discussion about Geonetwork, and how it is being used.
Yassine raised the question that Geonetwork uses SKOS and RDF and should
we reconsider our use of OWL? Luis argued that our current approach is
sound as it keeps the burden of connecting for new Atlases low. There
seemed to be some agreement towards the end of the discussion that we
would stick with the current approach.
- Luis described ontology registry products from MMI and from Stanford
and advised that ICAN should look at implementing something in this
regard.
Add MIDA admin boundaries to MIDA ontology
- results are only returned from Belgium because of Simon's node
- need to keep focus on coastal erosion as our continuing test case
- other topics may be incorporated in Prototype 3
- do we need to keep Admin boundary in or should we ask Simon to replace Admin boundaries with coastal erosion?
- we need to communicate to Simon on the emphasis on coastal erosion themes (Liz, Tanya, Yassine)
All other blue tasks under Ontologies
Tanya to take on superterms task
Need further discussion on need to separate mapping from local ontologies. Luis - Yes, they should be separated. MMI workshop at Huntsville shows some of the MMI products, Registry of Ontologies.pdf shows how easy it is to submit an ontology, a mapping to a registry and then you can see relations because you submitted the mappings
SPARKL interface available now. Will be enhancing the MMI registry or using the BioMed Registry (Stanford) which is incredible. Overall, it will be beneficial for ICAN to use their registry, good linkage to MMI too.
DOCUMENTATION
- We will need to get complete drafts of the various documentation needs
by the end of the year before we move into major recommendations for the
next phase of ICAN
Liz has great start with her outline of tasks
Next month's conference call can look at that and assign writing tasks
Yassine and Declan already have a lot of tasks, so we still need to be sensitive to divvying up the work
Liz will email cookbook doc to everyone for discussion
How much time do we want to spend documenting version 1? Perhaps merge tasks between version 1 and 2
Recs for next phase - on hold
Dawn will send around the benefits list to ICAN Tech
MISCELLANEOUS
- Skype calls - Liz will oversee
- Contact GeoNetwork people - Yassine still working on it; Paula Carerra no longer on GeoNetwork team; needs to contact Joeren Titchnor is the main contact person actually, and Tanya has his email. Yassine will still contact Paula tomorrow and try to contact Joeren as well
SUMMARY
- version 1 finished by end of October (fixing bugs, flash page)
next we will review what we have done, esp. in terms of ontologies (Yassine and Luis to discuss methodology and make sure it is standardized for new nodes; document that)
Liz working on general cookbook for new nodes coming online
Yassine working on new technologies such as Flex
End of December - documentation and review stage to discussion of version 2 in January 2009
WMS stage set up of server can be done now but bigger issue of plugging into atlas needs more development work and be pushed into January 2009, as well as WFS
WFS is still on the table for prototype 3 so need to discuss more about when it should fit in
New GEOSS participation
- in GEOSS response we said that we were going to be a community registry - one community registry to link to GEOSS, but we don't have an API or interface for this just yet
- we need a CSW for global atlas that encompasses all of the other CSWs - so that is the idea of the CSW mediation, not implemented yet, but will be in version 2 where the atlas exposes a CSW service so that you can send a query. You can send a CSW to an atlas, it rewrites to a local CSW request, local sends answer back to application making the request
- GEOSS meeting in September in Boulder; Dawn and Luis will report back
- GEOSS registry should be independent of atlas interface; so in theory it should be easy to set up a separate CSW node; perhaps aim for January
- Dawn and Luis will email us after the Sept 25-26 GEOSS meeting to fill
us in on the relavent GEOSS related deadlines.
NEXT CALL?
7 October 2008
Follow-up 20080627
Here is the newer version of the Oregon Coastal Atlas ontology:
www.coastalatlas.net/ont/20080627/theme
Follow-up 20080618
Here is the newer version of the Oregon Coastal Atlas ontology:
www.coastalatlas.net/ont/20080618/theme
Follow-up 20080611
Here is the URL for the Oregon Coastal Atlas GeoNetwork site:
www.coastalatlas.net:8080/geonetwork
Tanya still has plans to add some more metadata records, and to add more ontology keywords to the metadata. She also wants to add logos and thumbnail images to the records.
She will do this on Thursday after meeting with Dawn about the local ontology. She will also be adding the WMS URL to the "Online Resources" section for all the records.
Meeting 20080610
Skype conference call on ICAN Prototype
Tuesday, 10 June 2008, 1200 ET, 0900 PT, 1600 UTC
Participants:
Yassine, Declan, Katherine, Ned - CMRC
Dawn - OSU
Tanya - OCMP
Liz - Washington Dept. of Ecology
Luis - SURA
** Seek to finish prototype 1 week before workshop to have participants demo
** June 30 deadline for completion
Issues to discuss:
-- (1) Improving the global ontology --
- only 3 items at the moment - need to expand this a bit (Coastal Change Topic)
Human Responses to Coastal Erosion
Agents of Coastal Change
Effects of Coastal Change
- need to refine key words, we have 50-60 and it would be nice to refine this to around 10
- we did have levels below each of the 3, are we taking the 3 and break into sub-categories?
Yes, we are refining the categories and breaking them down
Tanya and Liz were discussing this already at the early stages
We started with top level only at early stages, but now would like sub-categories and to map to local ontologies
We already have a lot of terms from ontologies, could we create subclasses based on what we have in the other ontologies? We guaranteee that the categories that we are making result in a successful search
Create subclasses of 3 term, then map between them?
Tanya and Kathrin can exchange ideas
ACTION: Tanya will try to find list of sub-terms that we had earlier and forward it to the group, Kathrin will feed into process, Ned and rest of group will give input
** See Tanya's separate email "Global ontology suggested sub-terms"
----------
Hi Ned and Katherine,
In regards to refining the global ontolotgy terms (for coastal erosion), I referred to our earlier notes from October 2, 2007:
workshop1.science.oregonstate.edu/cv
If you recall, we arrived at the current 3 super terms in the end by adopting terms from the USGS controlled vocabulary for coastal erosion:
mrib.usgs.gov/user/cch/meta/hot_topics.txt
Since this controlled vocabulary is fairly well developed, I would suggest that we select our new subterms for the global ontology from the "next level down" on this list.
Does this sound like a good approach to you guys? To give you an idea of what this looks like in a more simplified view take a look at the following:
Agents_of_Coastal_Change
- Natural_Processes
- Human_Activities
- Sediment_Budget
Effects_of_Coastal_Change
- Barrier_Island_Migration
- Emergence
- Flooding
- Habitat_Alteration
- Saltwater_Intrusion
- Shoaling
- Shoreline_Accretion
- Shoreline_Erosion
- Submergence
Human_Responses_to_Coastal_Change
- Disaster_Recovery_and_Reconstruction
- Emergency_Response
- Legislation
- Mitigation_Strategies
- Policy
- Preparedness
The above only goes down one level, but it gets us pretty close to Yassine's suggestion of 15 terms in the 2nd level of the global ontology. Of course, these are not terms I am wedded to - so any suggested changes that you folks have are most welcome.
let me know what you think,
Tanya
-----------------
-- (2) Improving local ontologies --
ACTION: We need to have last look at local ontologies to check quality of relationships and themes. We may need to remove some themes as there may be too many.
- What criteria to use? relevance-relevance? Probably redundancy in the terms, and relationships can be messy sometimes (need to make sure that that is true). We need to provide new links between local and global ontologies.
ACTION: Split into 2 groups for refining of CMRC and OCA local ontologies respectively. OCA group will do separate Skype call or email exchanges.
ACTION: Thursday, June 12th works for local OCA email exchange between Tanya, Dawn, and Liz, and they will follow up with a late afternoon Skype conference call hosted by Dawn, if necessary
-- (3) Enabling Web mapping using MapServer
Question for OCA team regarding GeoNetwork server (e.g., server failed during IODE workshop in Belgium in February). That was close to a period of server transition for OCMP, but it should be up and stable now. Tanya just upgraded to GeoNetwork 2.2 and still needs to migrate metadata (can do this by end of day). Will use same keywords as defined in the ontologies, use the same URI string, just the keyword name, not the whole URI.
- Tanya will send URL for new
- There is a WMS set up, not very many layers yet. MIDA has WMS layers enabled and will soon tune it.
- will synch layers in GeoNetwork with WMS
- storing URLs to WMS in metadata - a sticking point. Where to store WMS URL within GeoNetwork (look into online resource - Yassine will send Tanya the exact section that talks about that)
- Current prototype has URL of WMS in the registry of atlases, so may not need to have the URL in the metadata. The name of the layer is key and should be stored according to instructions in online resource of GeoNetwork
Prototype does not do mapping yet, just metadata return. Mapping interface not as important. We talked about WCS at first. We need to get WCS up and running. WMS as a second stage can be discussed at the workshop. There are no huge interoperability needs at WMS side anyway.
Is prototype still pointing to MIDA records exclusively? Yes, needs to be reconfigured to point to true OCA metadata.
-- (4) Integrating the VLIZ atlas: Creating an ontology for VLIZ Atlas, etc. --
VLIZ is Belgium organization of Simon Claus
Interested to see if we could integrate boundary layers such as EEZ in interop prototype; worth doing in context of EEA conference. To have a third partner from Europe would be very good, should foster further enthusiasm.
Send URL for atlas, could we create a couple of keywords for the mainly base maps and layers
Wisconsin missing terms and developing a full-blown ontology, and they are quite far ahead in terms of coastal erosion theme. Would not be hard to bring up to speed, has a bunch of metadata lined up.
CMRC will mentor VLIZ, OCA will mentor Wisconsin Sea Grant
Adding terms to global ontology - everyone should be involved in that, both sides, US and Europe. We should add a new theme to the global ontology about admin boundaries and related. This is in addition to coastal erosion. You always needs those coastline, admin boundaries anyway.
ACTION: "administrative boundaries" is a good term to start of with - sufficiently restrictive that everyone knows what it means and can develop keywords based on this such as:
- country boundaries
- EEZ
- territorial sea for US?
Luis is the only one who has access to the global ontology. You can always download and make changes and send back to Luis to upload it.
Luis suggests putting this into MMI ontology registry (http://marinemetadata.org:9600/registry)
using the MMI services such as ontology registry, so you could have an ontology, change it, and then re-register it to MMI registry.
Locally, have a local ontology copy, work on it and then upload it
Host global ontology on CMRC server but then register it on MMI
Luis just sent email of 3 ontologies (2 local and 1 global) to assist
------------
Here is a screenshot of the 3 super terms and coa and mida term. Thoughts:
I think we need to define better what is Agent, Effect and Human Response. For example we have coa:Bathymetry under agent and mida:Bathymetry under effects. Is this an issue ?
------------
(4.1) Improving MMI ontology registry to meet ICAN needs [Luis]
- with another group coming into the interop prototype, where are they going to register their ontology?
registry of ontologies and registry of CSWs - we can search on an ontology to help create the user interface
- how to add another atlas? we need their name, their CSW, their ontology, and their ontology mapping, but we need a file or registry to know who are the atlases currently in. Right now this is just a text file on the CMRC server specifying all of the participating atlases. So this is not hard coded, which is good.
We might consider putting this text file in a place that all can access, and possibly in XML format.
(4.2) improvements on the user interface [Luis]
- one request to decrease width of the interface a bit (by 100-150 pixels). Too wide for some of our smaller screens. Perhaps just shrink map a bit.
- If we add sub-terms to global ontology, we might consider a drop-down menu. But it is straightforward to fix the size of the list. Needs to be a scrollable multi-list so that people can select many themes at once.
- Bounding Box Issue: If you do a search on any area with bounding boxes and no results are obtained, we should put a popup saying no records were found. Problem with bounding boxes must be with CSW in GeoNetwork. Known bugs page?
- Also check coordinates as stored in metadata. They must be stored in lat/long. It's possible that they could have been blank by mistake. Or check the SRS so that metadata record is indeed correctly chosen.
-- (5) Other issues --
Anyone familiar with a javascript to create an HMTL tree. Instead of a mere list of keywords, it would be nice to have a hierarchy of keywords. Any available libraries for creating trees?
If so, please send to the list
Follow ICAN workshop and EEA conference plans at
ican.science.oregonstate.edu
Should probably due weekly conference calls from here on in
Next call would be June 18, 19, or 20. Yassine will schedule this further, next week after we see how things are progressing.
Let's schedule in a standard TIME, which will be 5:00 p.m. and 9:00 a.m.
Meeting 20071219
Participants: Declan, Yassine, Tanya
Notes
- we discussed the status of the GeoNetwork installations, the Atlas Ontologies and the Prototype Interface.
- Both MIDA and OCA are in a similar position with regards to their GeoNetwork installations. OCA has moved the installation to it's final destination server, which is currently only accessible internally, but which should allow external access by early January. Both MIDA and OCA need to migrate appropriate metadata into their respective GeoNetwork installations. We discussed keeping the number of metadata records small and roughly parallel in numbers, so as to minimize future work with any revisions to keywords and URIs.
- Yassine showed the early stages of the functional prototype interface. Currently the prototype can perform a query on the ontologies of both Atlases (MIDA and OCA), and return keywords from each Atlas associated with the super-terms being used in the search interface. From here it is a (still to be coded) next step to connecting directly to the metadata in the respective GeoNetwork installations, once that metadata is in place.
- In addition to needing to populate GeoNetwork with metadata, both MIDA and OCA need to take another joint look at the super-ontology, and come up with subcategory terms for each of the 3 existing super-terms. This will enable the search interface of the Prototype to present a user with the option of a more refined search (eg. via a tree of terms), which in turn will lead to more specific search results. So this task (creating sub-categories for the super-ontology) will be one of the tasks that both MIDA and OCA should work on in the new year.
Since many people will be/are gone for the holidays, we figured that we would sort out the next conference call via email after the new year, with the second week of January being the first likely productive opportunity for a call.
Merry Christmas, Eid Mubarak, and Happy New Year!
Meeting 20071113
Participants: Liz, Declan, Yassine, Tanya, Luis
Notes
Interface discussion
--------------------------
web interface will have server registry that contains csw endpoints and the ontology related to the csw
the interface also knows about the super ontology
first screen
----------------
professional logo
search interface with keywords from global theme ontololgy
second screen
--------------------
list of results related to keyword that the user submitted.
each result will have thumbnail, title, description, link to full metadata, and link to WMS
Will we have an integrated map interface with data overlay from Ireland and Oregon? (perhaps in second phase)
Lat/Long WGS84 - common projection to support for WMS
Third screen
------------------
A simple WMS that will display a single dataset over base layers, with
no additional functionality.
In second phase we may choose to have an integrated map interface with
data overlay from Ireland and Oregon, with some basic functionality.
Lat/Long WGS84 - common projection to support for WMS
Building the interface:
-----------------------
Yassine is experimenting with Jena 2's inference engine:
jena.sourceforge.net/inference
Luis will send Yassine an example for importing multiple ontologies into Jena 2
MIDA is taking the lead in the prototype interface:
- Liz will work on the graphic aspects of the interface
- Yassine will work on the inference issues that generate the multiple quries to the CSW servers
- Declan will work on the CSW responses
We'd tentatively like to have a prototype up and testable by Christmas
Next conference call meeting is Nov 28th.
Meeting 20071031
Participants: Liz, Declan, Yassine, Tanya, Luis
Agenda
- ontologies status
- status of CSW /GeoNetwork implementation
- URI encodings in ISO. Summary of comments here:
General Comments
- The ontologies will be cleaned in the next weeks
- Tanya will test a CSW client in the next weeks
- Agreement on URIs encoding. See proposal 3 here.
- Discussion started about the web interface. It should have a query form and the results could be displayed as records, each record containing a tile, information and a thumbnail. There could an option “add this layer”. The areas are small compare to the a world map, so it will be difficult to displays oregon and an ireland layers in the same map. A world coast line was identified to be used as a common layer. A page in the wiki will be opened to continue this discussion.
Meeting 20071017
Participants: Liz, Declan, Iban, Tanya, Roy, Eoin, Luis
Agenda
- status of CSW /GeoNetwork implementation
- URI encodings in CSW, WMS and WFS
- review semantic mediation steps
- discuss ontology status and mapping
Notes for discussion were posted here: http://workshop1.science.oregonstate.edu/node/81
General Comments
- Tanya was able to get Metadata into Geonetwork, but still haven’t tested the CSW service.
- Agreed to put URIs inside CSW responses. Will check how SeadataNet is doing it and discuss it via email.
- Semantic mediation steps looks good. CSW will contain metadata about WMS. This metadata will contain the URIs. We will not be concern in the near future about WFS and WMS exposing URIs.
- Discussed ontology experiences. The ontologies currently contain more than what is needed and not all the terms will be map to the super ontology. This is OK if the mapped terms are currently representing available layers. Also one term could be mapped to more than 2 super terms.
- Portal task will begin in 2-3 weeks. Before that we will not need to access CSW. Firewall problems could be resolve then.
Meeting 20071003
Participants: Liz, Declan, Iban, Tanya, Roy, Eoin
General Comments
We had a short but good discussion on choosing the super terms, referencing the summary email from yesterday. In the end we chose to revert to using the three from the USGS controlled vocabulary:
- Agents_of_Coastal_Change
- Effects_of_Coastal_Change
- Human_Responses_to_Coastal_Change
It was great to have Roy on the line to say that he liked the clarity of these terms. We wanted to request that Luis create a new ontology from these so that Liz and Tanya can begin mapping the respective Atlas terms to that ontology.
Meeting 20070912
Participants: Liz, Tanya, Yassine, Luis, Iban, Declan
General Comments
- Creation of the ontologies took about 3 days. Problem cleaning the terms, creating the hierarchy and manually creating them in Protege.
- USGS hazard ontology is more geophysical and lacks of demographics.
- Main topics that we know we could have data for are: human responses and coastal vulnerability. The later could be further subdivide into shoreline types and different demographics.