Dear all,
Sorry for not having answer before and thanks for starting moving this. First Some news from the organizational side.
1) Lee just published a news item about the meeting on the TDWG website. Check . I sent him a joke as the beginning and he finished publishing it!, whatever. 2) People who have not send the request for travel please do it ASAP. We just got the news that the cost of the hotel will cost 65$ per day breakfast included so use this figure for your estimations. 3) the hotel will be booked very kindly by CRIA so please send me your staying dates at the hotel to ask for the reservations. 4)I am planning to stay until Sunday. Would you like that we organize something together to do on Friday and Saturday? We could go somewhere. Who would join this? Maybe I post this in a different email.
And now the most important stuff, replaying your messages:
Do we have any documentation of our goals and how we can achieve them?
I believe our goal is to try to see if we can achieve an implementation that meets the simple use case scenario outlined on the wiki here:
http://wiki.tdwg.org/twiki/bin/view/Geospatial/NicheAnalysisProcess
If I can summarise more succinctly I believe we have 3 goals:
- Build a foundation of 'glue code' in python to allow the various
services to exchange data 2) Build a simple web interface that allows the execution of the simple use case scenario 3) Document and report on the issues for interoperability we experience as it relates to TDWG standards / emerging standards, and web services out there in the big wide world.
In my opionion 3 is the most useful outcome - its a barometer reading on how interoerable the various TDWG initiatives are, and an indication of where we could focus more effort in the future.
I agree with you Tim. I actually see it very much as creating a workflow prototype where the scenario is demonstrated. Do you know Yahoo Pipes? (http://pipes.yahoo.com) I would say we have to create a pipe to chain different services using the different standards. This workflow has to be implemented on some sort of interface, with Python on the background and some cool interface, that makes possible to run it. As Tim we have to document how this worked out, where are the biggest troubles on interoperability among the standards, etc.
We have the basic scenario on the wiki which involves:
1) Use some taxonomic API service to resolve names 2) Get data from GBIF (probably using the new web services or WFS) 3) Run a model with the data 4) Present the results
But we can expand it as much as we want to demonstrate even more things. So we could say that the previous is just our base for the output but if it goes well we would like:
1) Use some taxonomic API service to resolve names 2) Get data from GBIF (probably using the new web services or WFS) 3) Get data from an specific data provider to complement the dataset using TAPIR for example 4) Retrieve a list of environmental layers from a Catalog Service that can be used for modelling 5) Run a model with the retrieved occurrences and the selected environmental layers 6) Present the interface with LSDI resolution of the occurrences when click on them. 7) Save the result on an archive (kind of spatial digital library) and register them in a catalog service.
I see here involved:
1) Taxonomic services, maybe TCS? 2) Occurrences API, TAPIR, WFS? 3) TAPIR 4) Catalog Services 5) Open Modeller API 6) LSID, maybe WMS and WFS? 7) catalog services
This complete workflow is a little bit ambitious, but I want to keep it in mind. What do you think?
Dave Neufled also proposed this:
"I’m imaging a python script located in Spain submitting a list of LSIDS and x,y coords to the EVE service in Boulder (environmental values extraction service) which distributes the query out to WCS servers at KU, CRIA, others and makes a CSV text file of specimen LSIDS, followed by the environmental values available back to the calling python client. Any interest?"
That also sounds cool and that looks more like a big service where different standards are used.
How exactly do we plan to work together? Huddle in a circle, everyone staring at their own laptops?
I would suggest we form 2 man teams and 'buddy program'. Day 1 we will do some planning. My suggestion would be that we hive off into smaller groups to mock up web ui, api etc. The following days we start implementing. Because the time is short we can do everything 'live' dropping code and web stuff into svn as we create it and deploy it on a test server at the same time. Pretty much I am suggesting some 'extremely extreme' programming.
I support you Tim. I would add that I am arriving some days before and will be working on setting up things so that we can start faster really working. What we need is that people really takes care of some parts of the "workflow" so they are informed before arriving on how they would do the task, it would be great not to have to do research there.
What do we expect to have at the end of the week?
A python framework that can be repurposed for e.g. non web apps A web application A report or outline of a report (to be finished offline after the meeting?)
The first one I am not so sure, it could be that the whole thing is just a fast prototype, but if we have some time before Tim we could maybe start setting it up a little bit better so that we get what you want. I am working right with a new technology called Flex, that compiles to Flash, to create RIAs. Maybe we can use this to create the interface as it looks quite impressive, it support several Maps APIs, it has very good web services support and it just looks better than HTML, but we can discuss this, I dont mind.
What do we expect to present to TDWG 2007?
A demo of the web app and the findings in our report.
Agree.
Can it be achieved in one week or will we need to continue working on it remotely?
I would think that anything we dont comlpete will take some commitment from others to invest further time which is probably unlikely. So we should aim to have something working in the week, with perhaps the report being communaly completed after the week?
I think the basic workflow demo at least must be accomplished, after this this sky is the limit :) I might have the possibility to put some resources latter to make the interface look nicer, I have several designers here at my company that could help us a bit.
Do participants have the ability to donate time after the workshop or will their efforts be supported by other projects? Aimee
Its possible that I could invest further time if the outcome was relevant to openModeller project - Id need to discuss that with Renato et al.
As I said I would not mind to spend some resources after, but we have to be realistic on the available resources.
Thanks Aimee and Tim for starting the thread.
See you all soon.
Javier.