Cached copy of Camp Roberts Planing 2010 Notes

After Saturday dinner at WhereCamp 2010 (April 3rd, 2010), some friends of Camp Roberts gathered in a quiet conference room to discuss the lessons we learned during the response to the Haiti earthquake. The following Camp Roberts Planing 2010 Notes were a tactical set of notes, intended to help inform the planning for the next two Camp Roberts, followed by Dane's raw notes.

Attending:

Schyler Erle, John Crowley, Mikel Maron, Jeff Warren, Jeff Johnson, Mike Migurski, Kate Chapman, Andrew Turner, Tim Carobruce, Sean Gorman, Josh Livni, Ragi Y. Burhum, Dane Springmeyer, Alan Glennon, others...

Proposed Projects

Walking Papers Atlas

Walking Papers was a success in Haiti, used by many UN agencies for printing maps based on the combination of OSM and underlying imagery. However, the process of printing detailed Walking Papers (WPs) for a large area required a manual, timeconsuming processes. The Humanitarian OSM Team (HOT) also noticed field workers tend to write on paper maps that are used for navigation, no matter what the intentions of the cartographer. From these two insights grew the notion of building a tool to create Walking Papers atlases for a given region, where the notations from driving can be scanned back into OSM and turned into annotations by the OSM community.

OSM Grid-Square Checkout Server (GCS) (job tasking)

Crowdsourcing OSM requires an ability to assign an arbitrary areas to many volunteers and audit the status of their tracings and other annotations. Analogously, USAR teams must checkout MGRS squares for S&R missions and clearing of areas. This tool would provide both communities—OSM volunteers and USAR teams—with a method for checking out a polygon from OSM for annotation, tracing, and/or USAR purposes. The tracking systems for development would enable an audit trail, and would need to clearly distinguish mapping activities from USAR activities. The tool would also create situational awareness around what parts of the map have been built and which are currently assigned to individuals. Analogously, this function would enable multiple international USAR partners to coordinate which parts of the map have been assigned and cleared (or reopened).

OSM Lite

In the field, downloads which are considered small in the office (18MB) are considered expensive and carefuly rationed (especially over BGANs that are split between many more people than usually intended). There should be mode in OSM that enables "lite" field operations, which can support tiles and UIs for low-bandwidth environment. This functionality may include 4-color tiles, reduced (or no) javascript, and the ability to ration which imagery or data gets downloaded by various filters and/or bounding boxes. An enabling technology would be the OSM field extraction tool (below).

Offline OSM

OpenStreetMap cannot assume internet connectivity during response operations. Networking may not be available, or may go down at any time. OSM requires a caching/synching method that will enable local instances of OSM to exchange data with the main OSM servers, while also allowing local access. Particular interest exists in using local tile caches.

OSM Field Extraction Tool

The Field Extraction Tool would enable a field worker to select a region and pull all information in OSM from that region, limiting the data download. In aggregate, these downloads of AOIs could be tiled to a local server to improve performance.

The Sneakernet II: High-speed "Geo Rag and Bones Man" File Transfer Network

Sneakernet has a venerable history, but can be improved. This method would work as follows:

Assumptions

Assume a two or more tents in an area in the field.

Assume a person with a mobile storage device with data that should be shared, such as geo data in layers

Method

Any person who walks between tents (rag and bone man, or RBM) will give data and accumulate new files as he/she moves along a chain of tents.

As the RBM moves between tents, he/she will carry all the new data as well as the original data.

Automated Processing, Cataloging & Release System (AutoProcessCRS, pronounce "AutoProcessers")

A project to ensure that no one has to 'crschmidt' dozens of terabytes of imagery during a disaster. An automated imagery processing, cataloging and release system designed to collate, tile, catalog, and release imagery to the open volunteer geospatial community.

Layer9

The United Nations improved coordination of disaster response through the creation of its cluster system, which aggregates representatives from all organizations according to nine areas of activity (shelter, water, sanitation, health, camp management, etc). The Layer9 project would create basic integration of all the location-based information from these clusters, including where health facilities are, where camps are, etc.

Starting to Prioritize

0-1 day

Issue Tracker

1-5 Days

Job Tracker Osm Multi master "Actually usable continuous OSM extracts" "Clip and Ship" (UI to various downloads)

6 days to forever

Second phase of clip and ship - OSM light interface with low qual tiles, tiny js, etc Walking Papers atlas

Integration

swift river - make sure we're in sync and that this app can handle needs for auto-geocding before manual intervention

Raw Notes

Camp Roberts - next steps planning

WhereCamp, April 3rd, 2010

Attending: Schyler Erle, John Crowley, Mikel Maron, Jeff Warren, Jeff Johnson, Mike Migurski, Kate Chapman, Andrew Turner, Tim Carobruce, Sean Gorman, Josh Livni, Ragi Y. Burhum, Dane Springmeyer, Alan Glennon, others...

Idea: discuss what could be worked on if/where there is another Camp Roberts Exercise.

Running notes on ideas:

walking paper atlas

• make it easier to print

-- surveying area - need whole area

• walking and navigation

OSM tasking server

proposal by Schyler

• area you want to break down into sub areas - to hand out
• digitizing, or ground serveys
• mgrs works, ability to be easily delegated
• paper componetn + online components
• ability to check out a grid square
• kate chapman: project gutenberg ocr (ability to check out grid square)
• go here, create an account, find a red quare, ability to redo a green square, and auditing chain
• so, review and re-review, having an audit trail available in an interface, coordinate in central or distributed way
• Ragi Y. Burhum: JTX (Job tracking extension) - esri tool allows some of these features
• Jeff Warren: (catholic relief services) - ability to grid out an almanac, ability to text in alerts, aggregated on a map
• Lots of crazy grid formats, mgrs is really pretty useful: "mgrs is global - we can invade anywhere."
• automated geocoding of locations. run yahoo placemaker on entire query to see if you come up with any locations - reduce load on crowdsourcing. does swift river do this already (ya probably). also need to leverage osm as much as possible

Lessons from Haiti Visits

Really important for humanitarian world:

• humanitarian osm ticketing system
• ocha thrilled, but nowhere to direct requests
• camp roberts could be change to develop and try out action protocals for central point of contact...trouble ticketing server
• noc, status board, (network operations center)
• hospitals - make sure as many people see request as possible
• easier export of KML and shapefile and garmin
• zoom into an area, get extract of just want you want in any format

- needs to be super lightweight

• osmlight web interface - tiny tiles, no map, search, very little javascript, low color tiles, "somehow slimming down openlayers"
• cleaning up shapefiles, avoiding topology problems. maplint, etc should be part of the workflow.
• need outside collaboraters to camp roberts exersize, so that we see the effect/successes real time and not in a test vaccum. Pick a place that needs mapped. coordinate real imagery donation, tell community its time to stress test, how to coordinate the cloud

Focus on Jeff Warren's work...

• IDP (internally displaced people)- apply jeff warren's ballon mapping tools to tracking IDP's. These things seem temporary and end up becomming cities. Taking slum mapping to its earliest possible stage.
• route for grassroots mapping: ability to get students/kids and innovating folks involved in kite making
• amy smith - mit lab - an illustrated guide to grassroots mapping
• in the absence of helium and with some wind, you can do well.

OpenAerialMap

Summary of basic needs and problems

Open Aerial Map

• "automating chris schmidt"
• warren: oam can cover both community sources imagery and top down release?

problems with oam:

• variable size contributers - MB vs terribytes
• cataloging - where to store, keep track, minimal metatdata (global minimum of upset across community) - "gradient dissent through communityu upset"
• huge upload need, vs small upload, vs WMS - but WMS by mass gis is "slow as donkey balls" (arcserver)
• proxies and fast access - could do this for landsat 7
• how is time going to be considered - access and availability of snapshots
• how to you backup tilecaches? (you don't have to as you have the source imagery, right?)
• processing pipeline - assuming its in a format the server can read, data needs to be warped, pushed into system, and tiled
• ability to warp and re-warp, fix alignment on the fly
• stiching looks hideaous - how to fix loss of trust do to fugliness of imagery (through omar?)
• have a caching server - seed an area needed, pack it up and go offline - one stop solution

 - being funded by world bank and others and written by opengeo - geonode (django adds front end to bolt on upload and cms)
 - actually called "geocrap", as it is a django front end to java tools

• taking things offline - minustah and ocha base - cheek by jowel - basically stovepipe in one tent and not very good at exchanging geodata
• cluster system used by UN - breakdown is functionally broken down, eg. food routing and medicine routing are separate. above that there is no good mech. for coordinating. how can we through locality create cross section across groups.
• rag and bone ma, "have geodata, here's geodata" - walk around all tents giving hard drives and usb sticks
• what is the tool for the rag and bone man, - he shows up with harddirve, and can just sync his data, maybe have OSM tools be able to check disparity
• instead of maps-on-a-stick, oam in a box. give original files (geotiff), not web mercators tiles
• "simple fricking lyr file for users to download."
• robust osm solution offline vs OL pointing at some tiles
• routeme - offmaps iphone map is good ui example for on-the-fly tile caching in device
• offline editing - multimaster sync problem. to avoid constantly sending osm over the wire, unique ids for edits
• if conflicts in editing on check-in, need auto flag for resolve and incorporate into tix system for resolution

PRIORITIES

• tasking server tops, disaster and not
• first pass of synch local instances of OSM might take an afternoon???

AJ ---

• COP - common operating picture.
• ability to deploy com ops on a stick AND UPDATE!!!! round-tripping, gather data and go back.

 - offline ability works well enough to be a round trip

• GeoPDF's were really used (delta state)- markup - lat long hover, switchable layers, queryable
• continuous GPS images. Latest is good, but history is great.
• some really wanted to know how - sneakernetting UBS sticks.
• TODD: how would you build a high speed burstable sneakernet

- trunks driving along a route, ability to have a USB pony-express

• geotagged photos - damage assessment - modile data collections
• pre and post lidar - rumble, drainage, pre-post blockage
• damage assessment tool - which building were still habitible based on groun truthing
• DHD - data hugging distorder - situation in haiti

Mike Arrives

• QR codes - how could we make them smaller
• books - select large bounding areas - tile it out
• chunk up areas - parcel out for focus areas
• SMS updating of OSM, ability to post the road name to OSM while they are out in the field? (mgrs + name of issue)
• ability to use stickers on paper maps that are automatically recognized and automatically digitized
• warren: arriving "I accidentally made some lattees"