karlmsmith
commented
6 years ago Comment by steven.c.hankin on 13 Jun 2008 00:33 UTC Filling in this note because we didn't have a TMAP meeting this week.
I asked Roland to take a crack at solving this trac ticket using GWT. A good small-scale project to see how well GWT does. In the first cut the solution should side-step the complexities where it can -- so that we have something in hand soon to show to Bernie Kilonski (Hawaii Sea Level Center). Some of the complexities that Roland and I identified on the phone today are:
- a "time series collection" is regarded as a single dataset in both our (theoretical) LAS XML data model and in practice, for those sites that keep the data in a database, rather than netCDF files. Can we find a uniform strategy to handle both methods of data management, so the LAS users does not see the difference? (In other words regarding all of the time series in the collection as a single "dataset" (object).)
- multiple-variable plot questions. If the user requests plots of sst, airt, and wind speed at once (in a single LAS product request), how should we handle this? The simplistic (short term) solution is to only allow requests for single variable time series plots.
Reported by steven.c.hankin on 10 Jun 2008 22:11 UTC We will soon be encountering (more) LAS sites that are collections of time series. Lets take the simplest case, where these time series are contained each in a single netCDF-CF file and there is a THREDDS catalog above them. What is desirable would be that in a single addXML command an LAS could auto-configure itself to this collection. Today already LAS could configure each of the individual files as a separate dataset. What it seems we need to add is an interactive XY map product that
In version 6 UI thinking the interactive XY map would be a compound product: i) a visualization product that depends on ii) a data access product. The data access ("ii)")would not be from a database. Instead it would simply harvest the XY location of the individual time series that are available from the LAS XML and create a netCDF file from these. Then in the visualization product ("i)") Ferret plots the points, and a Velocity template (JavaScript) sets up the clickable behavior.
The V7 UI might provide an alternative (and better) way to do this. For example, the list of XY locations could be returned by an Ajax service, so that the UI could display the time series locations on the navigation map (using Google Map). The key thing is that the system would autoconfigures itself to a THREDDS collection of time series files. Some examples are
Migrated-From: http://dunkel.pmel.noaa.gov/trac/las/ticket/522