Bill Howe's blog

A design goal for the Product Factory was to ensure that a permanent url existed for every product. Although every image and dataset generated through the factory is addressable in this way, "mashups" and HTML forms involving factory products must be manipulated interactively.

For example, a CMOP investigator was recently interested in inspecting Cast 77 and Cast 83 from the August 2007 cruise.

On her behalf, a data curator might gather up the relevant links like this:

August 2007, Barnes, Cast 77 profile plot and CSV dataset

Hourly data from a variety of sources extracted from the NANOOS database are attached.

We can easily create factory products to expose this data.

Further, these datasets ought to be extractable by a single query rather than one for each platform. Quarry supports this kind of interface.

Each platform generates measurements with a potentially unique signature. Quarry uses a generic RDF-like data model to enable uniform access to everything, but partitions incoming data by signature to maintain interactive performance.

We've use

Jim Gray, a Turing award winning computer science researcher at Microsoft Research, has used a "20 questions" methodology to gather data management requirements from scientists in Astronomy, Ecology, Materials Engineering, and more. Each of these efforts led to very successful projects.

The idea is this: You all, collectively, provide 20 science questions about your data that are difficult or cumbersome to answer using existing software. I take those 20 questions and your data and design an online system that can answer each one.

Yesterday, I talked about the goals of the product factory, and made a few recommedations for how to use the CMOP product factory -- that is, what kind of products we should add to our repetoire.

Today, I want to set a roadmap for refactoring the factory itself -- What changes do we need to make to the design of the Factory based on feedback?

Observations about the Factory

• The database-powered AJAX web interace feature turns out to be very useful in other contexts besides the Factory. Currently, it cannot be used in these other contexts.

I thought I'd write down some thoughts about the design of the product factory I've been pushing here at CMOP.

Original Goals

Goal 1: Provide user-configurable, real-time web products
How: Replace static, pre-generated products with products generated directly from query results.

Goal 2: Simplify the development of internal and external web applications involving CMOP observations.
How: Provide a web service so that including a configurable CMOP product in your application requires just one line of code.

Goal 3: Significantly reduce the amount of code required to create a new configurable CMOP product.

We're experimenting with Google Calendar as a low-cost meeting planning solution. Here's my calendar:

There were two keynote talks this morning: BJ Penn, Assistant Secretary of the Navy, and Richard Spinrad, Assistant Administrator at NOAA.

Mr Penn spoke first, but I want to start with Dr. Spinrad. He spoke about the interoperability mandate for IOOS and GEOSS, but stressed that the design must be couched in the value to society -- which I interpret as emphasis on user-centered design.

Dr. Spinrad spent a few minutes on the importance of unmanned vehicles (aircraft and seacraft) for the 3 Ds: jobs that are Dull, Dirty, or Dangerous.

I'm currently away at the MTS IEEE Oceans 2007 Conference in Vancouver, BC.

My talk is unfortunately in the last session of the last day, which is later than most "last sessions."

So, I've made up a flyer to advertise my talk. This strategy seemed to work pretty well when I was in a similar position at VLDB 2004 in Toronto.

I've also created a job description flyer for our Web Services Engineer position that I hope to hand out.

University of Utah, view from "The Point"

Last week, I had the opportunity to visit the Scientific Computing and Imaging Institute at the University of Utah, where I met with Claudio Silva, Juliana Freire, and their students.

Our goal was to exchange information and set a strategy for integrating VisTrails, the excellent provenance management software, into daily operations here at CMOP.

Here are profile timeseries plots generated for some coastal transects performed by the R/V Wecoma on the August Cruise (login required).

Two limitations of this product are the x-axis is time rather than longitude, and that there is no interpolation to produce contours.

Each of these plots was generated using the parameterized Cast Timeseries product specialized to the Wecoma and the August 2007 cruise.

CH Line