Ten years of the online exposure age calculator formerly known as…
It’s recently come to my attention that I, like everyone else, unfortunately seem to have completely missed the ten-year anniversary of the online exposure age calculators formerly known as the CRONUS-Earth online exposure age calculators(*). Fortunately, the possibly more significant milestone of 1 million calculations served hasn’t happened yet: I expect that to happen sometime around the end of 2017. In any case, though, here are the latest statistics on usage of the online calculators at hess.ess.washington.edu. The following plot is denominated in number of exposure ages calculated per month — that is, if you submit ten samples at the same time, it counts as ten — and I have made an effort to exclude samples provided as test data and also samples that are obviously synthetic (e.g., having zero or integer latitude and/or longitude). So this should be a decent approximation of how much actual use is happening. In any case the following shows more than ten years of usage by month.
The axes are for exposure age calculations (top), erosion rate calculations (middle), and production rate calibration (bottom). The color-coding reflects the various calculator versions: gray is the short-lived version 1; green the also fairly short-lived version 2.0; red the excessively-long-lived-and-well-past-sell-by-date version 2.2, and blue the recently introduced version 2.3. During the time period outlined in black in mid-2014, usage data weren’t being consistently logged due to various technical problems associated with hardware failures, operating system upgrades, and the switch from MATLAB to Octave. Thus, no data from that period are shown here. Shortly after that time period, the bars for two one-month periods go well off scale and are not shown here in their entirety: persons unknown, for unknown reasons, dropped tens of thousands of samples into the system all at once on several occasions. I have not yet investigated in detail whether they really pasted in that many samples by hand, or wrote some software to do that, or what exactly they were doing.
Agglomerating all the various functions and versions into a single number for total usage gives the following, again in units of exposure ages or erosion rates calculated per month:
There are a few interesting things about this.
First of all, overall usage continues to grow steadily. I guess this is not a total surprise, because even if the rate of new exposure age generation isn’t increasing, the total inventory of data that requires periodic recalculation continues to get bigger. But it does make you wonder exactly where the saturation point is.
As usual, calculating exposure ages dominates: the rate of erosion rate calculations is well behind. But usage rates are surprisingly high overall: average usage for the last year is about 10,000 samples monthly. That’s a lot of exposure age calculations.
Second, there hasn’t yet been any discernible impact from the introduction of two new online calculator services: the new CRONUS-Earth product, ‘CRONUScalc,’ developed by Shasta Marrero, and CREp, developed by Leo Martin, PH Blard, and colleagues. These have been available for approximately a year now without, apparently, making a significant dent in usage of the older system. This may just reflect the fact that right now everyone has to calculate everything three times, using all three systems, to make sure they’re not missing anything important. This would tend to imply that the brutal-competition-and-shakeout phase will come later. We can’t all be Facebook: someone has to be MySpace.
My bid to not become MySpace has been to put together a version 3 calculator with many technical and speed improvements, that does exposure-age and production-rate-calibration calculations and competes directly with CREp and CRONUScalc (but is faster, better, and cheaper, with more features, of course, even though CREp wins the swimsuit competition1 hands down and walking away). I am still working on version 3 and it doesn’t look like it has yet begun to cannibalize usage of the older version 2.x family, but the goal is to make that happen soon so that I can stop worrying about all the somewhat-out-of-date 2.x code. Version 3 usage, on the same time scale as above, is here:
In this plot, the upper panel shows exposure age calculations and the lower panel shows use of the production rate calibration interface (which has only existed for less than a month). Light green shows programmatic requests using the web service API, and dark green shows requests by actual people pasting data into the data entry webpage. The web service requests, which comprise the vast majority of usage, reflect the fact that the version 3 code is the back end for the antarctica.ice-d.org database: any web page served by the ICE-D server uses the version 3 web service API to compute exposure ages, and those numbers bulk up the light green bars here. As this setup generates lots of requests with not much user effort, I am a bit surprised that these numbers are not bigger. However, so far, the direct use of version 3 through the web page interface (that is, the dark green) is nowhere near that of the version 2.x family. I’ve pretty much completed the version 3 code at this point, though, so hopefully that’ll change soon.
Summary: 10,000 exposure age or erosion rate calculations every month is a lot of exposure age or erosion rate calculations.
1A “swimsuit competition” is a typical component of contests known as “beauty pageants” that are popular in many parts of America as well as a few other countries worldwide. U.S. President-elect Donald Trump has been closely associated with promoting and judging beauty pageants in the past. In a beauty pageant, young women compete on the basis of their appearance, personality, and accomplishments, with the aim of winning the right to represent their city, state, country, or other group either in other beauty pageants or for charitable or publicity purposes. The competition involves a series of events, usually including one or more in which contestants appear before a panel of judges while clad in garments such as swimsuits (typically one-piece) or formal dresses. These events highlight the appearance, grooming, and fashion sense of the contestants, in contrast to other events that may focus more on their skills and accomplishments. Although such events have incurred widespread criticism on the basis that they perpetuate obsolete gender roles and validate discriminatory behaviour based on appearance, the term “swimsuit competition” remains commonly used in colloquial English to denote contests or evaluation criteria that emphasize appearance over substance. For example, one might, hypothetically, state that “the selection process for GSA Fellows is a total swimsuit competition.” In the present blog entry, the phrase is used rhetorically to highlight the professionally designed web pages and online user interface for the CREp software, which contrast favorably with the near-total lack of attention to graphic design and visual communication displayed by other online exposure age calculator websites. It is not intended to imply that CREp would not also score well in, for example, the talent or personal interview portions of the competition.