Skip to main content

Unintended Benefits of Unit Testing: Documentation for Nothing and Testing for Free

I'm working on the geochrono[1] project a little bit at the time.  I unexpectedly came across a benefit of unit testing I'd forgotten about, documentation by testcase.  One of the first requirements for geochrono is:

The user will be allowed to add events or person chrono-locations by adding markers to a map at the appropriate location.
First implementation 
The user will be required to enter the year, month, and day of the month in three distinct textboxes before clicking on OK. The date will be checked as valid using code available at stackexchange[2].
Testcase: Send bad and good dates to date checker code.
Something like this:


The requirements and testcsases seemed simple enough.  Form a bad date and pass it into the date checking function.  Something like this should have done the trick:


assert.equal(isValidDate(new Date(1980, 100, 150)), false);

You get the idea, pass in a month that doesn't exist, (100), and a day of the month that doesn't exist either, (150).  The problem that cropped up almost immediately was that the Date object is remarkably resilient in creating a date it thinks makes sense even if you hand it garbage.  Consequently, the testcase checking to make sure a 'nonsensical' date was caught by the checker, failed.



There were a number of other combinations that resulted in valid Date objects being created.  This led to the following series of testcases that demonstrated the 'bad dates' that isValidDate will not catch:

QUnit.test( "hello test", function( assert ) {
  assert.ok( 1 == "1", "Passed!" );
  //The following are counter-examples showing how the Date object will
  //wrangle several 'bad' dates into a valid date anyway
  assert.equal(isValidDate(new Date(1980, 12, 15)), true);
  d = new Date();
  d.setFullYear(1980);
  d.setMonth(1);
  d.setDate(33);
  assert.equal(isValidDate(new Date(1980, 100, 150)), true);
  assert.equal(isValidDate(d), true);
  //If you go to this exterme, then the checker will fail
  assert.equal(isValidDate(new Date("This is junk")), false);
  //This is a valid date string
  assert.equal(isValidDate(new Date("November 17, 1989")), true);
  //but is this?
  //Ha!  It's not.  So, the secret to working with this version of
  //isValidDate is to pass in dates as text strings... Hooboy
  assert.equal(isValidDate(new Date("November 35, 1989")), false);
  //alert(d.toString());
});


So, as you can see, the testcases showed that to reuse isValidDate and get the checking behavior I'd like, I'm going to have to submit my dates in the form of a text string.

Of course, the next thing to figure out is how to guarantee that my code always ships in dates in the intended format, but I'll save that for another time.

References:
1.  http://copaseticflow.blogspot.com/2014/06/mapping-historical-events-in.html

2.  http://stackoverflow.com/questions/1353684/detecting-an-invalid-date-date-instance-in-javascript

Comments

Popular posts from this blog

Cool Math Tricks: Deriving the Divergence, (Del or Nabla) into New (Cylindrical) Coordinate Systems

The following is a pretty lengthy procedure, but converting the divergence, (nabla, del) operator between coordinate systems comes up pretty often. While there are tables for converting between common coordinate systems, there seem to be fewer explanations of the procedure for deriving the conversion, so here goes!

What do we actually want?

To convert the Cartesian nabla



to the nabla for another coordinate system, say… cylindrical coordinates.



What we’ll need:

1. The Cartesian Nabla:



2. A set of equations relating the Cartesian coordinates to cylindrical coordinates:



3. A set of equations relating the Cartesian basis vectors to the basis vectors of the new coordinate system:



How to do it:

Use the chain rule for differentiation to convert the derivatives with respect to the Cartesian variables to derivatives with respect to the cylindrical variables.

The chain rule can be used to convert a differential operator in terms of one variable into a series of differential operators in terms of othe…

Lost Phone

We were incredibly lucky to have both been in university settings when our kids were born.  When No. 1 arrived, we were both still grad students.  Not long after No. 2 arrived, (about 10 days to be exact), mom-person defended her dissertation and gained the appellation prependage Dr. 

While there are lots of perks attendant to grad school, not the least of them phenomenal health insurance, that’s not the one that’s come to mind for me just now.  The one I’m most grateful for at the moment with respect to our kids was the opportunities for sheer independence.  Most days, we’d meet for lunch on the quad of whatever university we were hanging out at at the time, (physics research requires a bit of travel), to eat lunch.  During those lunches, the kids could crawl, toddle, or jog off into the distance.  There were no roads, and therefore no cars.  And, I realize now with a certain wistful bliss I had no knowledge of at the time, there were also very few people at hand that new what a baby…

Lab Book 2014_07_10 More NaI Characterization

Summary: Much more plunking around with the NaI detector and sources today.  A Pb shield was built to eliminate cosmic ray muons as well as potassium 40 radiation from the concreted building.  The spectra are much cleaner, but still don't have the count rates or distinctive peaks that are expected.
New to the experiment?  Scroll to the bottom to see background and get caught up.
Lab Book Threshold for the QVT is currently set at -1.49 volts.  Remember to divide this by 100 to get the actual threshold voltage. A new spectrum recording the lines of all three sources, Cs 137, Co 60, and Sr 90, was started at approximately 10:55. Took data for about an hour.
Started the Cs 137 only spectrum at about 11:55 AM

Here’s the no-source background from yesterday
In comparison, here’s the 3 source spectrum from this morning.

The three source spectrum shows peak structure not exhibited by the background alone. I forgot to take scope pictures of the Cs137 run. I do however, have the printout, and…