February 2002 Programmer's Challenge


Mail solutions to: progchallenge@mactech.com
Due Date: 11:59pm ET, Friday, 1 February 2002

I don’t know which is more puzzling, the fact that people write these things, or the fact that I read them. Or perhaps it was just a slow news week, even for Wired. But when your daily headline email is headlined "Sexchart", you’ve just got to follow the link to see what’s on the other side. And I’m always on the lookout for a novel Challenge idea — you never know where they might appear.

You can check out the link (http://www.wired.com/news/culture/0,1284,48997,00.html) yourself, but the gist of it is this. Someone got the idea of showing just how, well, "connected" one of the more promiscuous members of the internet world was by representing individuals as dots on a two-dimensional graph, and connecting them with a line whenever they had, er, "hooked up". Over time, the graph grew to more than 1400 individuals and their assignations. The definition of what is required to connect two people with a line is provided in the referenced article, something about what sorts of disease can be spread by the contact, but that’s not relevant for our purpose.

So what does this have to do with the Challenge? If you look at the SexChart (http://www.attrition.org/hosted/sexchart/sexchart.9.25), you can see that it has some deficiencies. First, it’s done in ASCII, which limits the connecting line segments to a few orientations provided by a small number of characters. Second, the graph has grown up over time and, frankly, it has gotten a little messy. Certainly we can help. Your Challenge this month is to produce a better SexChart.

Entries this month will be complete applications, so there is no prototype for the code you should write. Your program must process a sequence of test cases, the number of which is provided in the file SexChart.in. The input for test case NN begins with a file (namesNN.in) containing the names of everyone in the graph. The first line in the file is the number of names it contains, followed by one name per line.

Kissin Tell
Alan Medusa

Wicked Pixie

The rest of the test case input is a file (hookupsNN.in) indicating who is connected with whom. Again, the first line in the file is the number of relationships that follow. The file might look something like this:

crank,Alan Medusa
crank,Handsome Harry
Metalchic,Kissin Tell
Metalchic,Handsome Harry

Handsome Harry,Wicked Pixie

Your objective in this Challenge is to place all of the names on a graph and connect them using a multi-segment in a way that minimizes the number of intersections among the connecting lines. If the line connecting names A and B intersects the line connecting C and D, you earn one penalty point for each place that they intersect. Collinear segments (not desirable) earn one point for every unit of overlap length. Your code should produce three output files. The first one (locationsNN.out) contains one line for each name on the graph, with the integer horizontal and vertical coordinates at which the name is placed:

100,150,Handsome Harry

25,150,Wicked Pixie

The second output file (segmentsNN.out) contains a sequence of points connecting each pair of names in hookupsNN.txt. The first and last point in each sequence correspond to the locations of the names being connected. So, for the line connecting crank with Handsome Harry, the segment file might contain this …


… followed by this for the connection between Handsome Harry and Wicked Pixie:


The final output file (logfile.txt) should contain one line for each test case, containing the amount of execution time in milliseconds required to process each test case, measured from before the input is read through the time the output files are generated.

Your solution should display the graph generated for each test case, showing the location where each name has been placed, and the line segments connecting each specified pair of names. You should provide the ability to resize the window in which the graph is drawn, and vertical and horizontal scrolling capability if needed to see the entire graph. You should provide menu control allowing the user to advance to the next test case. You can provide other capabilities as desired to improve the discretionary part of your score (see below).

Scoring will be based on minimizing the number of points your solution incurs. You earn one point for each nontrivial intersection of line segments in each graph that you generate. You incur an additional point for each 100 milliseconds of execution time your solution requires to generate the graph. The time required to display your graph after it has been generated does NOT count and does not incur points. However, the quality of your display along with any additional features provided in your program may earn you a bonus reduction of up to 25% of your points. The bonus will be awarded based on a subjective evaluation of the quality and attractiveness of your application.

This will be a native PowerPC Challenge, using any of the following environments: CodeWarrior, REALbasic (version 3.2.1 or earlier), MetaCard (version 2.3.2 or earlier), Revolution (version 1.1), or Project Builder. You may use another development environment if I can arrange to obtain a copy — email progchallenge@mactech.com to check before you use something else. You can develop for Mac OS 9 or Mac OS X. Your solution should be a complete Macintosh application, and your submission should provide everything needed to build your application, as well as documentation of the features you have implemented, to ensure that I don’t overlook anything.

Test data for this Challenge is now available.

You can get a head start on the Challenge by reading the Programmer's Challenge mailing list. It will be posted to the list on or before the 12th of the preceding month. To join, send an email to listserv@listmail.xplain.com with the subject "subscribe challenge-A". You can also join the discussion list by sending a message with the subject "subscribe challenge-D".