Four Important Announcements

Error in Algorithm for Assignment 6

Thanks to the resolute determination of Max Bogen and others, a mistake has been found in the algebra used to find the formulas for p₀ and p₁ from the numerical sums in Eqs. (6) of Assignment 6. The correct algorithm is now shown in both the original Assignment and the Solutions, which are now posted.

You will not lose marks for using the wrong algorithm that I gave you.

I offer no excuse for my algebra error, but it is an interesting example of how such errors can be incorporated into software libraries and remain undiscovered for years until some intrepid hero discovers them through more careful checking. This one slipped under the radar by giving almost exactly the right answer for the test case, in which the y-intercept p₀ happens to be very small. A fit to another straight line with a large intercept would have quickly revealed the flaw.

No Plotting Required in Assignment 7

I forgot to mention that plotting in FORTRAN is notoriously esoteric and requires special libraries that are not relevant to the main purpose of this Assignment, namely to get you familiar enough with FORTRAN to be able to use it for calculations -- which are its main strength. So you should not be trying to port the plotting part of your linfit.py script into FORTRAN; just get your FORTRAN program to read in the data, calculate the best linear fit using the (corrected) algorithm from Assignment 6, and print the results on your screen.

More Marking Feedback on Homework

I have finally realized that our "paperless office" philosophy leaves something to be desired when it comes to marking of Assignments: whereas a hard copy can easily be "marked up" with a red pen, providing a pretty good idea of what was done right and where marks were lost, this is very difficult with purely electronic media.

I will start working on an extension to my marking software that will Email each student a summary of the marking details, but this is a substantial task that will take a while.

In the meantime, I have asked the TAs to prepare a summary of the most common errors on each Assignment after they finish marking; this will be posted on the Assignments page of the website, probably with a link next to the Solutions link for that Assignment. If they are able to remember these details for the earlier Assignments, those will be posted as well. Hopefully this will provide enough clues that everyone will have a pretty good idea of where they lost marks, since most errors are very common.

"Project Week" with No Assignment?

Several people have requested a "week off" to get a solid start on their Projects. My concern has been that such a "break" might just be used to catch up on this and other courses, and thus have no effect on Project progress. (Sort of like how Lottery proceeds meant to support Education or Hospitals get dumped into General Revenues instead.) However, if everyone resolves to come to the Lab with their Project idea in mind and "brainstorm" with other students to get a real start on their Projects, it might be worth giving up a week's worth of course material. (Those not showing up would definitely lose Participation marks.)

There are 4 weeks left (not counting this week) before the final week of Project Presentations, and one is only a half-week (UBC is closed on Nov. 11, Remembrance Day). I therefore propose to use Tuesday 09 Nov as "Project Day" and eliminate that week's Assignment on Linear Equations and Interpolation (see the Course Outline).

This will mean one less Assignment (from which to select the best 8) and one less computational skill you will acquire in PHYS 210, so I think it requires a consensus. Please come to class this Thursday prepared to discuss this option and cast your vote.