husoski

A question popped up a few days ago about how to view data from the JSON files in Y!A user data downloads.  Here's a simple Python program to summarize the contents of the "answers.json" file, if anybody wants it.

I was somewhat surprised that my totals were a bit *less* that what's on the US Programming & Design leaderboard.  Very close, even though those numbers haven't changed in at least a year.  Back-end code for this board was worse than I guessed!

#

import os,sys

import json

import collections

ya_ansfile = open('answers.json', 'rb')

ya_ansdata = ya_ansfile.read()

ya_ans = json.loads(ya_ansdata)

# count all answers flagged as "best" ("favorite")

print(sum((a["userAnswer"].get("isBestAnswer")) for a in ya_ans))

# Get counts of answers and best answers by category:

allcount = collections.Counter(a['category']['name'] for a in ya_ans)

besties = collections.Counter(a['category']['name'] for a in ya_ans

                              if a['userAnswer']['isBestAnswer'])

# Print category totals sorted by category name:

print()

print("  total   best  category")

for cat in sorted(allcount.keys()):

    print(" {:6d} {:5d}   {}".format(allcount[cat], besties.get(cat,0), cat))

# Print category totals sorted by descending total answers:

print()

print("  total   best  category")

for cat in sorted(allcount.keys(), key=lambda c: allcount[c], reverse=True):

    print(" {:6d} {:5d}   {}".format(allcount[cat], besties.get(cat,0), cat))

All I get for several days is "Oops... Something went wrong", followed by a list of really dumb "trending" questions.

My lip goes into involuntary curls when I read this, but since I have been above median age (except, perhaps, in Boca Raton) for decades now, I have to ask...

Is this legitimate, now? Has "criterion" gone the way of "agendum" and "alumnus", or is this still considered ill-informed usage?

This problem from a recent edition of Halliday & Resnick (& whoever)...

In the x-y plane there are equal charges +q placed on the y axis at (0,D) and (0,-D). Let E(x) be the field at a point (x,0) on the x axis. Let a = x/D be the x position scaled to the y distance D.

(a) Find the value of a where E(x) = E(aD) is a maximum.

Note: This is eminently fair, and the answer is a = 1/√2. It's a pretty good elementary problem. The questionable part comes next:

Find the values of a where E is one half the maximum value. (b) and (c) ask for these two values. The Wiley website asks for something like 2% accuracy, and accepted a numerical solution obtained with the use of technology.

Is it fair to have a problem that can't be solved analytically using methods that the student has studied? Unless I'm missing something, this involves solving an unfactorable cubic, and nobody tortures lower-division math students with the cubic formula. They get an introduction to numerical integration in Calc I, typically, but most won't have taken a numerical analysis class.

So, is there a way to get a 2% answer to this without technology using methods that a student would be expected to have covered?

If not, is it sensible to even ask such a question? By today's standards, of course. When I took the course, it would have been unthinkable.

How many isosceles triangle can be made in the x-y plane that satisfy all of:

a. Integer coordinates,

b. Area = 9,

c. A vertex at the origin.

Two congruent triangles are different for this problem if they are not coincident. That is, if at least one vertex of the two triangles is not common to both.

This is motivated by an earlier question that I "pre-answered" and was prematurely awarded with a BA (undeserved). See:

http://answers.yahoo.com/question/index;_ylv=3?qid...

I ran into an odd thing yesterday when I was dusting out my system. After closing the hood and hooking up all the peripherals, I got a boot device error. The problem was that one of the SATA power connectors was disconnected. I reseated the connectors on both drives and they both feel a bit loose--much less insertion (or extraction!) force than I remember when building the system a little over a year ago. It seems like a mild problem to me and I wonder if it's a known problem with certain connectors, a thermal problem in the case, or what?

It's a CoolerMaster PS, around 500W, and WD Caviar 250G SATA drives.

This is the original valve installed in a new home around 1982. I can remove the handle and trim plate, but can't see any identifying information on the valve, plate or handle? It's a shower-only (no tub) single-valve setup: pull-push for on-off, twist for temperature control.