http://slashdot.org/comments.pl?sid=1499856&cid=30673056

I got a Ph.D. minus epsilon in Math, but never took a statistics course prior to dropping out during the dotcom boom to start a software company. I would trade most of the math I know for a solid background in statistics, which comes up all over the place, from the performance of data structures and algorithms, to dataflow analysis, to mining databases, to projecting performance failure rates, etc. This is not just basic statistics (which I managed to pick up on the street), but a deep knowledge of statistics can make a big difference. I just co-authored a paper on a very simple data structure for which the key step in analyzing its performance is to apply the Kolmogorov-Smirnov distribution.

TFA said (as essentially its main point):

Again, this puts hard science in the new position of saying, close enough is good enough. One hopes civil engineers never build bridges under this theory.

Actually, this is one of the most basic principles of how civil engineers build bridges. It's called "margin of safety." You don't build to the worst you can prove will happen. You build to the worst that you can't prove will not happen.

One thing I do agree with the TFA is that the public doesn't understand how science works (obviously neither does the author) and that is creating a huge public relations crisis. Science needs some articulate advocates who are actual real scientists (or at least deeply understand what science is), not the politicians (Gore) and editorialists (Henninger) who seem to be framing the discussion now.

Remember the saying that science proceeds by successive approximation to the truth.

A musician friend of mine told me to check out Les Paul, saying he did everything that Hendrix et al did before then. At first, I had trouble finding his CDs at the local music store until I discovered they were in the "Easy Listening" section. In spite of that (arguably accurate) classification, I have to agree with my friend. What an innovator!

The problem doesn't arrive because you are 40. It arises based on how far and recently you have advanced by 40. If you have been a mid-level programmer for the last 10 years and are no longer advancing in your profession, yeah, you can have problems because your seniority doesn't match your "seniority." Many people will want to hire someone who doesn't appear to have plateaued and therefore is perceived as having more potential (although many companies will respect that basing decisions on being under 40 is illegal, so they won't look at age per se).

  On the other hand, if you have been the lead architect for commercially successful products, demonstrably know what it takes to produce successful maintainable products, have continued to advance in your career, etc., your experience and seniority are likely to be a plus.

IOW, the risk of being in a profession for 20 years is that you'll have a long track record. The benefit of being in a profession for 20 years is that you'll have a long track record.

Actually, Section 3.6.3p1 of the C++ standard guarantees it. (Wonder why people who can't validate technical language claims feel qualified to mod posts that make them).

An Introduction to Number Theory is a fantastic book that assumes no familiarity with number theory. I used it to teach a high school number theory course with great results. Starting from essentially no prerequisites, it reaches important topics like the Chinese Remainder Theorem, and quadratic fields, as well as fun topics like Magic Squares and Continued Fractions. Perhaps the best part is the opening chapter on why you need proofs. He shows this by giving a half dozen examples of results that are "obviously" true (many of which were believed true for hundreds of years) that turn out to be false.

From the memo

While this is still considerable, it's much less than in the headline due to planned new hires. I hate it when the news mentions the number of people let go but never mentions how many are brought on. Headlines like these would lead you to conclude that America has lost many times more jobs over the years than it actually has.

I know very strong and effective managers who think personality tests can be useful.

For example, many of the best salespeople are very outgoing and social but at the same time don't care much what other people think of them (so they can take 30 rejections in a row without being discouraged). The is a pretty rare combination for obvious reasons. If a personality inventory shows that a candidate has these traits, that's viewed as interesting and potentially a point in favor of the candidate (not something to be applied rigidly or blindly but as useful information).

One other point is that the better modern personality inventories are usually pretty carefully designed to detect cheating with questions that the "honest" and "expected" answers are different, so they are much harder to game than many people are saying here.

Of course, putting out a crappy looking resume worked in this case, but it's a bad bet. I'd recommend going with a decent looking one.

Don't listen to the naysayers, it's only January...

I actually taught a class with "Modern C++ Design" as a primary text. Perhaps surprisingly, it turned out to be a great success. While the actual techniques in the book are not always suitable for use in most professional C++ programming shops, the revolutionary thinking and skills in it will serve you well whatever programming you do, even if it is in a different language, which is a hallmark of classic texts.