You could delete the word "mental" from your last sentence and it would still be pretty accurate.

In the era this was written, compilers didn't optimize worth a damn. I totally agree it's not elegant but Duff wrote that to solve a problem that the "elegant" solution just couldn't solve. Even still, Duff's device isn't really that hard to understand if you're familiar at all with loop unwinding, especially if you've ever done it in assembler (and C isn't that much more than a semi-portable assembly language). Sometimes you have to write ugly things because the pretty structured/OO way just doesn't cut it.

Even today, "optimizing" compilers sometimes don't. On the other hand, this doesn't occur as often as some people think. I operate on a policy of "trust your compiler, until you can't."

And they don't ice and plow roads during snow worth a good golly damn. We took an Arizona to Ohio road trip in December and there was a biggish snow storm moving into Illinois, Indiana and Ohio. About the time we reached Effingham. Illinois plow trucks on I-70 were pre-positioning themselves every 5 miles even though the storm had not hit yet. We hit Indiana and there was nothing. The storm actually hit when we got to the western edge of Indianapolis and it took 6 hours to get to the Ohio border (usually a 2 hour journey). The roads were terrible, there was no salt down, and I didn't see a plow truck until about 50 miles in after the snow started. Also, it seemed like every bridge seam in Indiana was not graded with the road. Almost every time I got on or off a bridge I thought I was going to lose a wheel.

When we hit Ohio, despite the storm actually getting worse, the roads were clear (just wet) and I constantly saw the trucks. Because of Indiana though, I was too exhausted to make it to my destination (Canton) and had to stop just west of Columbus. 6 hours of white knuckled driving a 30' RV in a snowstorm kind of takes it out of you. :-)

Also, if you ever drive west, Oklahoma highways suck too, ESPECIALLY around OKC. To hell with that city.

Heh. In 25 years in this field, I've been involuntarily unemployed for 7 weeks, total. With that said, I *do* have my bachelor's from a fairly respected state school. But the places I've worked few really cared about the degrees for mid-career and senior people (entry level is different). The group that didn't care includes a NASA-funded research lab at a university headed by a tenured professor, and an IT systems development group at another university, so there are even examples IN academia that are more interested in what you can do than where you may or may not learned it from. These places had a mix, from people who didn't have ANY college to PhDs, working together. "Can you get the shit done we need to get done and do it well?" is the most important criterion.

As I said, YMMV. :)

In fairness, most of the companies who have policies that work like that are fairly shitty places to work, especially in IT. Let them filter me. YMMV, of course.

And that's "we apologize", not "we're apologize". Derp. Even previewed and didn't catch it until half a second AFTER I pressed "post".

You're hitting it pretty close to the mark here. It comes down to the line between good traffic engineering and violating net neutrality is not a clear one. While I think Something Should Be Done(tm), I sometimes worry the cure may be worse than the disease. I work for a CDN and I can kind of see everyone's side simultaneously.

I think the difference here is in this case, it is Comcast's own customers that are requesting the traffic. Calling Comcast a "peer" like LVLT in this situation is a little bit murky, IMO. If it were purely about network performance, the "right" thing to do would be to charge their customers extra for Netflix streaming. But if they did that, they'd have open rebellion on their hands. And rightfully so, because all the consumer broadband providers hype "stream movies over the internet" in their marketing. So Comcast hits up Netflix for money instead. Somebody does have to pay for the packets, sure. But Netflix has in-demand content, and the cable companies are no stranger to paying for in-demand content, and they've made tidy sums off the consumer while doing it.

Purely from idle curiosity, I wonder what would've happened if instead of paying Comcast, they put up a message to every Comcast subscriber the first time they have a buffering event saying "We're apologize, but your internet service provider (Comcast) doesn't have sufficient network capacity to play Netflix movies at this time. Please contact Comcast technical support number at 1-8xx-xxx-xxxx for help." I think Netflix's reputation with the public is stronger than any cable company. They might have been able to get away with the bully pulpit.

"This saw sucks. I can't drive a nail with it. And what's with this useless hammer? Look at the jagged splintery bits I get when I try to cut wood with it!"

I don't get why there's a holy war between static and dynamic typing in the first place. Different tools for different problems. I don't trust any software engineer very far when they say things like "$LANGUAGE [sucks|rules] because it's [statically|dynamically] typed."

Well, to be fair, some of the guys that came up with scrum methodology were presenting it at OOPSLA 5 years before that, and if I memory isn't totally swiss cheeesed, I recall Kirk being heavily involved with OOPSLA back in the day. But it would be a couple years before I first heard the word "scrum" and a year after that before I realized "Oh, huh. I've done that." :-)

This is spot-on, and true in real cutting edge companies everywhere. If you're 40+ and give a damn about technology at all you don't want to work in an ageist place anyway. Most of them are doing "me too!" boring-ass shit you don't want a part of.

The ones that get it know that age is just a number, and while it may take my 46 year old brain a little longer to catch on to stuff than my 26 year old brain did, it still catches on just fine, and has a quarter century of experience to contextualize that new information against. In fact, I'd argue that's part of why it takes longer: I'm integrating the new information into a broader framework that I just didn't have in my 20s. You wouldn't believe the shit (or maybe you would) that's getting passed off as "cutting-edge" by some people today where I can say, "yeah it was cool when IBM/DEC/Sun/Cray/Pick another old company did that in the 1970s and 1980s too."

Disclosure: I know and have worked with the parent poster (Hi, Eric!). He's not blowing sunshine up your ass. He really is that good. :-)

I'm tempted to say "you misspelled Prolog" but since I've used Prolog, I think I'm actually the one misspelling it and your spelling is indeed more accurate.

"discreet" vs. "discrete". They mean somewhat different things. :)

Which isn't all bad... after all, the Bedouins didn't invent nuclear weapons.

Well, in fairness, if you have immortality, warp drive is less of a concern. What's 40,000 years if you live for a billion?

The one think the immortalists seem to miss is there's going to also have to be some huge advances in trauma medicine (unless you're talking we're to the point of uploading consciousness to robot bodies a la Moravec... that's so much change that if it were to happen hypothesizing on its results would be a series of science fiction stories, your guess is as good as mine what would actually happen.) you're not going to live MUCH longer. I can't find the article now, but I remember a statistical study where if you factor out all "natural" death, either murder or accident will get you sooner or later and life expectancy would still only be 5 or 6 centuries.