Forgot your password?

Comment: Re:Author is dumb (Score 1) 212

by lars_stefan_axelsson (#46765511) Attached to: How 'DevOps' Is Killing the Developer

Also lots of QA people can code - automated testing is mostly coding - and lots of developers can't test at all.

If my experience is anything to go by, then the QA people on average code about as well as the developers can test. Which is to say, everything from "decent," to "not at all," depending on how good the organisation was overall.

Comment: Re:Depends on who uses them (Score 1) 186

by lars_stefan_axelsson (#46765445) Attached to: The Security of Popular Programming Languages

It may be cliche, but how secure a language is depends on who is using it.

Well, this might be one of those times where "correlation is not causation" is appropriate. What I mean by this is that while the skill of the user certainly matters, so does the quality of the tool. "It's a poor workman who blames his tools" is a warning against poor workmen, not a carte blanche for shitty tools.

So of course, a tool can both help you and hinder you doing what you need to. It's no accident that you see top sportsmen obsessing over their equipment. Or why people don't clamour for manual gear boxes without synchronisation rings any more.

Now, when it comes to programming languages these interactions are very poorly understood. We don't know what makes a certain language (or features) work well for certain programmers in what situations. That said, we've amassed quite a body of evidence to suggest that some things work better than others. Functional programming languages with proper type systems for example seems to make large classes of errors just go away, by making them both obvious and impossible (since the type system wouldn't allow them). They also seem to increase productivity by a factor of four to ten, and increase system performance too boot. (Even though they're "slower" than C in micro benchmarks. Basically the same argument against C from the assembly crowd in the eighties, which we all know fell out in favour of that "slow, bloated, high-level mess" that is C...

Granted, much of this is still in the research state, and security among other things isn't important enough to make people change their ways (at least not yet), but there is a lot more interesting done, and still left to do, to just say that the tool/language etc. side of the equation doesn't matter. It clearly does.

Comment: Re:Get rid of income Tax (Score 1) 414

You forgot the bit about only certain countries having any factories left at the end of WWII...

Yes, just see what a horrible place Germany turned into... Losing all those factories really hurt them in the long run.

P.S. Their recovery was also paid for by the US taxpayers, and it seems they got a fairly good deal on that too.

Comment: Re:fixing the parent posting (Score 1) 307

I'm wondering if they got results without a choke, or at full choke. This might be statistically significant.

No, they actually correct for the "non random" spread of pellets in the paper using importance sampling. Which is a neat trick and really the takeaway from this paper. I.e. that you can make Monte Carlo methods work, even if you don't have a flat/uniform probability distribution.

Comment: Re:No shit (Score 1) 103

by lars_stefan_axelsson (#46747657) Attached to: Why the IETF Isn't Working

TCP is not reasonable.

Sure there are problems, many and obvious. It's showing its age, no doubt about it.

However, there's a reason that it won, as the stuff that came out of industry (esp. telecom) was and is so much worse that it would make your eyes bleed. Why? Well, because any reasonable solution would make the telephone company less powerful and it would make it more difficult to charge for ever little operation. Hence circuit switched everywhere, all the smarts in the center of the network, and please make that as centralised as possible.

The mistakes made in TCP (in hindsight) were at least honest. We didn't know any better then. The "mistakes" made in telecoms were made for a completely different reason. They were made from a position of power that was going to make damn sure that it didn't lose that power and technology and society be damned.

Remember, that if it weren't for TCP/IP you could have had all the X.25 or ISDN you would have been willing to pay for. Which wouldn't have been a lot... It's no accident that the pricing structure came first, and the protocols as a consequence of that, in the telecom protocols, with pricing not even part of the TCP/IP suite.

Be very careful what you wish for... :-)

Comment: Re:No shit (Score 1) 103

by lars_stefan_axelsson (#46747537) Attached to: Why the IETF Isn't Working

Well sure. That's an "abuse" of the system. But I bet most of these are at least honestly not working. And being passive beats being actively malicious every time.

That said. I've met plenty of people in industry that spent their days just carrying around a binder. Hell, I've been one of those guys, at least partially and at least some of the time. So again, it's not a unique to academia.

To summarize. If you haven't been in industry. Don't for a second think the grass is even a shade greener. If it looks that way to you, it's only because marketing got to it before the brown and wilting became too obvious. :-)

Comment: Re:No shit (Score 1) 103

by lars_stefan_axelsson (#46740081) Attached to: Why the IETF Isn't Working

Well, I dunno. I think it's more a question of the size of the organisation in that case. As more people are interested, and larger entities (whether corporate or academic) things will move slower, and slower and slower. And the larger the more "useless" political types, middle managers etc. will be attracted. Like flies... Case in point, remember Usenet before the year "September never ended"? Same effect.

In any case. My main point was that academia is after all at least mostly honest. The corporate players are often openly or covertly malicious. Like Stallman put it (paraphrase), it's not a question of how much faster you can run than your competitors, rather, how much you can slow them down, by tripping them up or shooting at them. And if that makes you stand still, that doesn't matter.

Comment: Re:No shit (Score 5, Insightful) 103

by lars_stefan_axelsson (#46739857) Attached to: Why the IETF Isn't Working

You can hate on corporate types for various thing, but anyone who acts like academics know how to get anything done has never worked in academia. I work at a university and fuck me do we spend ages spinning our wheels, having meeting after endless meeting, discussing shit to death, and finally doing things 10 years after they needed to be done.

Well having done both big corporate telecoms standardisation and academia, I know which place I rather work in... (And I ultimately put my money where my mouth is. Or rather, didn't put my money as it were, salary not being an academic strong suit).

Sure, the local bike shedding can be tiresome, but our actual work, i.e. research, is cut throat and a model of efficiency and sanity. (Don't laugh. Cry if you have to, but don't laugh). There's very little politics in that side of the "business" and if you think there is, don't ever, for the love of all you hold holy, get involved in the corporate world. That's not just moving to the bad side of town, that's leaving civilisation altogether.

We used to hold the IETF, current warts and all, as the highest standard to follow (pun intended), but also saw where we were headed with the increased pressure, as TCP/IP became important to the political types and not just a nerd affair for sensible, reasonable people any more. You know, the kind of people that can listen to argument, grudgingly realise that another suggestion has technical merit and go along with that, instead of pushing their hidden agenda at all cost, and above all else.

When you've seen how the big boys make their sausage, you'd be as surprised as we were that your phone and mobile internet works at all. It's nothing short of an all out heroic struggle by the engineers in the trenches that makes it so. The rest of the system tries with all its might to prevent that from happening.

Comment: Re:Some of the oldest trades become useful. (Score 1) 727

by lars_stefan_axelsson (#46739787) Attached to: Ask Slashdot: Are You Apocalypse-Useful?

It is not my job to be a brood mare. And I don't want my role in society to be marginalized because my primary role is being a brood mare. (Also, even without modern technology, we can probably keep infant mortality down far below historical norms just with modern knowledge - even low tech sterile conditions do a lot.) Having control over when to have children broadly gives women control of their lives.

That of course depends on the actual scenario, but what you consider your job to be will probably be a luxury that society can ill afford. After the dust has settled, we'll be in a situation where we'll be fighting tooth and nail to avoid complete population collapse. We'll need scores of young people to work us out of that hole, and if you're of child bearing age, you'll bear those children. It's after all a unique skill that very few people can be put to do. (I'd for example be pretty crap at it, as I fall for the simplest and most obvious of reasons, I'm a guy.)

That's not to say that you'll be forced at gun (or club) point, but rather that society will bring down quite a bit of weight on you to "make the right decision". So for example, expect contraception to be made illegal rather than made unavailable, and anybody that interferes with a pregnancy to hang from the nearest tree. (Of course, it'd be the smart thing to do anyway, as we'll spend quite a bit of resources on the pregnant and mothers of young, so "Why would she want to do that anyway?")

After a generation or two (if we make it that far) that'll be the norm, standard, and "the way it always was", like Rob Slade puts it; "Don't go out dear, it's not good for the baby." will very rapidly morph into "Don't go out". And remember this is a good thing. It's what we want to happen, because the alternative is much, much worse.

Now, again, to say exactly how things will play out is difficult, because it depends on the scenario. If the scenario is nuclear Armageddon many bright people spent careers thinking about the consequences, and some of this information is now in the open (see e.g. Rob Slades short introduction: "Nuclear warfare 101-103" , esp. 103 deals with the aftermath. Another good book about the possible aftermath of an EMP strike is "One Second after." It is based on the US Govt. EMP commission report (and has a wikipedia page). It deals mainly with the immediate aftermath, so it doesn't really reach the "we need you to have kids"-part, but is still a somewhat realistic assessment of what society could be like. (It for example contains a long scene where refugees are triaged according to useful skill. If you're not a doctor or electrical power engineer, take a hike.)

But again. The takeaway is that what you consider your job to be will probably come so low down on the list of things to consider that it won't even make the first chapter. And that in a time when people will be far pressed to make it to the end of the first page.

But don't let that discourage you. My assigned role in such a scenario is realistically to die as quietly, and quickly as possible, so as to not use any resources best spent on the deserving. Preferably without putting any undue stress on them, mental or otherwise. I think a heroic but ultimately futile act to save the needing is my preferred way to go, but even that isn't my call. You get at least to have kids (with great loss of control of your life, but hey, nobody is going to be "in control" of much of anything), I don't get a life at all... (And if you're past the age where child bearing/rearing is a realistic occupation for you, I'll even let you lead me in the charge unto the breech. How's that for an offer you'd probably do best not to refuse?

So on a more upbeat note, lets agree that civilisation is a good thing and make our damnedest to try and preserve it, rather than go stocking up on soap just yet.

Comment: Re:Hulk hogan could code too (Score 1) 578

by lars_stefan_axelsson (#46732683) Attached to: Michael Bloomberg: You Can't Teach a Coal Miner To Code

Try as I might, that kid just would not believe that he could do meaningful stuff. It's not that he didn't want to be code monkey like me, or didn't want to have a higher paying job. He just didn't think it was going to happen. I couldn't get him to try. That sort of resistance is weird, and I'm not sure I have a solid grasp of it's root cause. But if I had to call it something, I'd say it's the culture of the poor.

The usual suspect is learned helplessness. Or "you can run but you'll just die tired." If you've died tired time and time again, haven't seen anyone that hasn't, and everybody tells you by word and deed, to just die already, because what's the point of dying tired, what are you supposed to think? Or do for that matter?

Comment: Re:Knowledge (Score 1) 1037

by lars_stefan_axelsson (#46692667) Attached to: How the Internet Is Taking Away America's Religion

Um, absolutely not? Correlation has nothing to say about causation at all. It is not evidence of any causation at all. It can be (and often is) sheer chance.

Nope, that's too strong. If that was true then we wouldn't even use correlation as a tool at all. And we do. Because correlation is evidence of causation. It's not strong evidence, and it's not absolute evidence, but evidence all the same. If there's strong correlation, that's clear cause for further investigation, for where there is causation there is invariably correlation.

So whole correlation doesn't show causation, it correlates with it...

Comment: Re:Does everything need to be smart? (Score 1) 128

Sprinklers are something you really don't want to fail, because both scenarios are destructive. If the sprinklers fail to work as designed, your house burns down. If they go off without a fire, you have lots of water damage, which is almost as expensive to fix as fire/smoke damage.

And while that's true. Few people (if any?) have died from their sprinklers coming on. Many people have died in a fire...

So the cost of "fire" vs. "sprinkler" isn't symmetric.

Comment: Re:Because Hollywood. (Score 1) 544

by lars_stefan_axelsson (#46660015) Attached to: 60 Minutes Dubbed Engines Noise Over Tesla Model S

Having tires squeal on gravel is similar. Rather than background noise, the distracting element is that the sound just isn't what's expected. In an action shot, there usually isn't time to properly establish the scenery.

Which is interesting from a European background. We've definitely noticed for a long time, and our cars don't squeal nearly as much in the movies as US ones do. It was something I grew up with, learning that american cars apparently squealed their tires on ever turn. I even remember news pieces talking about the difference in tire technology between the US the Europe and how american cars actually squealed more IRL than ours... Wasn't so, it turned out.

So, given that we've gotten rid of bullets going an annoying "pew pew" in was movies, is there any chance of US cars not squealing their way around every single corner?

Comment: Re:Good, I guess (Score 2) 148

So CDMA scales beautifully with number of phones, while GSM does not scale at all. Consequently the CDMA carriers were the first to roll out 2g service. There was no way to fix GSM for data. They had to add on a different standard for data, which most carriers implemented with CDMA or wideband CDMA. That's right, the HSDPA data service on most 3g GSM phones was actually CDMA. That's why you could browse the web and talk on a GSM phone at the same time - it had one TDMA radio for voice, and a second CDMA radio for data. CDMA phones couldn't do that (unless they supported voice over IP) because they only had one CDMA radio for both.

As someone who developed GPRS for Ericsson back in the day, I don't even know where to start...

There were a number of different competing standards, in different parts of the world. That CDMA wasn't mandated in the US was not for lack of trying by the US manufacturers.

And, no, if we're talking about true packet data, i.e. not "phone modems", GSM/GPRS did emphatically not use a dedicated slot per user for data communication. Instead all the available "data" slots (and there can be many) were/are shared dynamically between all the users wanting to receive/transmit using dynamic reservation protocols (depending on, among other things, whether you have data to send/receive). Indeed EDGE is just GSM/GPRS with more data slots available, and with mobiles that can use more slots in sequence.

All this is moot anyway, as the explosion of demand for mobile IP, necessitated completely new systems anyway. And since they were new, they weren't hampered by what was already there. You say that UMTS is based on CDMA, which is true, but there are also FDMA and TDMA parts, and even versions of the UMTS protocols. So that UMTS is CDMA and that's superior to GSM which is TDMA does not follow.

I could write a book about the rest, but that'll have to do for now.

Comment: Re:[sarc]How wonderfully counter-productive![/sarc (Score 1) 207

Torture only works for confessions of things you already knew for sure.

Well, for completeness sake, there are specialised situations in which it can yield highly valuable results, and criminals for example, know that. I'm thinking of situations like "Tell me the combination to your safe, or else..." and the like. Time locks on bank vaults were for example invented to stop the all too popular kidnapping the bank manager and holding his family hostage, "or else".

But of course you're right, that as a means of intelligence gathering these situations are so uncommon, as to render the method completely useless.

There are worse things in life than death. Have you ever spent an evening with an insurance salesman? -- Woody Allen