obviously, the one where you skip static analysis and devote 100% of testing time to testing. it will be exactly 10% more reliable than both cases of your false dichotomy. :-P

besides, your perspective is more about productivity than about quality. that's ok, but it depends on your particular environment. actually static analysis gives you a specific insight into a code body which is good. but it's often greatly overrated and i find statements like "static analysis adds reliability" alarming. that's simply not true. it assists you in assessing reliability but if you stop there you're screwed, the actual important step is testing and if you get *that* right, you're ok. reviews are a great plus, not just for reliability, but also for learning.

and i like static typed languages, if anything just because refactor is a breeze with adecuate tools.

actually, it means you can employ unreliable programmers and think you can get away with it because your reports are green all over. odds are you won't.

static analysis tells if code is wrong. it will never tell if code is right. it's just a tool to assist in early improvement and by no means a substitute for reviews and testing. review and testing do add reliability. static analysis doesn't. it's just convenient, but you could do perfectly without, provided you test and review.

please, folks. do get some real insight into the languages you use before doing serious stuff others will have to maintain. thank you.

this is true in general, but regarding to javascript:
if you come from Java, you *must* unlearn almost everything to start doing anything half decent in javascript.
If you come from c++ not so much, but please, please ... read the manual first. javascript is *not* what you think it is.

the problem with javascript isn't javascript, it's that 90% of people writing javascript hasn't a clue about javascript.

as for dart, i don't know / don't care atm.

I don't know what sort of projects you've worked on, but an enterprise level project (specially a "large and complex one" as is discussed here) has a host of business analysts, functional analysts, architects, IT and consulting guys, field expers, accessibility experts, testers and developers skilled in several technologies, all committed to the same project. very few (if any) of all these have a thorough understanding of the whole system in full detail, not in 18 months and probably not in the entire life of the project. but for anyone to be productive in such an endeavour means breaking it down into manageable pieces and ... good management to make all these work together. if folks cannot make useful contributions whithin a month then they are useless or very poorly managed, or the system is not just large and complex but also badly architected or lacking a sound infrastructure.

dunno what to say. maybe if you give a concrete example we might find out why you guys are so utterly failing. could be guru corporativism, bad practices, bad organization or documentation, could be anything but definitely it's like there's a big elephant in your room.

now if 18 months seems weird, this seems complete nonsense. what has the concrete technology to do with complexity? if your front-ends are dumb and your backend is a nightmare, then you'll have to hire specially talented backenders. they might even point out what you are doing wrong. but if they can't contribute *something* of value in at least 18 months then the problem is on your side.

a month *should* be enough. more than a month means an awful programmer or average lousy management. large projects tend to have many managers with many of them being ladder climbers, this people can produce spectacular chaos when thrown together. so indeed longer catch-up periods are common but GP is still right about what it usually means: bad management.

which standard? html5 isn't a standard at all, let alone well known. it's a wannabee. besides, famo.us doesn't seem to seek to *replace* anything, it's just another option for an area currently lacking good enough support or standard.

if it works crossbrowser i can't possibly see a problem with it, just as with any other lib you are free to use or not.

won't work. many of dmca trolls have no problem in shelling out 10k, even several times a day. this would just alienate the guy or small company that may have a legitimate claim but can't afford it and also has reasonable fear that he will get trampled over in court. it's the same problem with patents: the rich will almost always win. patent fights nowadays only make sense between megacorps.

the only real solution is going full underground. free source simply cannot depend on private companies' goodwill, with or without dmca or you name it. it has to escape the system to be really free. that's the problem i always saw with github and similar sites: a cool idea, cool implementation, but moot in this society full off greedy incorporated assholes. we need to implement the idea differently, with a resilient anonymous network (that's right, it *has* to be anonymous, too bad for some crippled egos). if we manage to get to that no ignominy disguised in lame copyright will do shit.

... that less than 30% of Americans are mature enough to use the Internet.

isn't that theyism?

yes, it can. it involves a partial re-encryption, verification codes, zero knowledge proofs, a mixnet and a bunch other funny stuff. it can be verified individually by the voter, and the counting process is also auditable, verifiable, and anonymous. you'll find complete specifications on the norwegian election site.

that's another issue. i might even agree, but norwegian (and most countries') law strongly enforces privacy. privacy is precisely the hardest nut to crack.

great power comes with great responsibility. i guess we'll get there when we're ready.

the system is not opaque at all, it's fully transparent. unfortunately, it's also complex. it's verifiable by anyone with understanding of the math and the data manipulation involved. that's far from a majority but still quite a few, and it would take only one of them pointing out a flaw to invalidate the process.

a large majority of the population can't verify the whole process on paper either, there is also trust necessary in that due process is followed and collectively monitored correctly. this doesn't rule out fraud completely, but makes massive fraud very difficult.

i can't. but the voter can invalidate that vote many times, from anywhere, secretly, and ultimately he can vote on paper too if need be.

sorry, i'm really far fom angry, just baffled. who told you that? this was the first time (not really, we're actually talking about the second run of the same experiment, the first was in 2011, but on a smaller scale) that full evoting over the wild wild web was attempted on binding elections ever, with privacy, anonymity and verifiability. the real test was if this was at all possible and practical with present technology. i turns out it is. it's complex, it has issues, there's a lot of room for improvement but it works, it can be carried out and verified. participation was just another measure amongst many others ... i dunno, it may have been trumpeted in some official statement but your common sense alone should have sufficed for you to realize that even if that were the primary objective, results of a single pilot test between different types of elections in different contexts and different population sizes and with years of difference would have been far from conclusive if not merely anecdotical in all but very extreme scenarios.

the aim of this experiment was to see if this difficult and heavily debated stunt is at all feasible, and how it works out. it was possible because some guys just thought it was worth to try, and because some other guys thought they could handle it. many conclusions in lots of areas can be drawn from it, and more experiments will be necessary. just looking at participation comparing to some other random election is not only simplistic but makes no sense.

the difference now (that the system has proved feasible) is that current guys just don't find it interesting, necessary or worth the effort. that's just ok, but the talk about turnout is definitely just an excuse.

oh please, i don't merit this. i wasn't at all angry, just calling out bullshit. nothing personal.

you are wrong or lying. there was a "virtual" voting machine implementation. it was not used, like a bunch of other funny stuff, but it always was an option, that's the reason it was fucking implemented in the first place. you can look it up in the sourcecode since it is public. who the fuck are you anyway spilling all this bullshit?