Unfortunately we live in an age of idiocy where anyone will sell their soul for a smaller fee and less privacy. There's no winning against that so you have to deal with this from the inside of the enemy via "legitimate sabotage"

a) get into hiring positions in the government and hire the stupidest, least qualified and incompetent people you can get then quit and leave the projects in the shit.

b) spend money galore so it cant be used on anything else

c) leak insider information on the sly.

d) use the tools you built against the people who paid for them.

e) manipulate and shame other staff out of positions of power.

f) play people off against each other and create new rivalry which consumes all money and time.

I worked for the defence industry. This is how to break it.

You never know with git. Which is one of my big problems with it. The picture on master doesn't necessarily represent what happens on other repos across your organisation.

I don't understand the problem. Which VCS does tell you what your developers are doing on their machines before they commit to your main repository?

None. However svn prevents them doing anything major unless they tell someone.

You mean like Gerrit? Or do you mean something else?

It's bad, irresponsible science. And it's bad, irresponsible journalism to make it a giant orange pull quote.

"I think it's X. But I don't have any evidence yet."

If you don't have any evidence, you shouldn't be talking about it to a journalist. Who, quite predictably, is going to make a big deal out of it.

This sort of thing actively hurts the climate change cause. It's bad, and shouldn't be done.

By that standard, Einstein should have not talked to journalists about his 1905 relativity paper until the first prediction that wasn't null was experimentally verified in 1938.

That's simply not a comparable situation.

Somewhat amusingly, it was actually a remarkably comparable situation.

There was a significant populist anti-relativity movement in Germany in the 1920s: http://www.mpiwg-berlin.mpg.de/en/news/features/feature7

>Although they had previously played no role in German academic life, during the 1920s scores of self-proclaimed researchers alleged to have proved the theory of relativity to be scientifically incorrect. Because the arguments set out in hundreds of ensuing publications frequently rested on fundamental misunderstandings of Einstein’s new theory, their accounts have largely been ignored by traditional history of science.

>Einstein’s opponents were simply not prepared to question their own worldviews and instead sought alternative explanations for why their objections were disregarded by the academics. With time, many turned to conspiracy to account for their marginal status: plots favoring Einstein, so they imagined, explained his success and their marginalization. Having reached this point, any sort of resolution of the controversy had become impossible.

And from a review of the book ( http://arxiv.org/ftp/arxiv/papers/1111/1111.2181.pdf )

>Nevertheless, anti-relativists were convinced that their opinions were being suppressed. Indeed, many believed that conspiracies were at work that thwarted the promotion of their ideas. The fact that for them relativity was obviously wrong, yet still so very successful, strengthened the contention that a plot was at play—and some anti-relativists were convinced that the co-conspirators were Jewish. Jews were held to dominate both the newspaper business and the new discipline of theoretical physics; they could thus easily advertize one of their own (Einstein) and his fallacious work (relativity). Gehrcke, for instance, kept emphasizing that the successes of relativity could only be explained by a state of “mass hypnosis”, brought about by excessive and one-sided reporting.

Hmm…

That's interesting, but it certainly was not one of the largest political debates of its era.

That the situation climate change is in. So it's not a comparable situation.

"This world is a strange madhouse, every coachman and every waiter is debating whether relativity theory is correct. Belief in this matter depends on political affiliation." - Albert Einstein.

I guess there was a too high expectation of the audience, and it would better to just stick to what there is evidence of already.

Yes, actually, particularly when talking to a journalist about possible climate change effects that's precisely what scientists should do.

If you need proof this kind of thing hurts the climate change cause, look at this thread, which is nearly entirely a debate over her one statement and nothing about any actual science.

Nothing hurts the climate change 'cause' more than it being a 'cause' in the first place. Activism dressed up as science smells, and plenty of people who would otherwise be sympathetic can smell it.

This is exactly the problem. There should be no activism and no causes in science. Each of them implies bias and results in less than ethical practice and unreliable theories.

To ask for rationality and the application of the scientific method results in claims of denialism and instant burying of the opinion.

That's not acceptable.

Scientists are not magical logic faeries separated from the seething mass of culture by a wall of pure mathematics. Science is intertwined in life.

Once you have found that a species is being wiped out, you could stand on the sidelines and measure its decline, watching dispassionately, knowing that you can reach a good solid conclusion about what was killing them when the last dissection is fully documented, or you could chose to interfere and try and prevent the extinction.

However, if what is killing them is well funded human activity, you do not have a hope in hell of interfering successfully without getting political.

Personal attacks are not allowed on Hacker News.

The "problem" is that even with the science pointing in a particular direction, if monied interests don't like that direction, they turn it into a political issue with FUD (personal attacks on the scientists rather than the science, stirring up the "science is an affront to God" crowd, etc).

I disagree completely.

Thought experiment:

An astronomer sees a comet approaching earth, that will kill billions.

He wants to prevent those deaths.

He starts campaigning for methods to prevent that: comet deflection, shelter building, whatever.

Seems logical and humane to me.

There's a pretty big assumption in there. That the comet is approaching earth and not going to miss it by half a parsec or that the comet is infact a bit of dirt on the lens of the telescope...

All of which need testing first before you start waving your hands around and start saying "WE'RE GOING TO FUCKING DIE - RUN!"

We operate in an uncertain world. In many cases it is hard to obtain 99.999% certainties, and we must act with less certainty.

Not acting is a choice as well, and everyone bears the responsibility for that too.

Problem is we have about 50% certainty...

So, if we continue the thought experiment, and say thousands of people have been making observations of the comet for five years and have come to roughly 50% impact probability estimate, with multiple independent methods. And that impact would likely kill maybe a billion people.

And you say it is unacceptable for them to advocate doing anything?

http://neo.jpl.nasa.gov/risks/a410777.html

Current impact risks overview is here. 50% chance of a large object would certainly be unprecedented.

http://arxiv.org/pdf/0804.1126.pdf

Now, replace the comet watchers with paleoclimate researchers finding out what happened when the planet last time crossed 450 PPM (the other way around).

That's not what I'm saying. I can't make it much clearer than:

a) suddenly out of the blue, a Ford Transit mutates into a toaster.

b) someone is "pretty sure it is related to the impending comet impact".

c) Everyone goes YAY SCIENCE!!!

Sorry but derision is the only thing left. It's not science, it's religion.

a) suddenly out of the blue someone delivers a load of fish.

b) Some guy doesn't notice this and goes "pretty fucking sweet; I'm totally wasted on mushrooms and I reckon that bearded dude over there turned that loaf of bread into all those fish. Damn I've got the munchies."

c) HE'S THE FUCKING MESSIAH.

There is no causality chain established other than a hunch.

Ok, I thought you categorically opposed scientists being activists.

I do agree that some things are perhaps too easily reported or assumed to be caused by climate change when there is not much evidence to support that (yet).

Inaction is always easier than action, so which way do you think that people will choose?

If we do spend the time, money, and effort to do things like reduce emissions, will we be worse off than we were? Would it not make more sense to use renewable energy resources, and save up the 'easy' sources of energy (oil, coal, etc) for other applications?

Entirely agree!

But blaming everything on climate suddenly isn't the right answer. Perhaps someone lost a ship load of PCP in the sea and they just snuffed it. And they're trying to hide that.

Another hypothesis that could be tested rather than "fuck blame it on climate so we can get some funding and get some nice new Herman Miller chairs and a new MacBook each"

That's engineering, not science.

This cuts both ways. Demonizing one side ends all rational discussion. Where are the Tip O'Neils of today? We need them desperately.

I disagree. The tools are awful and terribly slow to execute at best. Then you occasionally get shafted with a relabel.

Can you be more specific? There's lots of different tools available - there might be some better options available than what you're using. And while tools that change things (e.g. setting booleans, changing context rules, etc) are unlikely to be instant in the near future, as the policy needs to be re-compiled, this has been improved, particularly for booleans. You can also batch updates together, which is a much nicer experience if you're trying to set a bunch of things at once. e.g. semanage -i <( echo -e "boolean -m --on httpd_use_nfs\n boolean -m --on httpd_use_sasl") That said, "this admin command I rarely use takes 30s to run" (30s seems to be about the average on RHEL 6/7) is to me an odd reason to try to avoid an important security feature.

I don't recall the last time I had to do a full relabel on a production system. Not saying it hasn't happened, but I can't recall an instance.

Overall, I spend less time managing SELinux (and that includes the custom policies I maintain) than I do managing IPTables. It's really not the nightmare it's made out to be.

Don't bother running a mail server. It's hell. I did from 1998 until about last week. I started running it on a cable modem with an old Compaq desktop, migrated to a dedicated server (which hosted a load of other stuff) and finally onto VMs.

I have spent hours getting myself of blacklists after entire IP ranges were reported. I've spent hours working out why the hell Postfix won't talk to dovecot on a local socket, upgrading Linux distributions, postfix and dovecot and being fucked over by config format changes and periodically losing entire mailboxes to IMAP bugs. Oh and SELinux - hours of it.

The thing that finally killed it for me was Yahoo. My landlord uses Yahoo mail for comms and Yahoo just decided to stop accepting delivery from my server. It took a phone call from me asking him why he hadn't sorted something for me and for him to tell me that he didn't get the email to discover this. So I dredged through the logs, found an error from yahoo's mail servers saying I need to hit a web form to prove I was a legitimate sender. So I did that, and nothing, not a sausage happened. Googling around says I need to wait 6 months before submitting the request again.

I have no power in this situation. I can't email my landlord. I need to do business, not piss around with politics.

So in a fit of anger, I blasted my VM on Linode, went to my local supermarket, bought a Moto G2 in cash to replace my Lumia 630 (which was doing IMAP) that I just smashed the screen on, signed up for Google Apps free trial and just moved the domain over to that.

Just can't be fucked with it any more. I run "inbox zero" (i.e. I delete religiously) so there is no cost for me to migrate.

I don't care if they read my email or shop me to the feds; I'm tired of herding servers and software these days. A decade ago I could quite happily spend all day doing it but I have better things to do now.

Funny enough my own server ONLY had problems with .. GMail so far.

From two days ago: "Let's try this GPG setup, write a mail to my coworker's private GMail account and get a verification that everything checks out on his side".

GMail refuses the PGP/MIME mail, bounces after end of DATA, as 'Spam'. What?

Send the same mail to my GMail account. Works.

Send the same mail to my brother's GMail account. Works.

Send to coworker again: Bounces after DATA, refers to a useless website that offers Google's policies for bulk mails again.

Send a plain text mail to coworker, complaining about Google's mess: Accepted.

So.. Takeaway:

- Google randomly rejects mails from my server, without anything I can do and without providing any information WHY it would do that. Not filing as spam, rejecting outright.

- Google only did that with a mail that is unreadable (PGP/MIME). Coincidence? Make of that what you will..

I will keep my server though. And loudly complain to people that run broken setups. In this case I complained to this coworker of mine and wrote lots of expletives directed at Google - and the issue is resolved for me now. People with GMail addresses that don't get my mail are frankly not my problem and for friends and family I might be able to exert enough pressure to fix the problem.

I moved to runbox a few months ago and I'm pretty happy. I'll echo your "don't even bother running your own mail server". Been there. Done that. No thanks. Horrible.

YMMV. I've also ran my own since 1998. It's worked almost flawlessly from the start. A stock Debian system that I've kept up to date just keeps on ticking.

The only trouble I've had has been sporadically with Hotmail. Still don't know why, I get the bounce and contact the recipient from somewhere else. Even Gmail has these problems from time to time.

I just wanted to stick in another datapoint in the discussion, so not every post is about how awful email is. Sometimes it just works.

right - for individuals who are trying to run a server on a shared server/network (VM), you're going to have a bad time... a real bad time.

however, it's not this bad if you have a 'real' infrastructure i.e. a colocated environment with your own IP space. it sounds like you were mainly dealing with the fallout from bad behavior from your neighbors on the shared network.

having said all that, it still might be worth a switch if you use google apps a lot (we do) because of the integration with gmail. we switched and now use IMAP to access our company google apps gmail. i really dislike the web interface.

Actually I had a nice dedicated box until 2013 at a reputable host in the UK. Unfortunately I no longer wished to fish out for £95/month for it. Still had problems even though the thing was well maintained and there were no bad neighbours. It was less noisy than a VM at Linode however, so you're right there.

yeah, i meant if you actually have ARIN IP space or get allocated at least a routed /24 from a tier 1 ISP... not a cheap dedicated box in some datacenter that suffers from the same problems as shared VM networks.

I have a compromise: run my own server but pass outbound messages through mandrill. Most of my mail is inbound anyway.

It took a few hours to configure exim + dovecot + sieve plugin, but it has been running unattended for over a year (upgrades are automatically installed using unattended-upgrades[1]).

Spam is not a problem, I just use a different address for each service, and ban it if it starts getting spam. As an additional benefit, I know who failed to keep it private.

[1] https://wiki.debian.org/UnattendedUpgrades

There is a hybrid approach -- run a local mail server for receiving mail, and configure your Gmail account set the "From:" address to your domain, and forward any mail to Gmail to your domain. This way you are still mostly in control of your mail, but with the sending reliability of a major service.

This basically negates half the privacy though. And given most of the mail you recitative will be sent from gmails servers anyway...

For myself, I'm not too worried about privacy (if I want that I'll use PGP). I'm more concerned about vendor lock in. Because as long as people are sending you mail to a domain you own, you can change the sending email provider to any other one that allows a custom From: address. Now from a practical side, I currently let Google Apps completely handle my domain's email, but I can yank it off there at a moments notice if I had to (i.e., if Google decided that I'm not allowed to host my domain there anymore).

Interesting approach and one I hadn't considered. Thanks for the heads up.

Spot on. I can't say enough good things about COM; it's enabled some quite amazing things for me over the years and really top notch distributed systems that are easy to manage.

I wish the FOSS world hadn't abandoned CORBA all those years ago.

I'm not sure why you were downvoted - it's a valid solution.

To be honest, I'd rather see DCE/RPC (DCOM's ancestor) make an appearance without some of the pain points of DCOM.

I wonder what Andrews and Arnold's approach to this problem will be. They're the last bastion of common sense in the UK.

See http://aastatus.net/1984

"Just to be clear on our policy here - when DRIPA comes in to force, and if A&A become subject to a retention notice for all customers, we aim to work on all practical legal means to minimise the amount of data retained under that legislation - making full use of the bad wording in the Schedule in the 2009 regulations where possible. We also aim to clearly publish what is retained under such a notice and what steps we have taken to minimise such data. Such steps may mean separate companies running email or other services, or even hosting some servers outside the UK, if those are practical steps we can take.

Why? Because blanket mass surveillance is illegal under EU law as it is against our basic human right to privacy as decided by a court, that's why!"

Fantastic - thanks for the link.

This. Plus its fucking miserable dying slowly over the space of 20 years. I've watched it happen to so many people. Avoid work, retire early, spend everything, eat everything, do what you must do as early as you can before probability becomes your worst enemy.

Edit: The people I've watched die had one regret only: Waiting until retirement to do what they wanted at which point they can't. Perfect cogs with pensions and private healthcare paid up through slavery for 40+ years. Nothing to show other than some bricks and a hip replacement and a nice plot in the graveyard.

The classic example is travel. Some people wait until their retirement before they travel the world, which means they can no longer do the physically demanding adventures and they have only a few years to enjoy the memories. Others travel before they start a career and family, allowing them to appreciate the memories for their whole life. Others still just find ways to travel their whole lives, and make it a part of who they are and what brings them fulfillment. It doesn't have to be travel, but I believe most people can find ways to be fulfilled their whole life instead of postponing it until retirement. I see plenty of examples in my surroundings of people who have psychologically locked themselves up in a non-fulfilling life. I can see the ways they could break out of it, but somehow they can't.

Just want to mention that some of the internals of this aren't great. There's absolutely no decoupling which makes testing very difficult and the contextual stuff is horrid. The components folder is also a cleverly hidden static ball of mud. I hope this isn't what they're teaching new developers as we'll have to spend a long time getting them to unlearn this in the future...

Could you give me an example of something that would be hard to test, and how it may look in another popular framework which is less decoupled? I want to understand this.

MVC is really easy to test. This example just completely ignores concepts as programming to interfaces, composition and single responsibility.

That's not much. Our baseline config HP DL380s come with 64GiB RAM (!!)

File servers should really have Looooots of ram. after all its cheap cache (compared to a fusion io card) Our file servers have 384 gigs of ram. The next gen will probably have much more.