istvan__

As well as network IO. This is why just by the load you can't tell what is going on and this is exactly the reason why I don't like it too much in production. A box should do a have a smooth 15minutes normalized load over time. (If your workload changes you can do autoscaling, I think we used the normalized load as the metric for scaling up and down).

Actually since you have no idea about what is causing the load (it can be wait on network IO) this is why I think that running your production system in that shape is not recommended. Out of curiosity in what situation is it ok to have significantly more things waiting to be running than your actual capacity? Seems like a bad capacity planning to me. Anyways, this is how it was done (keeping the normalized load around 1) in my previous gig where we had ~5000 nodes and it was working fine. I work on Hadoop clusters nowadays and any time we run into a load of 100+ there is a severe degradation in the service, timeouts etc happen. In reality high normalized load over time (not talking about 1 minute spikes) should be avoided, this is based on my experience.

I still remember the first time this came up and people were betting that Amazon is trolling everybody. In the same time, a far far away country:

http://www.bbc.com/news/technology-31129804

It is good to see that drones can be used intelligently for everybody's benefit. Hello faster delivery, good buy lost (stolen) items.

Also: https://gist.github.com/yminsky/c5b8088ce355a658174e

I would just add a rule into the loadbalancer config, if the request is coming from the music industry or a government agency redirect them to an empty website. (I am kidding).

I don't think that a load of 172 is a good idea. I know this is a benchmark that is measuring how fast you can go ideally but in production the question is how fast you can go with with keeping the latency within the SLA. As a general rule you want to run your boxes around 1 normalized load ( load / # of CPU cores). The rest of the article is pretty nice.

I absolutely agree with you, my problem is that if I bring these things up in an discussion about languages it results in "who are you kidding, we don't need that feature!" sort of comments. Personally I find it hard to do any sort of computation without the ability is inspect what is going inside.

Well I guess the Go community is a little bit too passionate about any update regarding Go, even if the introduced new thing is subpar compare to other languages, like this one in the post.

Damn machine learning software. :) Google needs to tune some of the parameters.

It is an ok article. I checked the license of LLVM though:

http://llvm.org/releases/2.8/LICENSE.TXT

http://llvm.org/docs/FAQ.html#does-the-university-of-illinoi...

If there is something very specific to Apple and they want to keep the copyright after writing that module in LLVM, we should respect their decision. There are several companies using copyright licenses or not open sourcing anything they are working on. Compare to them this entire LLVM situation is a success story from the free software point of view. LLVM became the default toolchain and compiler for many project because of its high quality and meaningful error messages (GCC anyone?). I can live without Swift, but I definitely need LLVM going forward.

I was just generally asking. But thanks.

You can check that with torrent. If it is not a single server and a single connection you should be able to saturate your network connection.

I think you just need to talk to a professional who can help you to get motivated and interested and the rest comes after that.

http://www.apa.org/helpcenter/therapy.aspx

I am pretty sure there is something you like to do. Get good at a tiny small particular thing first before you extend your interest to other things. If it helps, I used to be a test engineer who was tasked to run test cases manually. You know what I did? In me spare time I wrote an extremely dumb and simple executor that did my job. Management realized the potential, they let me write a better tool now with their approval. This was 10 years ago. Today I still struggle with Java and sometimes with Python but I found a handful of languages that I really like and I am focusing on getting better with those, kind of ignoring what is mainstream. I regularly fail job interviews (the last one few days ago, using codility) but I keep interviewing so I can do those better over time as well. I do one interview per week as a minimum. I try to contribute things to open source projects. Even stupid simple things like this (https://github.com/cloudera/hue/pull/68) it is not even code, it is configuration, but we needed that few jobs back.

I don't think I am particularly talented, but I spent so much effort on this (learning how to write software) that I can say today that I am a mediocre software engineer. Again, just think about it. I spent 10 years on this. In the meantime I did everything that I could, being Linux admin, systems engineer, data engineer, just to keep myself running. Almost forgot, I never received any formal education in computer science.

My co-workers run circles around me when it comes to programming, yet my persistence helps me to outperform them.

I guess what I am trying to say is: - don't give up - expect success over time (1-2 years as a bare minimum) - try to start with a simple thing, as simple as it gets (QA, DevOps, anything where you don't need to flip binary trees on a whiteboard :) )

This was my story, I hope it helps.

It seems they were using some of the tools:

https://2011.eurobsdcon.org/papers/gras/minix-bsd.pdf

"ABSTRACT MINIX 3 has imported a significant amount of userland BSD code. The trend began several years ago, but the pace has quickened markedly. We have already imported NetBSD’s buildsystem, NetBSD’s C library, the pkgsrc package management infrastructure, and various userland utilities from NetBSD and FreeBSD."

Weren't they using FreeBSD?

Yes, and there are some cases when it works. I would not run our production system with Ksplice. The microkernel architecture makes it possible to restart kernel functionality seamlessly. Emphasis on restart. Ksplice just binary patches the running kernel and overwriting the changed parts. Very different scenario.

An absolute must! I wish there was something for real.

I understand that it is much easier to sell something with unlimited (even if it is not) or free (even if it is not).

Absolutely, I still don't like it. :)

This is not about the speed, more like about the volume of data. They limit you after you reached a limit that should have never reach because it was advertised limitless.

We should stop lawyers from re-defining words like unlimited. We should make sure that if somebody says unlimited in advertisment or product description it really means unlimited. I know, I am an idealist. :)

This depends on the stack. If the cost grows faster than your revenue than it is a problem. If they have a lean stack where the cost grows slower than the revenue by the number of users they are good.

I guess than the next phase is to build a mesh network of wifi routers and create a free closed internet without access for government or whoever wants to ban free speech. The technology is ready, with solar power improvements we can do it.

Exactly. What is worse than non banning hateful mobs is banning only the ones the government disapproves...

Wait a second, you just want to ban all the feminists?? If you would like to see hateful things just pull up any of their actions.

Example:

http://www.independent.co.uk/news/uk/home-news/bahar-mustafa...

She is racist and sexist at the same time, a whole mob is, I don't see anybody taking these things seriously and she kept her job.

Similar situation different players:

http://www.dailykos.com/story/2015/06/10/1392082/-Woman-Invo...

Lady loses her job.

I don't see that limiting free speech if going to be good if it is implemented with the same hypocrisy and biased way as our society is (including media and people) now.

I strongly believe in free speech. The only thing that we might need to filter is direct hate speech that is calling for active violence. Anything else should be expressed freely. We need a distributed network for news that avoids these sort of laws to host content partially (few nodes together have the complete data) and changing over time what is hosted. TorrentNews or something like that.

RIP free speech. This is basically saying that highways are responsible for the damage bad drivers cause.

This was the best tech talk I have seen for a long time. Thanks for sharing it.

That is great, evolution always find its ways to get rid off the weaker population. Wearing a helmet (a good one) can save your life.

Some details:

"Is there evidence that helmets protect brains and heads? We have a statistics page that may help you research that. Don't miss the recent New York City data showing that 97% of their dead cyclists had no helmet."

I know what a data scientist would say, correlation is not causation.

More details, lots os studies and numbers here: http://www.bhsi.org/shouldi.htm

A good video in the subject: https://www.youtube.com/watch?&v=CLbpXkSoM2A