rumcajz

46 ms·

421.
by rumcajz 11y ago
What I meant is that looking at an average programmer's resume feels relly weird: A web shop. High-performance trading. Game development. Scientific computing. Telecom. And so on. No way anyone can be a domain expert it that many areas
422.
by rumcajz 11y ago
I wrote about the problem here: http://250bpm.com/blog:36 The fact that abstractions can often add complexity rather than remove it should be taught in schools. The problem with distinguishing between the two cases is that
423.
The Wrong Abstraction(sandimetz.com)
220 points by rumcajz 11y ago|119 comments
424.
by rumcajz 11y ago
See Stanislaw Lem: His Master's Voice https://en.wikipedia.org/wiki/His_Master%27s_Voice_(novel)
425.
Swiss to vote on guaranteed income for all(thelocal.ch)
2 points by rumcajz 11y ago|1 comments
426.
Getting rid of state machines (II)(250bpm.com)
1 points by rumcajz 11y ago|0 comments
427.
by rumcajz 11y ago
Good point. The advantage that I am looking to get from unikernels, one that no one seems to mention, is that the whole area of functionality that's above hypervisor but below posix/win32 is suddenly open to play with for applicat
428.
by rumcajz 11y ago
Same story here. Went from C++ to C for similar reasons.
429.
by rumcajz 11y ago
Inequality is bad for economy. A person that owns 100,000x of average doesn't spend 100,000x more. They have to invest the money. But there's nothing to invest in given that poor people don't have much money to buy stuff and
430.
by rumcajz 11y ago
Look at how the image is blurred at the edges. It's a virtual reality and not a very good one, at that.
431.
by rumcajz 11y ago
libmill provides fibers/coroutines as well. check "go" keyword.
432.
by rumcajz 11y ago
A problem that is easy to solve: Clearly define the scope of the project and when it's done, it's done. Period. Unfortunately (maybe with the exception of Knuth and DJB) nobody does it that way.
433.
by rumcajz 11y ago
On the other hand, programmers jumping from one industry to another build no domain knowledge. The result is that domain experts, such as accountants or telecom technicians, specify the requirements, which then get translated 1:1 to softwar
434.
The Big Mac index(discover.economist.com)
1 points by rumcajz 11y ago|0 comments
435.
by rumcajz 11y ago
Duff's device can be used as a test of familiarity with good old unstructured way of procedural programming. If you find it weird, you don't get it. More practice needed.
436.
by rumcajz 11y ago
Wealth inequality and income inequality are often touted as the incentive for people to do actual useful work. However, if you think about it for a second, it's clear that having more money in a bank or on a payslip doesn't make y
437.
by rumcajz 11y ago
> This is like saying the more developers you put on a project, the more entrenched you become...so developers are not a panacea. See Mythical Man-Month ( https://en.wikipedia.org/wiki/The_Mythical_Man-Month )
438.
by rumcajz 11y ago
An interesting side issue here is that bad conceptual model gets the more entrenched the more tests you add. Fixing the flaw becomes almost impossible because it breaks all the tests. Testing is often presented as a panacea but this is one
439.
by rumcajz 11y ago
> All I'm saying is that more freedom causes economic growth What about Russia in '90s?
440.
by rumcajz 11y ago
The original term is "sub-routine". See Wheeler, "The use of sub-routines in programmes" (cca 1952)
441.
by rumcajz 11y ago
> REST vs SOAP, the new Facebook API, whether Ruby on Rails is scalable, what do you think of Sun's new Java-based scripting language, Google's widgets API, Amazon S3, etc. Really? Why would anyone learn any of that unless they
442.
by rumcajz 11y ago
TCP roundtrip times over loopback in the presence of 1/10/100/1000/1000 parallel connections: Old version (poll): 10/12/22/435/10500 us. New version(epoll): 10/13/8/6/6 us.
443.
Libmill v1.2: 1700x improvement in performance(freelists.org)
4 points by rumcajz 11y ago|1 comments
444.
by rumcajz 11y ago
One feature worth pointing out is debugging support: http://libmill.org/documentation.html#debug
445.
by rumcajz 11y ago
It's a hack, but one that's relatively benign: allocate stack using malloc, move sp using variable-sized array.
446.
by rumcajz 11y ago
when using native i/o take care to use it in nonblocking manner. if it would block use libmill's fdwait() to wait while the file descriptor is unblocked. in case of long computations call yield() once in a while to give other coro
447.
by rumcajz 11y ago
It's basically the same thing, but simpler and more familiar (if you dealt with Go, that is) API.
448.
by rumcajz 11y ago
My primary use case was writing network protocols (the stuff I had to deal extensively when implementing ZeroMQ and nanomsg). On one hand you want ultimate performance, i.e. no context switchers, manual memory allocation et c. On the other
449.
by rumcajz 11y ago
I don't think there's a data race, given that the whole thing is single-threaded. Multiple processes should be used for parallelism (see step 7 of the tutorial). As for the default stack size, it's hard (but doable) to work w
450.
by rumcajz 11y ago
Simple stack switching. See here for a toy example: http://250bpm.com/blog:48

More ›