Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
runT1ME
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
58 ms
·
511.
▲
by
runT1ME
15y ago
Doesn't that imply that Apes/chimps, or even wolves and lions are able to recognize and identify weapons? Or that their fear of humans evolved at the same rate as we adapted to use tools?
512.
▲
by
runT1ME
15y ago
I'm not convinced the Actor model is a concurrency silver bullet. I'm using it in Scala, and you still have race conditions and potential deadlocks to worry about. It helps to 'segregate' your application so it's easier to reason about wha
513.
▲
by
runT1ME
15y ago
This is a ridiculous assertion. I believe it's the kind of attitude that keeps our profession from gaining respect among other professional groups. I've cooked meals for people that were impressed and happy, does that make me a great chef
514.
▲
by
runT1ME
15y ago
I always get a little suspicious about how little Twilio talks about their SIP stack. Barely a mention of it except for 'Asterisk/FreeSwitch/JSR289'. All of those ostensibly can provide the same functionality (SIP, Sip Termination) yet no
515.
▲
by
runT1ME
15y ago
VRapper is fantastic! And I use the Scala plugin, of course.
516.
▲
by
runT1ME
15y ago
I think DPP has a fundamentally flawed argument: He seems to be assuming that all those Java developers that fail to learn Scala are using Java successfully . Sure, their projects may compile, they may even be deployed and used on a daily
517.
▲
by
runT1ME
15y ago
Scala is pretty much the opposite of a dynamic language...
518.
▲
Ask HN: Why is my avg going down...
4 points
by
runT1ME
15y ago
|
5 comments
519.
▲
by
runT1ME
15y ago
I don't think there's anything to prevent more than one thread from being scheduled at any time. They just block when trying to run concurrently because of the GIL. >I'm not sure if you misunderstood the mail. The constant locking woul
520.
▲
by
runT1ME
15y ago
It's the same reason you pay a surgeon a lot of money. Sure, 80% of the time their job is somewhat cookie cutter, and maybe even a med student could do it. it's the 20% of the time when things get tricky that you really want an experienced
521.
▲
by
runT1ME
15y ago
>I believe his argument is that it would reduce thrashing between the caches. With the GIL ownership of a cache line containing the reference count for any given object will only have to be transfered at most once per timeslice. Ah. Ma
522.
▲
by
runT1ME
15y ago
>That is, if one processor writes to memory in a cache-line shared by another processor, they must stop whatever they are doing to synchronize the dirty cache lines with RAM. Thus, updating reference counts would flood the memory bus wi
523.
▲
by
runT1ME
15y ago
Do you have actual evidence of this? Can you point out a superstar engineer who will work for 65k + options? I can't think of any, and I know quite a lot of developers. I haven't even heard of one working for that little.
524.
▲
by
runT1ME
15y ago
Couldn't I theoretically manufacture a bunch of differently designed crappy tablets that barely work, but may have the shape/aesthetics that future ipads may move to? Seems to me this law would be easily exploited. Instead of a patent por
525.
▲
by
runT1ME
15y ago
40k for six months of my life in an area where I don't have a lot of familiarity or a network of business associates or friends isn't a great deal. Six months of my life to win a contract with the Chilean government to use my software, and
526.
▲
by
runT1ME
15y ago
In the words of Marc Fleury, Open Source is not a business model, it's a distribution model . Rather than spending time and resources convincing companies to download and try your product, you instead have to spend time and resources conv
527.
▲
by
runT1ME
15y ago
Why isn't this running Android? I think I might pay $299 for a nice 12'' netbook running linux...
528.
▲
by
runT1ME
15y ago
Functional and hybrid functional languages benefit greatly from a highly tuned GC. I don't see LLVM supporting something on par with the JVM's collector anytime soon so I think you'll lose a lot of the speed you see with Scala on the JVM
529.
▲
by
runT1ME
15y ago
Is there a reason you guys went with FreeSWITCH and not SIP Servlets/MobiCents or some other proprietary stuff he backend SIP code? Asterisk sucks for SIP...
530.
▲
by
runT1ME
15y ago
I disagree, getting a SIP server up and running in EC2 is no harder than doing it on it's own. Unless you're doing mixing/conferencing, you shouldn't see any delays from the virtualization.
531.
▲
by
runT1ME
15y ago
I'm a fan of Tropo and have recommended it to clients, and overall think Voxeo is the team to beat in the telephony market. However, I feel you're exaggerating a bit here: >VoiceXML is how the big boys build telephony applications. The
532.
▲
by
runT1ME
15y ago
>CLR+CIL are vastly superior to JVM, especially in memory management department. This is an extremely disingenuous statement. You may be referring to the fact that the CLR/C# supports the notion of structs, and allows native memory man
533.
▲
by
runT1ME
15y ago
All three can be used for the same functionality: scala> object Test { | def myMethod( f:(String,String)=>String) { | println("MyMethod result =" + f("a", "b")) | } | } defined module Test scala> Test.myMetho
534.
▲
by
runT1ME
15y ago
I disagree. The following all mean the same thing (off the top of my head): obj.( (a, b) => a + "." + b) obj.(_ + "." + _) obj { (a,b) => a + "." + b } Seems to me that's syntactic sugar. The flexibility of Scala is t
535.
▲
by
runT1ME
15y ago
Did you have concurrent collections turned on? It almost always reduces overall throughput of the collector. However in many workloads I expect the G1 to be superior to the current throughput collector....
536.
▲
by
runT1ME
15y ago
>The philosophy of "why build something, because company x, with infinite money, will just build it better" I think is a flawed one. Fascinating quote, but do you think you can articulate why it is flawed? Was it something your invest
537.
▲
by
runT1ME
15y ago
Wow, I think this article is taking quite a leap calling shrinking industries dying. I'm not sure anyone would argue that the typewriter industry is anything but dying, or the pager market. However, does anyone really think that wired tele
538.
▲
by
runT1ME
16y ago
Brazilian Jiu-Jitsu. It's the chess of the martial arts world. Practicing it keeps you fit, but more often than not, in competitions and live sparring practice, it's superior strategy that overcomes physical prowess. I hated lifting weigh
539.
▲
by
runT1ME
16y ago
>This is how most open source software companies make money. This usually includes some horrendous subscription fee and guaranteed SLAs. I never quite understood, why this works. My only explanation is, that many customers are comfortab
540.
▲
by
runT1ME
16y ago
I know a few developers there. If they're good enough, they still work there and get a very high salary... The problem is not that great developers are expensive, it's that they're expensive and very rare.
More ›