The bulk of nodes we sell are *first gen* ATOM. Yeah, first gen. And they idle mostly, since our workload is I/O intensive, not CPU. Even the highend gear we purchase is 2 gens old - but it's still higher end than the last high end gear.
1U Dual Quad Xeon L5520 with 72Gb ram and less than 500$ and room for 4x3.5"? Who could resist that, when we used to pay close to 200$ per month for a Xeon W3560, 32G Ram with 2x2Tb drives. And the CPUs offer more more power than the Xeon W3560, which was introduced a year earlier, yet those W3560 nodes were purchased just 1½ years ago from a vendor. The Dual Quad Xeons sit 95% idle -> so they don't consume that much power neither.
At home, i have a Core2Extreme 9770 with just 8G ram as my workstation - only lately has the CPU started to *nearly* max out, but i have laying around a spare FX6100 with 32Gb ECC i'll install some day when i can be arsed to. Still using that old of a CPU, and i'm not in any hurry to upgrade neither.
Indeed, at the end of they what matters for 80%+ is the 80% of users - ie. Average Joes. Very few people purchase the very top of the line, it makes absolutely no sense to pay triple for marginal gains - even if comparing within the intel brand. For a very few people it does make sense however.
We use quite a few AMD products in our DC - they are very solid, and very nice performance to price ratio. Because AMD is not as much used, we don't have the multitude of choices, but the highest end difference is:
Intel: Intel D2500CCE: Dual NIC integrated, 4Gb RAM Max Mobo: ~75€ RAM: ~25€ Total: 100€ Power: ~20W Cost per Gb of Ram: 25€
Intel system which compares: Intel DH61AG Mobo ~90€, CPUs begin at 150€ (low power version), 25€ 19V Powersupply RAM: 16Gb So-Dimm as new 110€ Total: ~375€ fluctuates depending upon supply (low power version supply fluctuates badly) Power: ~40W Cost per Gb of RAM: 23.44€
AMD: E350 from Gigabyte: 16Gb RAM Max but needs additional NIC Mobo: varies GREATLY from 50€ to 65€, RAM ~90€, Additional NIC: 7€ Total: 157€ average Power: ~30W Cost Per Gb of RAM: 9.81€
So the AMD motherboard falls in the middle very nicely, since we deal in DATA the only figure at the end of the day what matters is RAM/€. Also the price customers are willing to pay depends solely on storage + ram, not the cpu itself.
I haven't looked on CPU power metric at all - don't care. Even the D410 atoms in our use are like 80% idle, it's all I/O per euro what matters, so bulk of our cost is in disks, ram and networking.
However, intel is better on Wattage and Size metrics on the very low end, but this is mostly because AMD is not simply used as widely, so there isn't the niche boards available, AMD E350 board choices are *very* limited and supply is *very* limited as well.
However, in the past when we used Dual Opteron servers from a 3rd party DC, i want it curious that none of them actually worked, almost every one had broken CPUs, most of them crashed randomly etc. So i'm thinking the game was rigged at some level. They eventually removed AMD option completely, as only 10% of the servers worked. On our own DCs, the AMD gear has worked brilliantly, except where game is rigged. We tried to do software router using AMD CPU, turns out the game is heavily rigged towards Intel as Intel is the one developing the software routing codebase, many generations older Core2Duo was faster than FX6100, further only Intel NICs provided actual performance.
Yea, genius and madness tend to go hand in hand - same traits cause both. There's been research into it, and if i recall right they pretty much concluded it's almost like flip sides of a coin.
History shows us as well, many geniuses of our past were... A bit out of their mind, but because of that they are also geniuses -> they think about things no one else would, they see those things differently and approach differently.
I believe part of the reason genius minds go a bit mad, or some genius persons seems very obnoxious is because everyone else seems to stupid, and it gets so insanely frustfrating when people cannot understand what seems basic logic from the POV of the genius, and you have to deal with it every single F* day, but from the POV of the genius there is just so many morons.
Bottomline is - if they are a bit out of their mind, or seemingly completely bonkers - we should be glad and thankfull, because that is the person who is likely to come to a solution to a very hard problem concerning few or many, pushing our civilization forwards. Also because people fight against them calling "bullshit" at their "crazy inventions", they have to proof beyond shadow of doubt this is better, while it sucks for the genius it is better for civilization as a whole
Ever since 2.96 it's been crappier by every version. I still usually download the 2.96 TO DATE, because the newer ones are such crap, bloated adware shit.
2.96 was simple, yet powerfull, enough features but not bloated. Worked as a MP3 player brilliantly. Then came the stupid trying to play video shit (thus loosing your playlist), the GUI was changed to bloatware etc.
I did a mistake on last system setup -> i installed the latest. Now every now and then when watching netflix or videos on VLC -> it jumps out on top of everything. Yay, that's EXACTLY what i wanted.
So - in other words, you are saying, excel requires less than 40ms latency? watching a remote movie requires less than 40ms latency? Wow, did netflix just take fiber to everyhome to achieve that?
40ms being the usual latency or there abouts for adsl 2 to a geographically nearby server. if you allow 70ms you can go as far as from northern europe to central europe etc.
If the latency is below 41.67ms - it's imperceivable visually.
Does Counter Strike require less than 40ms latency? Yes and No. Many people played fine with cable with it's 60-70ms latency, and most people have 40ms latency nowadays. But i wish for the ADSL gen1 days with 15ms latency - it did make a difference. But if you have sufficiently good remote desktop code: Your latency is just above the network latency, so if you have 40ms latency, you might have 45ms if it's sufficiently good, if you are playing on remote desktop local server, that has less than 0.5ms latency, depending upon the number of switches between. Most likely less than 0.15ms latency - makes no difference anymore.
So why wouldn't it work?
It's a mental handicap why you think it's can't work!
one problem: Supply barely fluctuates. It is adjusted every 2 weeks, when there is vast network growth, supply is temporarily higher yes, when the network shrinks, the supply is temporarily lower yes, but the fluctuation during normal times is smalelr than you expect - maybe 20%. So one hour you might get 150 bitcoins, while another makes only 100. also the supply is about to go down, until maximum of 21mil is reached.
Actually supply is limited by 125 new bitcoins per hour, or well, slightly more, but adjusted to by weekly.
If you need the bitcoins today, you will spend that 400$. Mining -> Difficulty raises constantly etc. to get that bitcoin, you might need to spend 2 years mining for it, account for electricity, hw and time in the cost. Easily winds up cosnting more than 400$
Further, mining is a investment, not an acquirement. Mining takes time. Just because a miner costs 90$, doesn't mean you see 100$ even within a year. Plus electricity, plus increasing difficulty etc.
Your maths are wrong, and the latter example -> People will start doing what they *like* instead of what they *must* to have a good life. Someone wanting to be doctor, and not having the cash for education, will run a restaurant for now, while doing night studies on the cheap etc. prepping until the day his assets have appreciated enough.
In your example, banks accept savings in cash, while giving loans in cash. After 50% deflation event, they still ahve 90k liabilities, and 100k in assets.
Farmer on the other hand, because his asset is tangible, will loose virtually 50% of the number value, but also at the same time, the buying power of that asset remains same, ie a new farm would cost 50% less, ie. the same number value in assets.
Bank's balance sheet would in real world however likely show close to 50% decrease in assets (loans given out), as people have 50% less capacity to pay off their debts, however, as it's a compounding effect, they might loose close to 100% of their assets since only a few are now able to pay up, and go belly up, and in expectance of further deflation, new loans would dwindle, not many given out as people will rather hold on to their money.
Therefore, banks are made irrelevant, only function remaining would be gatekeepers as payment service providers - but who would want them to handle such things with their arbitrary blockings, chance of loosing your money etc etc etc. absolutely zero protection of your assets held in such a provider (ie. paypal), unless you absolutely have to in order to conduct business?
Just because there is 50% deflationary event, does not mean that bank can just say "we'll take half of your money", unless there is some legal loophole, in which case riots will ensue next.
what, i don't see any problems in what you describe. You hold BTC -> It gets more valuable over time. If you preorder something, or simply pay upfront for somethign to be delivered in 2 years - then that is a problem for the buyer.
lol. That's just limit of the current clients and accepted norm. Bitcoin itself, by protocol, is as diviseable as computers can represent, the current limit is imposed to have a limitation on transaction size in the blockchain, nothing else. The division maximum was already increased once.
Bitcoin has no limit on the division. Practical issues could happen but you can have 0.00000000000000000000000 00000000000000000000 0000000000000000000 0000000000 0000000000000000000 00000000000000000000000 000000000000000000000 00000 00000000000000000 000000000000000000000000000000 00000000000 0000000000000 0000000000000000000000000 00000000000000 000000000000000000000 000000000001 bitcoins if you wish to or even smaller denominations:)
About damn time someone was considering the impact of human heat sources, such as cars, electricity generation etc. I've been wondering this for a long time - why no one is considering this?! Basicly that heat has no where to go - some of it does radiate to space, but the heat what planet earth radiates away is rather limited, space being vacuum.
I wouldn't be surprised if the whole "global warming" would be all human heat caused, while pollution is minimalistic. Seriously tho, we shouldn't even blame pollution as long as we allow the super tankers, which of there is tens and each causes more pollution than 55 million cars. Global warming got changed to climate change, as heat generated in Brazil might cause UK to get cooler etc. due to changing weather patterns.
Someone talking about "europe's temperatures getting lower": Realize this is not isolated to continents, the locality of these is this whole planet. Heat generated in New York will affect australia, and heat in australia will affect Moscow. If one particular place is getting cooler, another place is getting warmer by the same amount. In other words, what matters is the total energy content (heat) of the planet. Specific locality (city) only matters in the where and how.
wow, one should not need to convince why doing quality work is important:O Just leave that company, or get all of those who oppose doing quality fired. There really is no middle solutions in situ like this.
lol, that's very easy. Not even basic mathematics and iteration. When you use modulus this indeed doesn't even require math skills to figure out. And these guys call themselves coders?
I must start passing guys this test myself!;D
Any other good 5mins or less coding tests you've used?
And code is written faster with clean code. Way faster. Code architecture, cleanliness is the key to make features faster and eliminate defects.
Dirty code will lend itself to a set of various problems, messy code even further.
Well architectured code base, with a clean writing style will make things happen faster, easier to maintain and every body understands a particular piece easier. On large systems tho, a great mind is required not to only comprehend but also understand the whole system with proper structure.
The rules are very simple: * Every function should have at least short comment of what it does, preferrably a docblock style * every function and variable name should be unshortened, camelCased (my preference) or hyphened without abbreviations or shortening the words, for example instead of uSesAuthGenId type it open userSessionAuthenticationGenerateId you also see this is incorrectly structured it should be more like users->authentication->generateId * Insert a comment approximately every 10 lines on average * Do not type a huge list on intended recursive function calls on single line, ie. $example = array_walk( array_map( $$exampleMethod1, $data1, $conf1 ), $$exampleMethod2, $mode ); is a bad example of code, doubly so if the callbacks use references and change the original data:P That should be written more open and commented * a single method should preferrably not exceed 100 lines, if it does it's doing too much * You should never ever need intendation further that 4 levels, intendation comes from if clauses, for/while/foreach loops etc. for example: foreach($data AS $thisKey => $thisValue) { if ($thisValue['type'] == 'type') { switch($thisValue['mode']) { case 'thing': if ($something != $something2) { foreach($thisValue['someArray']) {.... Is bad * Abstract, but do not abstract too much. Abstract things which require more than say 10 lines, but if you are abstracting single line things via 8 different methods/functions you are doing something HORRIBLY WRONG (even Zend FW falls for this) * When you abstract, do not duplicate code, try to reuse code intelligently, do not copy & paste... *EVER* * Separate, isolate certain types of code, for example: View, Business Logic, Flow control. MVC is not just a "fancy word" forcing you to "a nasty framework i need to work around". Infact MVC is a very old concept. Learn it, digest it, understand it. For example anyone saying Smarty is not MVC, or Smarty is MC+V or stupid things like that do *NOT* understand what MVC is and how to write structured code. * Above all: Reuse, Reuse, Reuse, Reuse. Don't do practically the same thing in 50 different ways, do the same thing 1 time but do it excellently and reuse it elsewhere. Goes for architecture & structure, for layer separation, and line level code. For example, each model method shouldn't have it's own handling of MySQL result set, they should ask another model for the results.
And the most important lessons today are: Creating code is 7 times more reading than writing, it's 10 times better to spend 50 minutes planning/designing a particular feature then writing it in 5minutes rather than use 45 min writing it and 0 minutes planning it (First change arises you will know why).
One can create a particular system using just ~25k lines of code instead of 200k lines of code if it's well structured and thought out. One can also spend creating that either half a year in man hours, or 5. It's all in *executed* high worth lines of code, not about quantity of written low worth lines of code.
My basic benchmark for qualitive productive? If code style is properly adhered, how *few* lines of code and what's the ratio of comments vs. code.
Earnings and value of money is also relative, 2$ in India buys hell of a lot more than in the US, nevermind in Finland you can't even get a decent cup of coffee with 2$
So living on less than 2$ per day doesn't tell anything if not put into context of the local economy. For example how much does carton of milk cost? What's the cost of rice per kilo in those areas?
I'm quite sure that 2$ buys at least a kilo of rice in that area and some soybeans to go with it. But how about in US? I know here in Finland you can barely get a kilo of rice with that.
Nokia N900 was AMAZING - Sure it had it's issues to be used as a phone, it simply wasn't polished properly - but that was a real proper geek style smartphone, complete with QWERTY.
I loved it's debian based. The only phone i've REALLY wanted in the last decade or so until the N950 which turned out to be justa developer phone. Then they dropped the production despite demand being high:( Auction prices for used N900s were same as new phones when they were last available for sometime.
N950 if sold to public i'm sure would have sold A LOT. These few models are the only ones i've seen from Nokia in a really long time which combines their HW excellency with some software excellence.
I would buy either N900 or N950 still new to date if they were available - not interested in used phones of this kind, can't be arsed to go through finding out how to reset the system to completely original from factory to avoid 'extras'. I was waiting and hoping for N950 to hit the shelves like the rising moon! I was hugely disappointed when it came clear it's developers only. I would even accepted drawbacks like short battery life, no camera what-so-ever, no bluetooth and even major ones like way too slow CPU, too little RAM, no MicroSD slot!
I even considered getting N9 despite lack of qwerty phone but the demand for Qwerty outweighed, i might need to SSH into servers on the road, and doing even basic maintenance on touchscreen keyboard... no thanks, so i got E7. Since E7 nokia hasn't released any phone worth even considering for my use.
I don't want an Android or IOS phone due to obvious privacy concerns, usability concerns etc. I want Nokia Hardware + Meego! and i'd pay extra for that.
To me, other manufacturers are not even proper choice, Nokia is a phone manufacturer first, everything else then, unlike the other companies. This means Nokia pays attention to things which don't come up anywhere in the development cycle with the likes of Samsung, Apple, HTC etc.
Sure, Nokia hardware ain't perfect neither, but they are *phones* from start to finish, and this shows up in things like durability (mechanical) and battery life which are most obvious, and many many little details. For example, my E7 i've dropped to concrete floor, on asphalt on multiple occasions, but absolutely no damage, the most wear on that phone is on the plastic end parts from being in my pocket. Screen is still like from factory, and the screen is actually so sensitive that you can use thin gloves with it, which i hear other phones are not. Once on a trip i was away for 6½ days without charging my phone before getting back to home and putting it back to charger. Friend with whom i was on the trip had to charge his android phone every night.
So please nokia! Bring back Meego, make a brilliant phone with Aluminium or Titanium case, nice screen, qwerty, protected durable USB connector, plenty of storage + ram, decent extra low power cpu and meego -> You will sell a lot. I can't be alone willing to put out extra 100-200€ for a Nokia phone like that! Even if it's "low end" hardware spec i would still happily put the extra 100-200€ !
like the other commenter said, you have a serious disconnection from reality. There *needs* to be a "top dog" who dictates the qualities of the project, otherwise it ends up being endless debate and project systematic individual style hacks introduced, in the end no one knows how it was supposed to work and how it does work now.
Certainly, it will create some grief on the short term when someone is adamant about bad quality and breaking things, in the long run however -> things work better as a result of that.
Those doing bad job needs to be punished, and those doing a good job needs a reward.
Still Linus was right in doing this. Any bad coder out there like that - who doesn't test properly and breaks things horribly deserves some wrath. I've seen too many bad, so called coders out there, even hired some. Excellent CV has not been a guarantee that even basic processes of the person doing the job is anywhere near the required level for the job.
Especially lack of testing, even if manual 5min testing, is something which is completely unacceptable. In this case, the person probably did not need to do more than compile the kernel, boot a KDE environment and try to play a song. I've not looked into this patch which caused the trouble, but usually you can immediately tell from glancing at the code what to test.
And your example is wrong - in this case Linus is stopping crap to be introduced to mainline and making sure quality is maintained - anyone with basic reading comprehension skills can see that, and infact, for sensible persons that's positive marketing -> knowing that "heads will drop" if quality is not maintained.
By far not sufficient when you get to that level of required security. If there is no alarms, monitoring etc. and reinforced walls, a thief can potentially get inside without anyone noticing through another wall, ceiling or floor.
Practical security:
* Use linux with GRSEC
* All network daemons turned off
* Firewall all ingress, don't even allow ping etc.
* Firewall all egress, only make sure what's ultimately needed is accessible, potentially building a whitelist if possible
* No excess software what-so-ever, just what is ultimately needed
* ROOT account: No logins, create another account which can only be locally logon to, which can sudo. Password 16 chars, potentially automatically rotating. Possibly also having 2 factor authentication. You can trivially create this step by even creating a PHP Script as the shell:)
* USER account: Limited to only what is required, potentially chrooted to the exact data which is required to be accessible etc. Depends on the usability required
* Watch logins: More than 2-5 failed logins, shut the system down immediately using "magic" SYSRQ, wrong username? Instantly
* Full disk encryption, on top of which potentially using a bit obscure filesystem to make it that much harder to break. The required data should have 2nd level encryption unless doing that creates a potential attack vector on the first level encryption
Hardware:
* Potentially use hardware where you can review the firmware/bios if possible
* HW firewall "integrated" to the motherboard, motherboard network connectors are removed and hardwired to this HW firewall, so that even a skilled person would require atleast 20mins to bypass the HW Firewall
* HW Firewall configured in the same sense as the SW firewall, potentially with additional protections.
* Super Epoxy glue all connectors, modules etc. including the HW firewall buttons and it's mainboard into the motherboard etc. -> Stops quick tampering.
* Disk drives and CPU needs cooling, so CPU heatsink could use heat transfer glue to the CPU and super epoxy from the sides on to the motherboard. Disk drives can have little spacing with the super epoxy.
* The whole case is epoxied together/welded. No connector should be accessible, but peripherals mounted permanently with super epoxy to avoid inserting capturing devices directly.
* Braided stainless steel sleeves for all cabling to make splicing in harder.
* Epoxy on the other side of the peripherals as well;)
FW Config: Potentially disabling all unencrypted connections, verifying against known certificates, no other connections allowed, if possible. Potentially also limiting data transfer rates so that if anyone tries to transmit data outside -> it will take long enough for security to take notice. GRSEC configuration is very involved, but can be teached. Process list should be verified and checked against.
This will create a secured SW + HW environment. If you cannot use a motherboard/devices which firmware you can verify, the extreme FW measures taken (both SW + HW) should ensure no data gets transmitted without permission. It is highly doubtful that same organization can be behind a security hole in the motherboard AND the HW FW, but you can also create your own HW FW using things like Arduino where you would be the person creating the firmware as well.
Epoxy: Modern cars are glued together, so just use similar industrial strength epoxy.
In the end it's all about making accessibility slower if it's a highly skilled attacker with knowledge about the system upfront, which can potentially stop the attempted attack all together if it's deemed too secure. BUT Security via obscurity is still not security, i see people changing their SSH ports, blocking Ping etc. but that doesn't really add to security, as the information can still be gathered very quickly.
Slashdot Mirror
Yea and 730+ from 1.6 4AGE engines. One CA18DET from Nissan 200SX is widely known to produce about 730whp as well, so about 780 from the crank.
4AGE is considered very lightweight at 110-120kg, add turbo etc. to it for 10-20kg.
Damn right!
And this goes for servers as well.
The bulk of nodes we sell are *first gen* ATOM. Yeah, first gen. And they idle mostly, since our workload is I/O intensive, not CPU.
Even the highend gear we purchase is 2 gens old - but it's still higher end than the last high end gear.
1U Dual Quad Xeon L5520 with 72Gb ram and less than 500$ and room for 4x3.5"? Who could resist that, when we used to pay close to 200$ per month for a Xeon W3560, 32G Ram with 2x2Tb drives.
And the CPUs offer more more power than the Xeon W3560, which was introduced a year earlier, yet those W3560 nodes were purchased just 1½ years ago from a vendor.
The Dual Quad Xeons sit 95% idle -> so they don't consume that much power neither.
At home, i have a Core2Extreme 9770 with just 8G ram as my workstation - only lately has the CPU started to *nearly* max out, but i have laying around a spare FX6100 with 32Gb ECC i'll install some day when i can be arsed to. Still using that old of a CPU, and i'm not in any hurry to upgrade neither.
Indeed, at the end of they what matters for 80%+ is the 80% of users - ie. Average Joes.
Very few people purchase the very top of the line, it makes absolutely no sense to pay triple for marginal gains - even if comparing within the intel brand. For a very few people it does make sense however.
We use quite a few AMD products in our DC - they are very solid, and very nice performance to price ratio.
Because AMD is not as much used, we don't have the multitude of choices, but the highest end difference is:
Intel:
Intel D2500CCE: Dual NIC integrated, 4Gb RAM Max
Mobo: ~75€ RAM: ~25€
Total: 100€
Power: ~20W
Cost per Gb of Ram: 25€
Intel system which compares:
Intel DH61AG Mobo ~90€, CPUs begin at 150€ (low power version), 25€ 19V Powersupply
RAM: 16Gb So-Dimm as new 110€
Total: ~375€ fluctuates depending upon supply (low power version supply fluctuates badly)
Power: ~40W
Cost per Gb of RAM: 23.44€
AMD:
E350 from Gigabyte: 16Gb RAM Max but needs additional NIC
Mobo: varies GREATLY from 50€ to 65€, RAM ~90€, Additional NIC: 7€
Total: 157€ average
Power: ~30W
Cost Per Gb of RAM: 9.81€
So the AMD motherboard falls in the middle very nicely, since we deal in DATA the only figure at the end of the day what matters is RAM/€.
Also the price customers are willing to pay depends solely on storage + ram, not the cpu itself.
I haven't looked on CPU power metric at all - don't care. Even the D410 atoms in our use are like 80% idle, it's all I/O per euro what matters, so bulk of our cost is in disks, ram and networking.
However, intel is better on Wattage and Size metrics on the very low end, but this is mostly because AMD is not simply used as widely, so there isn't the niche boards available, AMD E350 board choices are *very* limited and supply is *very* limited as well.
However, in the past when we used Dual Opteron servers from a 3rd party DC, i want it curious that none of them actually worked, almost every one had broken CPUs, most of them crashed randomly etc. So i'm thinking the game was rigged at some level. They eventually removed AMD option completely, as only 10% of the servers worked.
On our own DCs, the AMD gear has worked brilliantly, except where game is rigged. We tried to do software router using AMD CPU, turns out the game is heavily rigged towards Intel as Intel is the one developing the software routing codebase, many generations older Core2Duo was faster than FX6100, further only Intel NICs provided actual performance.
Only way to properly compare the pricing is to include mobo, cpu, ram and gpu.
No other way around that.
So in other words: People are the issue.
Lol
Yea, genius and madness tend to go hand in hand - same traits cause both. There's been research into it, and if i recall right they pretty much concluded it's almost like flip sides of a coin.
History shows us as well, many geniuses of our past were... A bit out of their mind, but because of that they are also geniuses -> they think about things no one else would, they see those things differently and approach differently.
I believe part of the reason genius minds go a bit mad, or some genius persons seems very obnoxious is because everyone else seems to stupid, and it gets so insanely frustfrating when people cannot understand what seems basic logic from the POV of the genius, and you have to deal with it every single F* day, but from the POV of the genius there is just so many morons.
Bottomline is - if they are a bit out of their mind, or seemingly completely bonkers - we should be glad and thankfull, because that is the person who is likely to come to a solution to a very hard problem concerning few or many, pushing our civilization forwards.
Also because people fight against them calling "bullshit" at their "crazy inventions", they have to proof beyond shadow of doubt this is better, while it sucks for the genius it is better for civilization as a whole
If he'd sign a message using the wallet that is known to be owned by him...
Ever since 2.96 it's been crappier by every version. I still usually download the 2.96 TO DATE, because the newer ones are such crap, bloated adware shit.
2.96 was simple, yet powerfull, enough features but not bloated. Worked as a MP3 player brilliantly.
Then came the stupid trying to play video shit (thus loosing your playlist), the GUI was changed to bloatware etc.
I did a mistake on last system setup -> i installed the latest. Now every now and then when watching netflix or videos on VLC -> it jumps out on top of everything. Yay, that's EXACTLY what i wanted.
So - in other words, you are saying, excel requires less than 40ms latency? watching a remote movie requires less than 40ms latency? Wow, did netflix just take fiber to everyhome to achieve that?
40ms being the usual latency or there abouts for adsl 2 to a geographically nearby server.
if you allow 70ms you can go as far as from northern europe to central europe etc.
If the latency is below 41.67ms - it's imperceivable visually.
Does Counter Strike require less than 40ms latency? Yes and No. Many people played fine with cable with it's 60-70ms latency, and most people have 40ms latency nowadays. But i wish for the ADSL gen1 days with 15ms latency - it did make a difference.
But if you have sufficiently good remote desktop code: Your latency is just above the network latency, so if you have 40ms latency, you might have 45ms if it's sufficiently good, if you are playing on remote desktop local server, that has less than 0.5ms latency, depending upon the number of switches between. Most likely less than 0.15ms latency - makes no difference anymore.
So why wouldn't it work?
It's a mental handicap why you think it's can't work!
one problem: Supply barely fluctuates.
It is adjusted every 2 weeks, when there is vast network growth, supply is temporarily higher yes, when the network shrinks, the supply is temporarily lower yes, but the fluctuation during normal times is smalelr than you expect - maybe 20%.
So one hour you might get 150 bitcoins, while another makes only 100.
also the supply is about to go down, until maximum of 21mil is reached.
Actually supply is limited by 125 new bitcoins per hour, or well, slightly more, but adjusted to by weekly.
If you need the bitcoins today, you will spend that 400$.
Mining -> Difficulty raises constantly etc. to get that bitcoin, you might need to spend 2 years mining for it, account for electricity, hw and time in the cost. Easily winds up cosnting more than 400$
Further, mining is a investment, not an acquirement.
Mining takes time. Just because a miner costs 90$, doesn't mean you see 100$ even within a year. Plus electricity, plus increasing difficulty etc.
Your maths are wrong, and the latter example -> People will start doing what they *like* instead of what they *must* to have a good life.
Someone wanting to be doctor, and not having the cash for education, will run a restaurant for now, while doing night studies on the cheap etc. prepping until the day his assets have appreciated enough.
In your example, banks accept savings in cash, while giving loans in cash.
After 50% deflation event, they still ahve 90k liabilities, and 100k in assets.
Farmer on the other hand, because his asset is tangible, will loose virtually 50% of the number value, but also at the same time, the buying power of that asset remains same, ie a new farm would cost 50% less, ie. the same number value in assets.
Bank's balance sheet would in real world however likely show close to 50% decrease in assets (loans given out), as people have 50% less capacity to pay off their debts, however, as it's a compounding effect, they might loose close to 100% of their assets since only a few are now able to pay up, and go belly up, and in expectance of further deflation, new loans would dwindle, not many given out as people will rather hold on to their money.
Therefore, banks are made irrelevant, only function remaining would be gatekeepers as payment service providers - but who would want them to handle such things with their arbitrary blockings, chance of loosing your money etc etc etc. absolutely zero protection of your assets held in such a provider (ie. paypal), unless you absolutely have to in order to conduct business?
Just because there is 50% deflationary event, does not mean that bank can just say "we'll take half of your money", unless there is some legal loophole, in which case riots will ensue next.
what, i don't see any problems in what you describe.
You hold BTC -> It gets more valuable over time. If you preorder something, or simply pay upfront for somethign to be delivered in 2 years - then that is a problem for the buyer.
lol. That's just limit of the current clients and accepted norm. Bitcoin itself, by protocol, is as diviseable as computers can represent, the current limit is imposed to have a limitation on transaction size in the blockchain, nothing else. The division maximum was already increased once.
Bitcoin has no limit on the division. Practical issues could happen but you can have 0.00000000000000000000000 00000000000000000000 0000000000000000000 0000000000 0000000000000000000 00000000000000000000000 000000000000000000000 00000 00000000000000000 000000000000000000000000000000 00000000000 0000000000000 0000000000000000000000000 00000000000000 000000000000000000000 000000000001 bitcoins if you wish to or even smaller denominations:)
About damn time someone was considering the impact of human heat sources, such as cars, electricity generation etc.
I've been wondering this for a long time - why no one is considering this?!
Basicly that heat has no where to go - some of it does radiate to space, but the heat what planet earth radiates away is rather limited, space being vacuum.
I wouldn't be surprised if the whole "global warming" would be all human heat caused, while pollution is minimalistic. Seriously tho, we shouldn't even blame pollution as long as we allow the super tankers, which of there is tens and each causes more pollution than 55 million cars.
Global warming got changed to climate change, as heat generated in Brazil might cause UK to get cooler etc. due to changing weather patterns.
Someone talking about "europe's temperatures getting lower": Realize this is not isolated to continents, the locality of these is this whole planet. Heat generated in New York will affect australia, and heat in australia will affect Moscow.
If one particular place is getting cooler, another place is getting warmer by the same amount.
In other words, what matters is the total energy content (heat) of the planet. Specific locality (city) only matters in the where and how.
wow, one should not need to convince why doing quality work is important :O
Just leave that company, or get all of those who oppose doing quality fired. There really is no middle solutions in situ like this.
lol, that's very easy. Not even basic mathematics and iteration. When you use modulus this indeed doesn't even require math skills to figure out.
And these guys call themselves coders?
I must start passing guys this test myself! ;D
Any other good 5mins or less coding tests you've used?
And code is written faster with clean code. Way faster.
Code architecture, cleanliness is the key to make features faster and eliminate defects.
Dirty code will lend itself to a set of various problems, messy code even further.
Well architectured code base, with a clean writing style will make things happen faster, easier to maintain and every body understands a particular piece easier. On large systems tho, a great mind is required not to only comprehend but also understand the whole system with proper structure.
The rules are very simple: :P That should be written more open and commented .... Is bad ... *EVER*
* Every function should have at least short comment of what it does, preferrably a docblock style
* every function and variable name should be unshortened, camelCased (my preference) or hyphened without abbreviations or shortening the words, for example instead of uSesAuthGenId type it open userSessionAuthenticationGenerateId you also see this is incorrectly structured it should be more like users->authentication->generateId
* Insert a comment approximately every 10 lines on average
* Do not type a huge list on intended recursive function calls on single line, ie. $example = array_walk( array_map( $$exampleMethod1, $data1, $conf1 ), $$exampleMethod2, $mode ); is a bad example of code, doubly so if the callbacks use references and change the original data
* a single method should preferrably not exceed 100 lines, if it does it's doing too much
* You should never ever need intendation further that 4 levels, intendation comes from if clauses, for/while/foreach loops etc. for example: foreach($data AS $thisKey => $thisValue) { if ($thisValue['type'] == 'type') { switch($thisValue['mode']) { case 'thing': if ($something != $something2) { foreach($thisValue['someArray']) {
* Abstract, but do not abstract too much. Abstract things which require more than say 10 lines, but if you are abstracting single line things via 8 different methods/functions you are doing something HORRIBLY WRONG (even Zend FW falls for this)
* When you abstract, do not duplicate code, try to reuse code intelligently, do not copy & paste
* Separate, isolate certain types of code, for example: View, Business Logic, Flow control. MVC is not just a "fancy word" forcing you to "a nasty framework i need to work around". Infact MVC is a very old concept. Learn it, digest it, understand it. For example anyone saying Smarty is not MVC, or Smarty is MC+V or stupid things like that do *NOT* understand what MVC is and how to write structured code.
* Above all: Reuse, Reuse, Reuse, Reuse. Don't do practically the same thing in 50 different ways, do the same thing 1 time but do it excellently and reuse it elsewhere. Goes for architecture & structure, for layer separation, and line level code. For example, each model method shouldn't have it's own handling of MySQL result set, they should ask another model for the results.
And the most important lessons today are: Creating code is 7 times more reading than writing, it's 10 times better to spend 50 minutes planning/designing a particular feature then writing it in 5minutes rather than use 45 min writing it and 0 minutes planning it (First change arises you will know why).
One can create a particular system using just ~25k lines of code instead of 200k lines of code if it's well structured and thought out.
One can also spend creating that either half a year in man hours, or 5.
It's all in *executed* high worth lines of code, not about quantity of written low worth lines of code.
My basic benchmark for qualitive productive? If code style is properly adhered, how *few* lines of code and what's the ratio of comments vs. code.
Earnings and value of money is also relative, 2$ in India buys hell of a lot more than in the US, nevermind in Finland you can't even get a decent cup of coffee with 2$
So living on less than 2$ per day doesn't tell anything if not put into context of the local economy. For example how much does carton of milk cost? What's the cost of rice per kilo in those areas?
I'm quite sure that 2$ buys at least a kilo of rice in that area and some soybeans to go with it. But how about in US?
I know here in Finland you can barely get a kilo of rice with that.
Nokia N900 was AMAZING - Sure it had it's issues to be used as a phone, it simply wasn't polished properly - but that was a real proper geek style smartphone, complete with QWERTY.
I loved it's debian based. The only phone i've REALLY wanted in the last decade or so until the N950 which turned out to be justa developer phone. :(
Then they dropped the production despite demand being high
Auction prices for used N900s were same as new phones when they were last available for sometime.
N950 if sold to public i'm sure would have sold A LOT.
These few models are the only ones i've seen from Nokia in a really long time which combines their HW excellency with some software excellence.
I would buy either N900 or N950 still new to date if they were available - not interested in used phones of this kind, can't be arsed to go through finding out how to reset the system to completely original from factory to avoid 'extras'.
I was waiting and hoping for N950 to hit the shelves like the rising moon! I was hugely disappointed when it came clear it's developers only. I would even accepted drawbacks like short battery life, no camera what-so-ever, no bluetooth and even major ones like way too slow CPU, too little RAM, no MicroSD slot!
I even considered getting N9 despite lack of qwerty phone but the demand for Qwerty outweighed, i might need to SSH into servers on the road, and doing even basic maintenance on touchscreen keyboard ... no thanks, so i got E7. Since E7 nokia hasn't released any phone worth even considering for my use.
I don't want an Android or IOS phone due to obvious privacy concerns, usability concerns etc. I want Nokia Hardware + Meego! and i'd pay extra for that.
To me, other manufacturers are not even proper choice, Nokia is a phone manufacturer first, everything else then, unlike the other companies. This means Nokia pays attention to things which don't come up anywhere in the development cycle with the likes of Samsung, Apple, HTC etc.
Sure, Nokia hardware ain't perfect neither, but they are *phones* from start to finish, and this shows up in things like durability (mechanical) and battery life which are most obvious, and many many little details.
For example, my E7 i've dropped to concrete floor, on asphalt on multiple occasions, but absolutely no damage, the most wear on that phone is on the plastic end parts from being in my pocket. Screen is still like from factory, and the screen is actually so sensitive that you can use thin gloves with it, which i hear other phones are not. Once on a trip i was away for 6½ days without charging my phone before getting back to home and putting it back to charger.
Friend with whom i was on the trip had to charge his android phone every night.
So please nokia! Bring back Meego, make a brilliant phone with Aluminium or Titanium case, nice screen, qwerty, protected durable USB connector, plenty of storage + ram, decent extra low power cpu and meego -> You will sell a lot.
I can't be alone willing to put out extra 100-200€ for a Nokia phone like that! Even if it's "low end" hardware spec i would still happily put the extra 100-200€ !
like the other commenter said, you have a serious disconnection from reality.
There *needs* to be a "top dog" who dictates the qualities of the project, otherwise it ends up being endless debate and project systematic individual style hacks introduced, in the end no one knows how it was supposed to work and how it does work now.
Certainly, it will create some grief on the short term when someone is adamant about bad quality and breaking things, in the long run however -> things work better as a result of that.
Those doing bad job needs to be punished, and those doing a good job needs a reward.
Still Linus was right in doing this.
Any bad coder out there like that - who doesn't test properly and breaks things horribly deserves some wrath.
I've seen too many bad, so called coders out there, even hired some. Excellent CV has not been a guarantee that even basic processes of the person doing the job is anywhere near the required level for the job.
Especially lack of testing, even if manual 5min testing, is something which is completely unacceptable.
In this case, the person probably did not need to do more than compile the kernel, boot a KDE environment and try to play a song. I've not looked into this patch which caused the trouble, but usually you can immediately tell from glancing at the code what to test.
And your example is wrong - in this case Linus is stopping crap to be introduced to mainline and making sure quality is maintained - anyone with basic reading comprehension skills can see that, and infact, for sensible persons that's positive marketing -> knowing that "heads will drop" if quality is not maintained.
By far not sufficient when you get to that level of required security.
If there is no alarms, monitoring etc. and reinforced walls, a thief can potentially get inside without anyone noticing through another wall, ceiling or floor.
Practical security: :)
* Use linux with GRSEC
* All network daemons turned off
* Firewall all ingress, don't even allow ping etc.
* Firewall all egress, only make sure what's ultimately needed is accessible, potentially building a whitelist if possible
* No excess software what-so-ever, just what is ultimately needed
* ROOT account: No logins, create another account which can only be locally logon to, which can sudo. Password 16 chars, potentially automatically rotating. Possibly also having 2 factor authentication. You can trivially create this step by even creating a PHP Script as the shell
* USER account: Limited to only what is required, potentially chrooted to the exact data which is required to be accessible etc. Depends on the usability required
* Watch logins: More than 2-5 failed logins, shut the system down immediately using "magic" SYSRQ, wrong username? Instantly
* Full disk encryption, on top of which potentially using a bit obscure filesystem to make it that much harder to break. The required data should have 2nd level encryption unless doing that creates a potential attack vector on the first level encryption
Hardware: ;)
* Potentially use hardware where you can review the firmware/bios if possible
* HW firewall "integrated" to the motherboard, motherboard network connectors are removed and hardwired to this HW firewall, so that even a skilled person would require atleast 20mins to bypass the HW Firewall
* HW Firewall configured in the same sense as the SW firewall, potentially with additional protections.
* Super Epoxy glue all connectors, modules etc. including the HW firewall buttons and it's mainboard into the motherboard etc. -> Stops quick tampering.
* Disk drives and CPU needs cooling, so CPU heatsink could use heat transfer glue to the CPU and super epoxy from the sides on to the motherboard. Disk drives can have little spacing with the super epoxy.
* The whole case is epoxied together/welded. No connector should be accessible, but peripherals mounted permanently with super epoxy to avoid inserting capturing devices directly.
* Braided stainless steel sleeves for all cabling to make splicing in harder.
* Epoxy on the other side of the peripherals as well
FW Config: Potentially disabling all unencrypted connections, verifying against known certificates, no other connections allowed, if possible. Potentially also limiting data transfer rates so that if anyone tries to transmit data outside -> it will take long enough for security to take notice.
GRSEC configuration is very involved, but can be teached.
Process list should be verified and checked against.
This will create a secured SW + HW environment.
If you cannot use a motherboard/devices which firmware you can verify, the extreme FW measures taken (both SW + HW) should ensure no data gets transmitted without permission. It is highly doubtful that same organization can be behind a security hole in the motherboard AND the HW FW, but you can also create your own HW FW using things like Arduino where you would be the person creating the firmware as well.
Epoxy: Modern cars are glued together, so just use similar industrial strength epoxy.
In the end it's all about making accessibility slower if it's a highly skilled attacker with knowledge about the system upfront, which can potentially stop the attempted attack all together if it's deemed too secure.
BUT Security via obscurity is still not security, i see people changing their SSH ports, blocking Ping etc. but that doesn't really add to security, as the information can still be gathered very quickly.