« first day (1003 days earlier)      last day (4016 days later) » 

1:20 AM
check out this super duper question
r
1
Q: Calculate network addresses

NickI'm asked to divide up 10.64.0.0/15 into three subnets for |Group | Min # of Hosts| |Group 1| 70,000 | |Group 2| 30,000 | |Group 3| 12,000 | so I start by figuring out that 32-15 = 17 => 2^17 = 131072 are the total amount of addresses. Begin to split that number into p...

I think it'd be better placed at serverfault.
anybody here brave enough to even attempt it?
divide up 10.64.0.0/15 into three subnets for
Group 1: 70,000
Group 2: 30,000
Group 3: 12,000
With 17 bits left.. it seems like it might be possible 'cos 128K addresses.
 
 
2 hours later…
3:08 AM
@JourneymanGeek Hackintosh.
 
@Griffin: 2 things. Not the same, and you'll end up spending a lot of time getting the damned thing to work
 
What's different?
 
@Griffin: quite a lot of hardware isn't supported
my old laptop had it installed. Both the mic and audio out are stuck on MAX VOLUME, wireless isn't supported ect
 
3:40 AM
Does anyone actually use the 172.16.x.x range?
 
any javites here?
 
4:10 AM
Heh. Got the fourth Reversal badge on MSO.
 
@JourneymanGeek Know how to use lists in java?
 
If I said yes, what would you say?
 
I'd say help me :)
What to import for list?
 
org.apache.commons.collections.list.TreeList
 
4:18 AM
I mean the inbuilt(?) ArrayLists
 
java.util.AbstractList
If you're in Eclipse, just press Ctrl-Shift-O
 
cannot find symbol - class ArrayList
 ArrayList<Integer> y = new ArrayList<Integer>();
On that
Am I doing something wrong
 
Did you import java.util.ArrayList at the top?
 
yep
Wait no
Cool!
It works. Thanks
class add3
    {
        public static void main(int n)
        {
            int currentSum=0,currentProduct=1;
            double totalSum=0;
            for(int i=2;i<=n;i++)
                {
                    for(int x=1;x<=i;x++)
                        {
                        currentSum+=x;
                        currentProduct*=x;
                    }
                    totalSum+=currentSum/currentProduct;
                    }
            System.out.println("Result="+totalSum);
            }
WAP to find the sum of the foll series
((1+2)/(1*2))+((1+2+3)/(1*2*3))+........+((1+2+....+n)/(1*2*.......*n))
 
Heh. I forgot how ugly Java was :P
 
4:33 AM
XD :P
I'll make you remember
 
Damn, just lost 40 reputation from a deleted user.
 
can someone please tell me how to fix that program?
 
Why is it broken?
 
Instead of giving 2.5 when I enter 2, it gives me 1
 
your inner loop runs twice then, once with 1, once with 2
currentSum will be 3
currentProduct will be 1*1*2 so 2
so you should get 3/2
But, because you're dividing integers, you get 1 instead of 1.5
Make sure you (double) currentSum / currentProduct
Anyway, with that I'm off on holidays, see you guys.
 
4:46 AM
ok
 
5:04 AM
@slhck: I lost 20 :/
 
I lost 10 :P
 
Bob
5:26 AM
Hm. Lost 10 rep...
@slhck Enjoy your trip?
 
 
1 hour later…
6:44 AM
@AmithKK: Don't use an inner loop. :)
1+2+3 can be computed using (3(3+1))/2; 1*2*3 can be computed using 3!
The currentProduct can grow large too, you probably don't want to use int for it either; for the sake of being safe you might even need it for currentSum depending on how high you the values are that you accept for n.
 
@TomWijsman Oh god! just forgot that formula
 
Bob
@TomWijsman Where were you when I needed to optimise things? ;)
 
Is there something like that for multuplication?
 
If there is a function for 3! then that might be more optimal than running over a loop, but I don't expect too much benefit.
Well, technically, one could implement that with a cache such that after a few calls to it it can start using the highest computed value and compute further.
:P
 
Bob
I'm too slow :P
 
6:53 AM
But yeah, you can implement the cache yourself here easily by just keeping around the previous computed value.
That way you don't have to permute at all, just a simple multiplication.
 
7:35 AM
I've seen a lot of WTFs throughout the network, but this has to take the cake
/submits to #gemsfromstackexchange
 
what the ..
"make one call, change your life" sounds like Abhishek bachchan in idea ad.
 
Bob
7:56 AM
@Sathya what is this I don't even
 
 
2 hours later…
9:46 AM
@Sathya: got rid of two of those on AU. This dude actually had done this before.
 
Bob
@JourneymanGeek why??
 
@Bob: cause it could be done?
 
 
1 hour later…
11:29 AM
@JourneymanGeek he was on Super User yesterday, @OliverSalzburg deleted his post yesterday
 
 
1 hour later…
12:53 PM
Oh snap, Wheezy was released
 
1:43 PM
Hey everyone. You know how some GPU reviews sometimes include power consumption charts? Are those charts predominantly of the whole system and not the GPU alone?
 
@ЯрославРахматуллин I would assume they're the GPU alone
 
If that is the case, I find it very odd that a system with a card that's rated at 116 watts by Nvidia hardly ever turns out to consume more than 160 watts.
Yeah, that's what I'm assuming too
But how would you measure it? Perhaps they subtract what they recon rest of the test-bed draws?
 
I believe they're generally the whole system
they keep the other parts constant, and measure consumption at the plug
 
I recall intel did per component power readings on their tablets
it was impressively messy'
 
2:05 PM
Anything other than "at the plug" must be at least more complicated than the average hardware reviewer can manage, and probably way too expensive to set up.
 
But I don't get how they get such low numbers. I should probably do some experiments of my own...
I've always been pushing my PSUs to the limit. And now finally I've managed to over-tax a 5 year old 650 watt PSU. It has endured two CPU and GPU upgrades, along with two more drives than originally planned.
I always wanted to try two-psu in one comp. It's far easier than I thought :) With additional 350 watts (offloading all hdds and fans), the workstation is now stable again after preliminary stress-tests. My life is complete once again. hahah
 
you'd just need to connect two or three wires
ground and sense, and power clean with another one, optionally
 
2:27 PM
That Debian in-place upgrade went well... screen was blank, no longer accepted input, I had to reset, now the whole system is broken :P
 
Bob
2:54 PM
@JourneymanGeek Would be a little difficult, though. CPU power draw can vary pretty wildly depending on usage.
@OliverSalzburg Sounds like every time I've tried to upgrade K/ubuntu so far :P
 
What's a "wandler", Oli
 
@Bob In that tomshardware article, they say that they saturate the CPU with Prime95 to work against that variation
@ЯрославРахматуллин In what context?
 
cpu, including wandler
 
Google says it means "converter"
No idea how that applies to CPUs
And no idea why the word in front of it would be in English :D
 
It is German. It says inklusive...
 
3:00 PM
It's not enough context :P
 
tauschen Energie zwischen einem System und der Umgebung in mindestens zwei Energieformen aus
converter
aka transformator
 
Bob
@OliverSalzburg Ah. Not sure how accurate that would be, but it might work :P
Heh.
I have so many windows open that the alt+tab switcher is showing 5 rows of 7 columns and has a second page
 
:D
Fresh installation of Wheezy is looking nice. Well, I guess, technically it's Gnome ;P
Iceweasel comes with AdBlock preinstalled
 
3:18 PM
nice
Wheezy ships without an OpenVZ kernel >_<
 
@allquixotic You mean it doesn't exist at all?
 
@OliverSalzburg right
 
@allquixotic Should it? :D
 
@OliverSalzburg Squeeze did.
 
Bob
@allquixotic well, isn't it up to OpenVZ to support newer kernel versions?
 
3:27 PM
@Bob yeah, but instead of supporting newer kernels, they're still supporting 2.6.32 atm, and mainlining more and more of their stuff so that you can kinda sorta get a working OpenVZ kernel on mainline if you enable all the cgroup stuff, but still no venet driver and a few other things missing
 
Bob
@allquixotic They're apparently moving to 3.5...
Also, BuyVM seems to be using 3.5 kernels on their Las Vegas nodes
root@vps-lv:~# uname -r
3.5.0-042stab076.7
Mid 2012 OpenVZ blog post:
> What's next? We will be rebasing OpenVZ to Linux Kernel 3.5 (most probably) and will try to re-use CRIU for checkpoint and restore of OpenVZ containers, effectively killing a huge chunk of out-of-tree kernel code that we have in OpenVZ kernel.
@allquixotic How would I check the venet driver?
 
how did they get a "stab" (stable) kernel that isn't published? >_>
 
Bob
The networking interfaces still come up as venet with ifconfig
@allquixotic I have no clue :P
Welcome to Ubuntu 12.10 (GNU/Linux 3.5.0-042stab076.7 i686)
 
maybe a BuyVM manager is an OpenVZ contributor
no fair :<
they need to release that shit
 
Bob
Could always ask them? :P
 
3:33 PM
w... wh... wait...
i686
dafuq
 
Bob
Yea. I chose a 32-bit template.
 
oh
i was like wait, all their servers have <= 4GB RAM?
 
Bob
XD
Nah, this is a tiny container so I wanted to minimise my memory footprint
o.O
looks like they've got Wheezy and Raring templates up already...
Ah.
@allquixotic Apparently it's not really 3.x
@SeaStarStar Forged by openvz alas :( We'll be pushing 2.6.32 back into beta soon.
Users wanting to dist-upgrade their debian 6 installs, be sure to reboot first so your kernel reports as 3.2.0 kernel!
 
Bob
4:10 PM
I've been wondering about that.
 
4:39 PM
sounds like a short sighted article
eventually servers will support ipv6
ipv6 isn't so necessary at the moment that's why people are still on ipv4
 
It sucks to be referee: uk.eurosport.yahoo.com/news/…
 
If ipv6 had been designed as an extension of ipv4, as that article suggests, then it'd probably be really messy and inefficient
the only mess i see that article mention, is servers not supporting ipv6, but they will eventually.
 
Bob
@barlop It's already really messy and inefficient.
 
@barlop I think ipv4 should become ipv6.
 
@Bob what do you mean?
@Boris_yo yeah sure just add two dude!
 
Bob
4:41 PM
The problem is the requirement of cutting over completely, which we've now got half a dozen different barely-working bridging techniques to try to alleviate.
 
@barlop Yeah my opinion as well.
@Bob You mean transition?
 
@Bob what requirement to cut over completely? it doesn't have to happen -now-.
with good programmers there's no reason why bridging techniques should barely work.
they should test their code!!
 
Bob
@barlop It's not broken code. It's the lack of good bridging servers.
Since not everyone can get IPv6 addresses.
 
@Bob and if they can't then software must support both ipv4 and ipv6. what's the problem?
software should have programmers that can deal with that.
 
Bob
@barlop Once again: not everyone can get IPv6 addresses. And if IPv4 ever really 'runs out', then you'll also have people who can't get IPv4 addresses.
 
4:44 PM
@Bob Does America have jobs for CCNA? And what if one knows computers and network but does not possess CCNA certification?
 
@bob i'm not saying there should be a complete transition now, so the fact that some can't get ipv6 addresses is irrelevant.
 
Bob
Software can support both IPv6 and IPv4. Great. But the nature of the problem, and what the article is complaining about, is the requirement to have a publicly visible IPv6 address and a publicly visible IPv4 address to work with both.
@barlop It's perfectly relevant.
Those who can't get IPv4 will have to use IPv6.
Which will be inaccessible to those who can't get IPv6.
 
@bob people will be able to get either ipv4 or ipv6, if not then they should look for another ISP!!
 
Bob
The problem is the requirement to have both if you want to access both.
@barlop That is stupid.
I'm in Australia. Maybe one ISP here has IPv6 support.
 
@bob you're telline me the reality is that some can't get ipv6, and in the future they may not be able to get ipv4. so i'm answering regarding that reality
there's no point calling it stupid, if that's the reality we're in that has to be accepted.
and it's not unworkable.
 
Bob
4:46 PM
And I'm telling you that you are relying too much on the user's ability to make informed choices.
And whether they even have the choice.
 
you're telling me they often don't have the choice..
'cos they're limited by what the ISP can give them
i'm not relying on them having a choice
 
Bob
I've had to set up a VPN connection to a VPS in the US, with all the glory of 300 ms latency, jsut to access IPv6 only sites.
 
ipv6 only sites, is stupid
 
Bob
At this point in time, such sites are more for testing. But in future, there will be more of them.
Many more.
 
well, that is stupid
but in the future it will be less stupid
 
Bob
4:48 PM
sigh
 
but ipv6 isn't the problem there. the problem is a programmer not making his server compatible with both ipv6 and ip4
 
Bob
....
 
What do you propose Bob if you consider it to be stupid?
 
Bob
No.
The problem, as described perfectly in that document, is that the programmer cannot make that decision.
 
the programmer obviously -can- make the decision for his server to support both
 
Bob
4:49 PM
Nope.
The standard says that IPv6 cannot communicate with IPv4.
 
oh ok. that is stupid
 
He can support but it does not mean it will work?
 
Bob
That means the server admin must acquire and set up an IPv6 address in order to communicate with IPv6. And an IPv4 address to communicate with IPv4.
 
that sounds like nonsense to me..
How can one protocol declare that it's not compatible with another?
 
No backward compatibility? Okay bad.
 
Bob
4:50 PM
So the server will need two addresses.
 
fine
and the server has two addresses then the server -is- compatible with both
 
Bob
@barlop Well, when you design the protocol to be independent, rather than an extension of the existing one...
 
but it is compatible
 
Bob
@barlop At the moment, that is the only ideal we can have.
But many server admins are lazy.
And IPv6 is a pain to set up, at times.
 
server admins now..
 
Bob
4:51 PM
I know that. I've tried.
 
so the programmers put the ability in.. and it's just incompetent admins.
well, the people paying the admins should pay more then
 
Bob
There's also the issue of many providers not offering IPv6.
 
so the server should support both. programmers and admins.
 
Bob
it's not easy to switch a massive site over to another provider.
best you can do is set up soem form of proxy
 
and that makes compatibility.
 
Bob
4:52 PM
which often cannot handle large loads because they never anticipate it
 
get a better proxy
 
Bob
...
 
this is 2013
 
Bob
It's not always a case of shouting 'do it better' at people.
Yes, this is 2013.
Your point being?
That document was from 2002. When IPv4 exhaustion was already well anticipated.
 
is the cost of the technology for the proxy so prohibitive?
 
Bob
4:53 PM
We have barely started transitioning some things to IPv6.
The vast majority of servers are still not IPv6-capable.
 
"640K is more memory than anyone will ever need on a computer," Gates
"ipv4 addresses more range than anyone will ever need" ipv4 founder
 
sounds liek a programming/admin issue. not a technology issue. 2013 technology is fine i would think. The cost of not doing it better might push people to do it better.
 
Bob
@barlop You try convincing management to increase your budget.
 
you don't have to.
 
Bob
Ify ou want a better proxying server, yes you do.
 
4:54 PM
^^^
 
Bob
If you want to move massive amounts of infrastructure, yes you do.
If you want to set up an entire new networking interface, yes you do.
All of thisi s a lot of work, especially for bigger projects.
 
you don't have to nag them to do what's to their benefit. just tell them what their options are. what happens if they don't.
 
Bob
It's idealistic to drop all backwards compatibility and start with something clean.
In the real world, it's not successful.
The transition is a pain.
 
1
Q: Vector graphics formats: SVG , EPS or PDF

AIBHow does the vector graphics formats svg , eps and pdf compare against each other(Quality, filesize etc)? If I can generate(eg matplotlib) plot in any of the above formats, which format should I choose for output? (Optional) Assume that the output graph is to be published using Tex. I know th...

Diamond studded closure
 
@bob you have compatibility via a proxy. It's idealistic to wish that doing that isn't work, but it is work. And it's adequate as a solution while both are about.
 
Bob
4:59 PM
@barlop Which relies on an additional server, and a lot of setup work.
 
@Bob ok and when ipv4 dies out completely people may be glad they're not lumbered with an ipv4 extension. I've heard ipv6 is easier to subnet.
 
Bob
@barlop Lumbered? In what way?
The thing is, we are getting close to the point where new IPv4 addresses will be hard to find/rather expensive.
 
well, I don't know much re the advantages of ipv6, but there may be advantages.
 
Bob
When that happens, since we don't have some clean transition set up, peopel won't be able to easily acquire new addresses.
Then they're stuck.
 
they can set up a transition though, proxy.
it's going to become more standard to set up such a proxy. and easier to do
 
Bob
5:01 PM
@barlop Fact of the matter is, the IPv6 address space is big enough that a standard subnet is 2^32 times bigger than the entire IPv4 address space.
You could easily set aside a subnet to contain the IPv4 space (actually, one RFC did that, though only for use within private networks...)
 
well, they might be splitting the subnets smaller than that size then.
so there's still going to be a need for subnetting with ipv6.
 
Bob
I'm not saying there isnt going to be a need.
 
like for security.. And subnetting may be easier with ipv6.
that may be an advantage with ipv6.
 
Bob
I'm saying it's very easy to contain the IPv4 address space within IPv6, and therefore allow IPv6 to directly route to IPv4 addresses with IPv6 capable software
 
fine so it's very easy.
so making the compatibility is less of a difficulty then.
 
Bob
5:04 PM
that way, anyone with an IPv6 address can still communicate with those on IPv4 with only a software upgrade required, not relying on acquiring and setting up IPv6 addresses
of course, that may result in added software complexity
 
@bob I was copying something with teracopy yesterday.. and it said something about timestamps being invalid
have you ever heard of that?
 
Bob
:\
never used teracopy
just never bothered
 
ever heard of a message like that from any copying program?
 
Bob
uhh... no?
I only use the Explorer builtin and s=the command-line builtins
never heard of anything to do with timestamps
 
do you ever do a copy where you copy from source to dest e.g. doing a backup, copying only the files that are modified at the source with a later date/time?
in windows or linux?
 
Bob
5:10 PM
no
I use a proper backup program on the desktop
and just replace all otherwise
 
what proper backup program do you use?
 
Wow eBay now displays never ending feed of items. On their homepage. Innovating...
 
Bob
@barlop Acronis True Image
 
@Bob How much you paid for it? Was it on sale?
 
5:26 PM
@bob do you use Acronis true image for imaging only, or for file copying backup as well?
 
Bob
@barlop One full image, daily incremental backups of certain directories rotated weekly
 
 
1 hour later…
6:36 PM
weird sound issues somehow related to USB devices being (un)plugged... i sense a reboot soon
 
 
2 hours later…
8:21 PM
Please try to limit yourself to one question per question. Feel free to open several questions for every question you have. — Oliver Salzburg 4 secs ago
I love comments like that :D
 
@OliverSalzburg open several questions for every questions you have?
:D
 
I'd be happy to learn about a better phrase to use in cases like that ;D
 
"Please open up a new question for each distinct question you have."
 
@bob are you sure it's not a differential backup that you do(rather than incremental)?
incremental seems to be doing little mini backups which then need to be combined. "when you want to carry out a complete restore, the most recent full backup and all of the subsequent incremental copies must be restored."
do you really do that? i'd choose differential backup over that.
differential backup "Differential backups copy those files that have been changed since the last full backup took place. ".
 
 
1 hour later…
9:43 PM
I wasn't aware that this has been going on for that long.
 
 
1 hour later…
10:49 PM
One question per post please, and keep in mind that there is no post limit!
swoosh -- 3 points
 
11:07 PM
@ЯрославРахматуллин Nice! Thanks
 
holy sheet, calculating the scalar product between two points in skyrim gives you the distance between them! I was expecting it to not work due to interior and exterior cells. If I can get the distance that way, why can't ObjectReference.getDistance() do the same?
 
@OliverSalzburg You'll love geekwagon.net/projects/xkcd1190
4
 
@TomWijsman Neat, but I've seen it now ;D
The slider is pretty hot though :D
 
Yeah, kind of has all the features you can imagine I think.
 
11:22 PM
It's pretty hot alright
 
Wonder what's going to happen next, the situation feels close to the first frame
whereas the difference is that they have that bag now.
 
I hadn't watched it since I read about it again on Slashdot earlier
 
I check it every other day, so huge jumps in progress.
 
what an odd narrative...
 
Bob
11:37 PM
@barlop Since I do a new full backup every week, yes, I am sure it is incremental.
Most of the time there are no changes day to day anyway
or only very minor changes
 

« first day (1003 days earlier)      last day (4016 days later) »