What would be truly a shoe-in for a nobel would be if you could engineer
...a hybrid of crawling ivy, a maple tree, and a venus fly trap... which crawls around the forest eating other plants, then gets tapped for its gasolicious sap every March.
mm-MM! Let's go to the sugar bush. And kids -- stay away from the trees with teeth!
The DC rolls up the cost for AC and everything else (except bandwidth, which is cheap-ish) into the power bill. But yes, that's pretty pricey. I think they are trying to convince people that don't really need to be downtown to move to the other location where everything is cheaper.
Of course, I'm grandfathered into for most of my equipment, so I only have a few hundred VA on the new price plan.
But, geez, it really changes the board room dynamic: "Should we get another server in the rack? It'll cost WHAT a month? Maybe we'll just light up a zone on a lightly loaded machine..."
Around here, you can't get GPRS without Edge, so I have no idea what the display says in GPRS-only mode.
And 110K per second -- what I've clocked my Edge connection at -- DOES feel like dialup. Sure, it may be faster on paper, but waiting 10 seconds for a google map to load is crazy slow to me. Hence the "roughly dialup" comment.
Of course, I could be time dialated, the last time I actually used a dialup modem for 'net access was 1999 or so.
Look in the top left-hand corner(ish). If it says "3G", you have a 3G connection. If it says "E", you have Edge. Edge is roughly dialup. It works, but it's slower'n'piss.
I have a Rogers iPhone, I live in a rural area, and get 3G on one side of my house and Edge on the other. The difference is astounding.
...which comes right back to TFQ: the submitter is running a high-performance scientific workstation.
Personally, I think he should put through a budget request for the most expensive computer he can find and say he needs it to keep working at his current pace after WDE is installed.
You also have the problem of rabid freaks on usenet droning on and on about Cassinni and how all life on earth will cease to exist if something goes wrong during the launch.
Personally, I think we should be mining plutonium on the moon and doing final assembly on-site. With robots that look like ants.
> There is nothing fundamentally wrong with the Javascript > language, like there is in visual basic.
Oh, sure there is.. they're not show stoppers.
Pick up Doug Crockford's new book, "JavaScript: The Good Parts" (O'Reilly) and read the "Bad Parts" appendix. I agree with most of his points. Best $30 I've spent in a while, and I've been writing JS code (in browser and out) for over a decade.
The big one being automatic semicolon insertion. BAD! BAD!
!== vs != is another. != should have been !==, !== left out, and so on.
I don't know about you, but in 1988 I was still using a 300 baud modem with no autodialer. 1200 bps was a fantasy that finally came true in 1989 for the low, low price of $249. 2400 bps (or better) arrived for the masses in early 1992.
I don't know about Wintel (or even Linux/x86), but there is no reason to limit yourself to 4GB of VM in a 32-bit environment. You only need to limit yourself to 4GB of address space per process.
Segmented addressing (all the vogue on the 80286 -- is it still on x86?) allowed 20-bit addresses in 16-bit address space.
I have even seen OLD (10-12 years old) Sun machines which could take 5GB of RAM but only had 32-bit sparc CPUs.
> Now, you cannot seriously consider abandoning virtual memory and all that comes with it > (inter-process protection, kernel protection from user-space errors amongst others), can you?
Can you imagine the hit you'd take on fork() without VM? You'd have to re-write every single pointer in the data segment.
Which would mean you'd need to _know_ what was a pointer and what wasn't.
There's other reasons for big swap, on Solaris though. (I don't know about other OSes, don't use 'em much).
One, which recently bit me in the foot, has to do with forking. The system basically pre-allocates swap against process space size (vm size, not rss), even though the pages may not actually get physically allocated. This is because the kernel wants to make sure that when you want to write to the memory, it's going to be able to allocate pages in the VM for you to write to -- remember, we're forking so we're doing copy-on-write.
I bumped into this a few months ago retrofitting a bind-listen-accept-fork daemon to make use of a 200+ MB data structure which was populated into the parent and read in the children. Even though I had plenty of swap "free", the OS had enough of it bookmarked for use in case I wrote to it that it refused to fork all the children I needed.
For the curious, the solution to my particular dilemma was to write my own [trivial] allocator backed by mmap; IIRC I used MAP_NORESERVE | MAP_SHARED to convice Solaris not set aside excessive amounts of swap [excessive because I had knowledge the OS didn't] and allow my program to fork enough times to handle the requisite load.
Slashdot Mirror
mm-MM! Let's go to the sugar bush. And kids -- stay away from the trees with teeth!
The DC rolls up the cost for AC and everything else (except bandwidth, which is cheap-ish) into the power bill. But yes, that's pretty pricey. I think they are trying to convince people that don't really need to be downtown to move to the other location where everything is cheaper.
Of course, I'm grandfathered into for most of my equipment, so I only have a few hundred VA on the new price plan.
But, geez, it really changes the board room dynamic: "Should we get another server in the rack? It'll cost WHAT a month? Maybe we'll just light up a zone on a lightly loaded machine..."
In this case, yes, because all my power supplies have a power factor very close to 1.0.
Man, I gotta get me one of those UPSes!
Preferably one that only draws 2-3 VA.
Then I could plug all my servers into it, and save a BOATLOAD of money!
(My datacenter charges about a buck a VA)
I'd like to see them try with Medeco bi-axial keys
Did you just imply that senior T-Mobile engineers can't even afford their own apartments?
Around here, you can't get GPRS without Edge, so I have no idea what the display says in GPRS-only mode.
And 110K per second -- what I've clocked my Edge connection at -- DOES feel like dialup. Sure, it may be faster on paper, but waiting 10 seconds for a google map to load is crazy slow to me. Hence the "roughly dialup" comment.
Of course, I could be time dialated, the last time I actually used a dialup modem for 'net access was 1999 or so.
Look in the top left-hand corner(ish). If it says "3G", you have a 3G connection. If it says "E", you have Edge. Edge is roughly dialup. It works, but it's slower'n'piss.
I have a Rogers iPhone, I live in a rural area, and get 3G on one side of my house and Edge on the other. The difference is astounding.
You forget three things:
1. Guitar Hero builds/uses precisely the mental pathways as are used by real musicians reading music.
2. Guitar Hero teaches meter. And there are PLENTY of non-musicians out there (and even wannabees with a borrowed amp and an old Strat) without meter.
3. Guitar Hero is fun
...which comes right back to TFQ: the submitter is running a high-performance scientific workstation.
Personally, I think he should put through a budget request for the most expensive computer he can find and say he needs it to keep working at his current pace after WDE is installed.
And Bob Dole could get its antenna to deploy -- and function -- even if it was on the brink of death
You also have the problem of rabid freaks on usenet droning on and on about Cassinni and how all life on earth will cease to exist if something goes wrong during the launch.
Personally, I think we should be mining plutonium on the moon and doing final assembly on-site. With robots that look like ants.
Funnily enough, that's not true in SpiderMonkey -- IIRC -- ints less than 2^28 are stored as tagged longs. More like Visual Basic than VIC-20 BASIC. :)
But that's just an implementation detail and doesn't address the real (no pun intended) problem you raised.
> There is nothing fundamentally wrong with the Javascript
> language, like there is in visual basic.
Oh, sure there is.. they're not show stoppers.
Pick up Doug Crockford's new book, "JavaScript: The Good Parts" (O'Reilly) and read the "Bad Parts" appendix. I agree with most of his points. Best $30 I've spent in a while, and I've been writing JS code (in browser and out) for over a decade.
The big one being automatic semicolon insertion. BAD! BAD!
!== vs != is another. != should have been !==, !== left out, and so on.
I sure hope he has the ringer turned right off, or he's gonna go deaf if someone calls!
Send it to your hotel DHL overnight before you leave, and do the same to get it home.
Problem solved.
> Your law school is teaching you some strange maths...
He's probably majoring in IP and Copyright and using the new math they taught him in Damage Assessment 101.
I prefer Panaphonics.
It wasn't thinly veiled at all. Yeesh.
This is basically astroturfing.
No, wait. *coins new term*
This is bashtroturfing !
Wow, dude, relax.
Or, at least.. stick your head in a microwave and give yourself a tan.
I don't know about you, but in 1988 I was still using a 300 baud modem with no autodialer. 1200 bps was a fantasy that finally came true in 1989 for the low, low price of $249. 2400 bps (or better) arrived for the masses in early 1992.
That man page vanished sometime in the mid nineties.
I don't know about Wintel (or even Linux/x86), but there is no reason to limit yourself to 4GB of VM in a 32-bit environment. You only need to limit yourself to 4GB of address space per process.
Segmented addressing (all the vogue on the 80286 -- is it still on x86?) allowed 20-bit addresses in 16-bit address space.
I have even seen OLD (10-12 years old) Sun machines which could take 5GB of RAM but only had 32-bit sparc CPUs.
> Now, you cannot seriously consider abandoning virtual memory and all that comes with it
> (inter-process protection, kernel protection from user-space errors amongst others), can you?
Can you imagine the hit you'd take on fork() without VM? You'd have to re-write every single pointer in the data segment.
Which would mean you'd need to _know_ what was a pointer and what wasn't.
*shudder*
There's other reasons for big swap, on Solaris though. (I don't know about other OSes, don't use 'em much).
One, which recently bit me in the foot, has to do with forking. The system basically pre-allocates swap against process space size (vm size, not rss), even though the pages may not actually get physically allocated. This is because the kernel wants to make sure that when you want to write to the memory, it's going to be able to allocate pages in the VM for you to write to -- remember, we're forking so we're doing copy-on-write.
I bumped into this a few months ago retrofitting a bind-listen-accept-fork daemon to make use of a 200+ MB data structure which was populated into the parent and read in the children. Even though I had plenty of swap "free", the OS had enough of it bookmarked for use in case I wrote to it that it refused to fork all the children I needed.
For the curious, the solution to my particular dilemma was to write my own [trivial] allocator backed by mmap; IIRC I used MAP_NORESERVE | MAP_SHARED to convice Solaris not set aside excessive amounts of swap [excessive because I had knowledge the OS didn't] and allow my program to fork enough times to handle the requisite load.