I couldn't imagine buying flowers, or gift baskets. I couldn't imagine buying sunglasses. I'm one of those people who has to simply buy some things in person.
It's the delivery bit. A while ago my wife ordered flowers online for her grandma's funeral, which we could not attend -- it was in another country. The only alternative would have been to ask someone else to buy the flowers for us.
How do you budget with GNUcash? I thought that was something it didn't do?
With scheduled transactions (automatically inserted a month in advance): I've scheduled all the regular payments as accurately possible, and then I've scheduled a weekly '(various)', for GBP xx.00. At the end of each week, after I've input the actual expenses, I delete that '(various)', of course.
That gives me a reasonable estimate of how much is left at the end of the month.
My 1986 International Physics Olympiad CD is still ok -- I listened to it yesterday! That was a junket from the English team, sourced from Philips/Dupont, and labeled as a CD-ROM.
It turns out I was lucky: I got Beethoven's Ninth Symphony, others got CD(-ROM)s containing just junk. But it took some 5 years before I could first listen to it -- no CD player yet!
I tend to skip all articles about Google on The Register, because they are all written by Andrew Orlowski -- who seems to have a personal vendetta against Google. I suspect it doesn't rate his personal website high enough:-).
To put what Gelernter is talking about in more clear terms, ask this question; "what if a network connected computer could somehow be the shape of a newspaper?" You could turn its pages, like an old newspaper, but its actual scope would be the size of the internet.
...Continued on page 32323243243...94?
That would be a lot of page turning!
And you know what? I think an XML v1.1 document would be incompatible with any non-updated program, no matter what the changes in v1.1 are -- for if the program wasn't upgraded, it can't know what XML v.1.1 means. And there must be some difference, otherwise it wouldn't have a different version number
They're right, Gecko is pretty nice, but big deal. AoL is quite a large part of USA based ISP's but do you think EVERYBODY who has deviated from the de-facto standard of IE is going to change thier website so a bunch of AoL idiots can read it?
There are a lot of 'AOL idiots':
on my website, www.accordionlinks.com, AOL provides
a few percent of the audience, four times more
than number two.
AoL's hated for a reason. Thier service doesnt require the brain power needed to use other ISP's. AoL users are generally regarded as morons.
If they are morons, it might be
easier to part them from their cash, so that might make them a good target audience for some...
> printf("%d = %d\n", (int) (((60/6)*0.3) + (10*0.7)), (int) ((( a/b)*0.3) + ( c*0.7)));
[..]
On all Linux distros [..] I get 9=10.
Repeat after me: Floating point calculations are
imprecise. It seems like you've got it the wrong way around, by the way: the calculation that is
done at compile time is "exact".
You can simplify the above further to int c = 10; ... (int) (3 + c * 0.7)
which gives you 9 -- because 0.7 happens to be
rounded down in binary! (60/6) is an integer
calculation, so that gives you exactly 10 --
but with floating point calculations you get rounding errors.
That the first expression, which is calculated by the compiler does result in 10, is because
that is done by the
optimised compiler, which calculates the whole
expression in the FPU without storing intermediate
results. And the i386 FPU uses 80-bit numbers
internally, so that it does work out OK in this
case. Or perhaps you're just plain lucky.
If you want the answer to be 10, just change
int c = 10; to double c=10.0000000000001;.
When it comes to numerical mathematics,
lesson one is: You cannot trust the rounding
of floating point calculations.
So is there anything new in this book?
Yes, to me, anyway: for instance, how Alan Cox
got to be the maintainer of the networking
code in kernel 0.96xx instead of Fred van Kempen
(wasn't he the guy who owned the linux.com domain?)
This is fantastic! If I read and understand this correctly, they have 100km fiber runs _without_ a repeater! That's truly excellent. Most of the cost for long distance runs [after the right-of-way] is in the repeaters and powering them, not in the mediaAnd they can run it at 40 GHz. That's 4 THz/km. Normally, fiber is limited by "smearing" over long lengths--the light pulses get spread out over the length of the media.
(you mean THz * km: longer distances are more impressive). 40 Gbit/s per channel with 100 km repeater spans is not bad, but it is definitely not why they got postdeadline paper in OFC 2000 (Optical Fiber Communications Conference, paper PD23)! There is another OFC postdeadline paper about "320 Gbit/s single channel pseudo-linear transmission over 200 km of non-zero dispersion fiber" (also from Bell Labs/Lucent Technologies) -- there the amplifier spacing is 100 km as well. So that's 64 Tbit/s*km. Probably, the reason that that paper had two amplifier spans rather than three was that they could not get it working for three amplfier spans:-).
Once you allow for amplifiers (which are more expensive than just fibre, but not too expensive), you can basically go unlimited distances, e.g. 10,000 km at 10 Gbit/s is almost run of the mill. And as long as it is just amplifiers, the telecomms companies don't mind putting in a few more, with a slightly smaller amplifier spacing, if that will make the system more robust. What they really hate are regenerators, where you demultiplex the signal, put it through a receiver, (probably do error correction), and then retransmit it again: that is horribly expensive. But amplifiers are relatively cheap.
The smearing you are talking about is not a major problem anymore, with properly designed systems (and fibres): with dispersion management and dispersion compensation the net linear dispersion can be brought down to zero.
There's a new reason to start reading comments at 0 and -1...when there are funny trolling posts like this one that slip through the moderators. People see something that is anti-rob anti-slashdot or anti-"linux party line" and they moderate it down.
But this is a high-quality troll. Has Jesus Christ ever checked out segfault.org? You'd fit in over there
It is a nice troll, but do we need it three times? I moderated exactly the same posting down in the Ebay may bid for Sotheby's thread, as "Off-Topic", (its current rating is +3 there), and it first appeared on the Linux 2.3.48 thread yesterday (where its current rating is +4). So perhaps I should have moderated it as "Redundant" instead?
Perhaps he is trying to burn off his Karma by repeating himself...
And Slashdot is rune more or less independently anyway. Come to think of it, I can't figure out what Andover ever got out of the deal, besides being able to say that they'd bought Slashdot.
Would AV have bought Andover if Andover had not bought Slashdot?
But I have a hard time understanding why Mozilla takes so long to take off when something like Linux or Free BSD, whose source code size is probably bigger than Mozilla's (correct me if I'm wrong), are still very active. It's definitely not the lack of people, or the lack of skills, but I'd like to understand where it comes from.
You are wrong: kernel-source-2.2.14.tar.gz is 16 MB, kernel-source-2.3.39.tar.gz is 17 MB, mozilla-source.tar.gz is 20 MB. So Mozilla is bigger, and it could be argued that it is more complicated: it is cross-platform, has to work with different GUI and system libraries, and has lots of interaction with the user, etc. Ok, the kernel is cross-platform as well, but in a sense its behaviour is simpler, and more modular -- want to add a new device or file system? Write a module that implements open(), close() etc. and you are done. And you don't have to interface with fickle users directly, just with the C library:-)
Besides, how long did it take Linux to come to a usable state? I did not start with version 0.3xx, I started at 1.2.13, by which time Linux was probably five years old or so. Mozilla will be ready in less time than that, I guess.
Better still: get The Annotated Alice - Alice's Adventures in Wonderland and Through the Looking Glas, with annotations by Martin Gardner.
Jeroen Nijhof
Re:(response, getting slightly OT and onto USB)
on
SuSE Coming on DVD
·
· Score: 1
So, what's the solution? #define UNINFORMED_SPECULATION I guess Linus is either working on a modified devfs or an alternative scheme that will satisfy the naysayers. #endif
Or he is waiting for someone else to come up with the "right" solution, as he did with the big memory on Intel.
> It drives me nuts that I can't cut and paste from emacs to Netscape.
Select in emacs and middle click in Netscape works, but if it's URLs you want to copy for jumping to, there's something much better: browse-url-at-point. It will make netscape display the URL at the cursor position (using netscape -remote).
I found out about it because it was bound to C-c C-z u in html-helper-mode, but I found it so useful that I have bound it to C-c u globally -- it's pretty useful in mail mode as well!
browse-url-at-point is part of the standard elisp libraries, for (gnu) emacs 20.3, at least; html-helper-mode is not (although it might be for xemacs), see http://www.santafe.edu/~nelson/tools/ for that.
Slashdot Mirror
It's the delivery bit. A while ago my wife ordered flowers online for her grandma's funeral, which we could not attend -- it was in another country. The only alternative would have been to ask someone else to buy the flowers for us.
With scheduled transactions (automatically inserted a month in advance): I've scheduled all the regular payments as accurately possible, and then I've scheduled a weekly '(various)', for GBP xx.00. At the end of each week, after I've input the actual expenses, I delete that '(various)', of course. That gives me a reasonable estimate of how much is left at the end of the month.
Jeroen
My 1986 International Physics Olympiad CD is still ok -- I listened to it yesterday! That was a junket from the English team, sourced from Philips/Dupont, and labeled as a CD-ROM.
It turns out I was lucky: I got Beethoven's Ninth Symphony, others got CD(-ROM)s containing just junk. But it took some 5 years before I could first listen to it -- no CD player yet!
Jeroen
I tend to skip all articles about Google on The Register, because they are all written by Andrew Orlowski -- who seems to have a personal vendetta against Google. I suspect it doesn't rate his personal website high enough :-).
Because then they would have to pay out on IBM's, RedHat's and everyone else's counter suits?
In the Linux version, that feature is available, but slightly differently: just middle-click (as in 'paste') anywhere in the browser window!
Jeroen Nijhof
Shoudn't this be filed in the humour section?
There is only a rant on that page, no examples.
And you know what? I think an XML v1.1 document would be incompatible with any non-updated program, no matter what the changes in v1.1 are -- for if the program wasn't upgraded, it can't know what XML v.1.1 means. And there must be some difference, otherwise it wouldn't have a different version number
Jeroen
Could <head explodes> be a legal rule invocation? Rule with an argument?
Jeroen Nijhof
There are a lot of 'AOL idiots': on my website, www.accordionlinks.com, AOL provides a few percent of the audience, four times more than number two.
AoL's hated for a reason. Thier service doesnt require the brain power needed to use other ISP's. AoL users are generally regarded as morons.
If they are morons, it might be easier to part them from their cash, so that might make them a good target audience for some...
What, and not include the linux kernel?
Jeroen Nijhof
From the DOJ press release,
DOJ press release:
Mr. Sklyarov will be prohibited from violating any laws during the year
...
Jeroen
[..] On all Linux distros [..] I get 9=10.
Repeat after me: Floating point calculations are imprecise. It seems like you've got it the wrong way around, by the way: the calculation that is done at compile time is "exact".
You can simplify the above further to
... (int) (3 + c * 0.7)
int c = 10;
which gives you 9 -- because 0.7 happens to be rounded down in binary! (60/6) is an integer calculation, so that gives you exactly 10 -- but with floating point calculations you get rounding errors.
That the first expression, which is calculated by the compiler does result in 10, is because that is done by the optimised compiler, which calculates the whole expression in the FPU without storing intermediate results. And the i386 FPU uses 80-bit numbers internally, so that it does work out OK in this case. Or perhaps you're just plain lucky.
If you want the answer to be 10, just change int c = 10; to double c=10.0000000000001;.
When it comes to numerical mathematics, lesson one is: You cannot trust the rounding of floating point calculations.
Jeroen Nijhof
Jeroen
> I don't see anything "extreme" about it
The hype, perhaps
Jeroen Nijhof
(you mean THz * km: longer distances are more impressive). 40 Gbit/s per channel with 100 km repeater spans is not bad, but it is definitely not why they got postdeadline paper in OFC 2000 (Optical Fiber Communications Conference, paper PD23)! There is another OFC postdeadline paper about "320 Gbit/s single channel pseudo-linear transmission over 200 km of non-zero dispersion fiber" (also from Bell Labs/Lucent Technologies) -- there the amplifier spacing is 100 km as well. So that's 64 Tbit/s*km. Probably, the reason that that paper had two amplifier spans rather than three was that they could not get it working for three amplfier spans :-).
Once you allow for amplifiers (which are more expensive than just fibre, but not too expensive), you can basically go unlimited distances, e.g. 10,000 km at 10 Gbit/s is almost run of the mill. And as long as it is just amplifiers, the telecomms companies don't mind putting in a few more, with a slightly smaller amplifier spacing, if that will make the system more robust. What they really hate are regenerators, where you demultiplex the signal, put it through a receiver, (probably do error correction), and then retransmit it again: that is horribly expensive. But amplifiers are relatively cheap.
The smearing you are talking about is not a major problem anymore, with properly designed systems (and fibres): with dispersion management and dispersion compensation the net linear dispersion can be brought down to zero.
Jeroen Nijhof
But this is a high-quality troll. Has Jesus Christ ever checked out segfault.org? You'd fit in over there
It is a nice troll, but do we need it three times? I moderated exactly the same posting down in the Ebay may bid for Sotheby's thread, as "Off-Topic", (its current rating is +3 there), and it first appeared on the Linux 2.3.48 thread yesterday (where its current rating is +4). So perhaps I should have moderated it as "Redundant" instead?
Perhaps he is trying to burn off his Karma by repeating himself...
Jeroen Nijhof
As in Please use a mirror?
Jeroen Nijhof
Would AV have bought Andover if Andover had not bought Slashdot?
Jeroen
You are wrong: kernel-source-2.2.14.tar.gz is 16 MB, kernel-source-2.3.39.tar.gz is 17 MB, mozilla-source.tar.gz is 20 MB. So Mozilla is bigger, and it could be argued that it is more complicated: it is cross-platform, has to work with different GUI and system libraries, and has lots of interaction with the user, etc. Ok, the kernel is cross-platform as well, but in a sense its behaviour is simpler, and more modular -- want to add a new device or file system? Write a module that implements open(), close() etc. and you are done. And you don't have to interface with fickle users directly, just with the C library :-)
Besides, how long did it take Linux to come to a usable state? I did not start with version 0.3xx, I started at 1.2.13, by which time Linux was probably five years old or so. Mozilla will be ready in less time than that, I guess.
Jeroen
Jeroen Nijhof
#define UNINFORMED_SPECULATION
I guess Linus is either working on a modified devfs or an alternative scheme that will satisfy the naysayers.
#endif
Or he is waiting for someone else to come up with the "right" solution, as he did with the big memory on Intel.
Jeroen Nijhof
Not necessarily. One of my favourite verses is 1 Thessalonians 5:21: Prove all things; hold fast that which is good.
Jeroen Nijhof
> It drives me nuts that I can't cut and paste from emacs to Netscape.
Select in emacs and middle click in Netscape works, but if it's URLs you want to copy for jumping to, there's something much better: browse-url-at-point.
It will make netscape display the URL at the cursor position (using netscape -remote).
I found out about it because it was bound to C-c C-z u in html-helper-mode, but I found it so useful that I have bound it to C-c u globally -- it's pretty useful in mail mode as well!
browse-url-at-point is part of the standard elisp libraries, for (gnu) emacs 20.3, at least; html-helper-mode is not (although it might be for xemacs), see http://www.santafe.edu/~nelson/tools/ for that.
Jeroen Nijhof