Here's a better translation, albeit only based on a bit of high-school German and a small dictionary. E&OE.
The rumour on the grapevine is that the secretive processor manufacturer Transmeta will finally reveal the date for the introduction of their long-awaited processor on the first day of Comdex, 15th November (as already announced by Transmeta employee Linus Torvalds). But c't has already heard that the new processor will be launched under the name "Crusoe" on 19th January 2000 (a Wednesday, not a "Friday", by the way). Owing to its very low current consumption, Crusoe will be aimed primarily at the notebook market.
These distributed projects are all very well, but can one have too much of a good thing? How many people will take part in each one, I wonder, if they become commonplace?
I must say, I find their web site thoroughly unconvincing. I mean, all their screenshots are only of flat windows from awkward angles. Even their descriptions don't explain why they're doing it, what advantages a 3D window manager has over a 2D one.
Maybe this is the way of the future. But they haven't convinced me of that yet.
I've just rewritten the CGI interface for my program analog in Perl -- it was in C before. Indeed I even learnt Perl to do it. So let me talk about your question in terms of my example.
I guess the question is, what sort of application are you talking about? Analog itself has to be blindingly fast, so C is the obvious choice. But its CGI interface isn't of that type. Speed -- or efficiency if you prefer -- really isn't an issue. So C loses one of its big advantages. I guess a lot of CGI applications are of this type.
Instead, it has to take some small amount of user input, munge it around, and spit it out again. That's exactly the sort of thing that Perl is perfect for. Converting my application from C to Perl, I saved an enormous amount of code length. Again, I guess this is typical of many CGI applications.
Of course, one has to be very careful about security in CGI applications. I think it's swings and roundabouts here. Both languages have similar theoretical problems. But some dangerous things are easy in Perl, and thus probably get done more readily. (E.g. backticks instead of system(), open("|...") instead of popen()). This is a worry: but weighed against this is Perl's -T mode, and its ability to check input for safety very quickly with regexps. In either language you have to know what you're doing here and be very careful to apply it!
And finally, you shouldn't underestimate Perl's readability. Whether in an open source or a closed source model, one day other people will want to maintain your code. Readability is good.
In summary, I'm far from a Perl expert, having written exactly one thing in it. But for my application, and I suspect for the majority of CGI applications, efficiency wasn't the key constraint. Ease of writing and maintaining was.
As I understand it, "By Appointment" is only available if a member of the royal family uses the product personally and for some period of time. Using it personally presumably wouldn't be applicable here.
I guess the problem with a program like this is what one would do with the false positives.
Suppose you could identify a set of personal characteristics and say "90% of killers are in this category." This is unlikely: but even then it still wouldn't follow that most people in that category are killers.
So what do you do? Unless a reasonable proportion of the positives are true positives, the program isn't really very useful. The danger would be that the school would target all the positives to avoid litigation, even though most positive kids were quite safe.
Here's an idea I've wondered about. Maybe someone can tell me if it's sensible.
Wouldn't it be possible for my browser to have an option to fetch inline images in a page only if the image was on the same server as the page? That way I could see the useful content, but discard all banner ads. Or would I discard too much useful stuff too? Comments?
It's not even November yet and already we're starting to get stories of Christmas? Ugh.
I love Christmas the family time, but I hate Christmas the overcommercialized holiday. I could probably attempt to write a lengthy diatribe about how America is too commercialized...
At least in the US you have one advantage over us in the UK: Thanksgiving. With a major holiday at the end of November, the Christmas season doesn't really start until after that. In Canada, it starts after Hallowe'en (Thanksgiving being at the beginning of October there). In the UK, we don't take that much notice of Hallowe'en (yet), and Christmas seems to start in the middle of October, and getting earlier every year...
It's Jonathan Swift (author of Gulliver's Travels), not Thomas. The text can be read here and here and here (probably more).
Swift argues that babies could be a delicacy for the upper classes, and a source of revenue, instead of a resource drain, on the working classes. Sounds morbid, but it's quite amusingly done too.
And it is relevant to this debate. Swift too was trying to argue against contemporary attitudes which counted certain people as worthless, although in his case it was the poor rather than the disabled.
You can read a short introduction to the proposal here.
There's no free lunch (aka Metered calls are good)
on
ISP War in the UK
·
· Score: 1
The popular debate about the price of internet access misses one crucial point: none of it is free. The ISPs who don't charge a fixed charge are usually described as "free". At the same time people want "free" local calls.
Neither of them is free, in different ways. The first has no fixed charge but is metered through the telephone bill. The second would be unmetered but presumably have a higher standing charge.
Now, you can make a case that telecoms are overcharging. But that is an entirely separate issue from whether calls should be metered or unmetered. And given a choice between everyone being metered or everyone being unmetered, I'd go for metered every time.
Bandwidth is a scarce resource. By charging even a small amount for it you
limit demand; and
provide an incentive for service providers to upgrade their links, because if they can carry more traffic they immediately generate more revenue.
I don't think it needs a high charge. I would guess that if the charge is metered instead of unmetered, the total charge would be lower for almost everybody, and the network would be less clogged up so that everyone could get their work, or play, done faster.
Oh, and hasn't anyone noticed that this model is popular with customers? It's essentially the model that Freeserve and their 100 imitators use, that has brought the internet to 100,000 new Britons a week for the last year.
3com may not know it, but they've just stopped me from buying someone else's network card.
Write, call, tell them somehow! And anyone else in the same position. Seriously, if people do this, 3Com will continue with the same strategy. Other companies might well find out about it as well, and do the same. But if people don't give positive feedback for a decision like this, well, who knows?
In England, where there is a government censorship board which can potentially cut any movie, Eyes Wide Shut is opening tonight completely unaltered.
What's the difference? In the case of government censorship, at least the people doing the censorship have some accountability -- they can be voted out of office...
No, we in the UK can't vote our censors out of office. In general we only vote for our Members of Parliament, no other public officials. And most of our officials don't even change when a new party is elected. They're meant to be outside the political fray.
There must be some much deeper cultural reason for our showing films like this uncut, although I'm not sure what it is. I'm always confused about whether the USA is a very liberal or a very conservative country!
I get e-mails telling me that my program analog isn't Y2K compliant because I use (y % 4 == 0) as my test for leap years. They tell me it should be something like
(y % 4 == 0 && (y % 100 != 0 || y % 400 == 0))
I reply explaining that analog isn't Y2100 compliant. It's got so bad I've had to put an explanation in the source.
This article seems to make the usual confusion between the distribution and the kernel. Most people are working on a distribution or on a piece of software, rather than on the kernel itself.
If the Hurd microkernel is better, I hope I'll still be able to use the Debian distribution and the Apache web server: or whatever's better than them by then. Probably you won't be able to see easily which kernel I'm using anyway.
The point is that Linux is already only a part of the open source movement. We tend to think of it the whole environment as "Linux". But really the kernel is just one more modular part of the whole.
It was good to see that the BBC pointed out that relevance of the search results is probably more important than the number of pages in the database -- and that Google seems somehow to have the most relevant results time after time. Alltheweb didn't do very well on my standard test word, so I'm sticking with Google.
Slashdot Mirror
This is a fascinating document. But it just emphasises what a fundamental advance public key cryptography was.
Yeah, centimetres and inches.
Now compare that with last week's interview. Oh well, you win some, you lose some, I guess.
These distributed projects are all very well, but can one have too much of a good thing? How many people will take part in each one, I wonder, if they become commonplace?
Please don't post NY Times articles when you know of another one, because some of us object to registration!
Does anyone have a pointer to another article, please?
I must say, I find their web site thoroughly unconvincing. I mean, all their screenshots are only of flat windows from awkward angles. Even their descriptions don't explain why they're doing it, what advantages a 3D window manager has over a 2D one.
Maybe this is the way of the future. But they haven't convinced me of that yet.
I guess the question is, what sort of application are you talking about? Analog itself has to be blindingly fast, so C is the obvious choice. But its CGI interface isn't of that type. Speed -- or efficiency if you prefer -- really isn't an issue. So C loses one of its big advantages. I guess a lot of CGI applications are of this type.
Instead, it has to take some small amount of user input, munge it around, and spit it out again. That's exactly the sort of thing that Perl is perfect for. Converting my application from C to Perl, I saved an enormous amount of code length. Again, I guess this is typical of many CGI applications.
Of course, one has to be very careful about security in CGI applications. I think it's swings and roundabouts here. Both languages have similar theoretical problems. But some dangerous things are easy in Perl, and thus probably get done more readily. (E.g. backticks instead of system(), open("|...") instead of popen()). This is a worry: but weighed against this is Perl's -T mode, and its ability to check input for safety very quickly with regexps. In either language you have to know what you're doing here and be very careful to apply it!
And finally, you shouldn't underestimate Perl's readability. Whether in an open source or a closed source model, one day other people will want to maintain your code. Readability is good.
In summary, I'm far from a Perl expert, having written exactly one thing in it. But for my application, and I suspect for the majority of CGI applications, efficiency wasn't the key constraint. Ease of writing and maintaining was.
It's Her Majesty, not Her Royal Highness, by the way. HRH is only for lesser royals.
(Unfortunately my program analog is used on Her Majesty's site, but fails the same test!)
Suppose you could identify a set of personal characteristics and say "90% of killers are in this category." This is unlikely: but even then it still wouldn't follow that most people in that category are killers.
So what do you do? Unless a reasonable proportion of the positives are true positives, the program isn't really very useful. The danger would be that the school would target all the positives to avoid litigation, even though most positive kids were quite safe.
That's nice. But is there any good configuration file available for it, or do I just have to work my own out as I go?
Wouldn't it be possible for my browser to have an option to fetch inline images in a page only if the image was on the same server as the page? That way I could see the useful content, but discard all banner ads. Or would I discard too much useful stuff too? Comments?
If this is a US criminal offence, presumably I as a European can still squat? Or can I be extradited for it???
Swift argues that babies could be a delicacy for the upper classes, and a source of revenue, instead of a resource drain, on the working classes. Sounds morbid, but it's quite amusingly done too.
And it is relevant to this debate. Swift too was trying to argue against contemporary attitudes which counted certain people as worthless, although in his case it was the poor rather than the disabled.
You can read a short introduction to the proposal here.
Neither of them is free, in different ways. The first has no fixed charge but is metered through the telephone bill. The second would be unmetered but presumably have a higher standing charge.
Now, you can make a case that telecoms are overcharging. But that is an entirely separate issue from whether calls should be metered or unmetered. And given a choice between everyone being metered or everyone being unmetered, I'd go for metered every time.
Bandwidth is a scarce resource. By charging even a small amount for it you
I don't think it needs a high charge. I would guess that if the charge is metered instead of unmetered, the total charge would be lower for almost everybody, and the network would be less clogged up so that everyone could get their work, or play, done faster.
Oh, and hasn't anyone noticed that this model is popular with customers? It's essentially the model that Freeserve and their 100 imitators use, that has brought the internet to 100,000 new Britons a week for the last year.
I just want say thank you to Tim O'Reilly for answering all these questions so thoughtfully and so comprehensively. We all appreciate it.
There must be some much deeper cultural reason for our showing films like this uncut, although I'm not sure what it is. I'm always confused about whether the USA is a very liberal or a very conservative country!
And the game's called Cluedo in Europe. If only they'd stuck to that, they wouldn't be having this problem now. :)
Actually, there is a good reason -- speed. It has to make this calculation lots of times.
This article seems to make the usual confusion between the distribution and the kernel. Most people are working on a distribution or on a piece of software, rather than on the kernel itself.
If the Hurd microkernel is better, I hope I'll still be able to use the Debian distribution and the Apache web server: or whatever's better than them by then. Probably you won't be able to see easily which kernel I'm using anyway.
The point is that Linux is already only a part of the open source movement. We tend to think of it the whole environment as "Linux". But really the kernel is just one more modular part of the whole.
It was good to see that the BBC pointed out that relevance of the search results is probably more important than the number of pages in the database -- and that Google seems somehow to have the most relevant results time after time. Alltheweb didn't do very well on my standard test word, so I'm sticking with Google.