Don't write in C or C++ unless you absolutely have to. If you absolutely have to, make decent plans for dealing with the eventual security problems - like write the crap anonymously or using the name of some politician you decided you didn't like;).
There are plenty of other languages where a whole area of C/C++ problems just don't exist. Pick a decent language or two (Please NOT PHP! Forth is not much better than C). It'll still have problems, but at least you'd have a lot less "common programmer bug = attacker executes _arbitrary_ code of attackers choice" which should be totally unacceptable for modern code.
I personally use perl, and so far none of the bugs found in the programs I write for work would allow an attacker to "execute arbitrary code". It's more likely that any such bugs would likely be a bug in perl or the 3rd party libraries I use (nyah spread the blame;) ).
If you don't like perl, you could use python or similar.
Perl of course isn't fast (fast enough for a dhcpd[1]). But other languages like LISP, *ML are nowadays about as fast as C.
If you don't like the other fast languages, write the bulk of the stuff in a slow safe language that you like. Then write the performance critical programs in C and try to keep those really small and secure.
[1] Why a dhcpd in perl? We needed a fair number of features not in the isc-dhcpd and other dhcp servers. And go look at the source code of isc-dhcpd and a few other dhcp servers written in C sometime... Let me know what you think of the code;).
Why not Python? Existing Python libs don't allow us to make a dhcpd that does what we need.
If I were the Luxembourg Gov, I'd try to get my citizens to all have passports of other friendly countries.
IMO the best defense for you guys in event of a "conventional war" (ala WWII) is to all leave your country, hide somewhere else and then when convenient take the fight to the attacking country (why screw up your own country?).
If that's not possible and a lot have to stay behind then it's Mr Sniper/Saboteur time...
Problem is you'd probably be called terrorists;).
I guess the other option is you could just leave and not come back. Bah.
1) Just sail over with a big bunch of armed people 2) start a revolution (win of course) 3) execute dicators/kings if necessary 4) hold elections 5) establish democracy
They've pissed off the UK gov, so the UK gov is likely to ignore the extermination of Sealand.
Hasn't happened yet because someone else could easily do the same thing back to you and it's not worth the trouble for all that.
Gow would a random visitor using a free network know whether it's.lan or.local or.localnet or.localdomain? There's no standard at the moment. And that's my point.
How would one check the TOS of an intentionally free WAP, or see if there's a bulletin board, or maybe some cool service?
Heh no trouble finding porn. Spammers even send free samples to me, my mom, etc.
Anyway, my proposal is just a matter of helping people create defacto standards, a stepping stone to make further interesting stuff easier.
You suggest.localdomain, Apple (Bonjour) suggests.local, some suggest.localnet. I suggest that ICANN and the world stop mucking about and reserve one ala RFC1918. I don't really care what it is (as long as it's not too long or stupid), just frigging reserve one already.
You said: 'You enter "here" as an nntp server and you go to the local webserver, not the NNTP server. DNS does not care about ports, only addresses (DOH!).'
If you specified "here" as an NNTP server you would be trying to access the local server on port 119, I see no reason why you'd end up at the local server's webserver on port 80. They are different ports, capable of running different services without confusion. So what's the problem? Same goes for ssh and other stuff. If you'd prefer news.here that's fine.
What would airconditioner.here mean? That could be related to what http://here/ told you wouldn't it? I doubt it's polite to change settings of random airconditioners (nor would they necessarily allow you to do so - many ways to control it).
Believe me, I have had a lot of time to think about this and after all these years it still makes sense. I submitted a draft to the IETF in 2001 (and earlier to ICANN). If you think my proposal is convoluted/complex you should see the other proposals which create new proprietary "lock in" protocols to do fairly specialized things in similar areas.
My proposal uses _existing_ technology already available, and allows extensions. People can easily add stuff to it.
Sure DHCP sends a search path, but not everyone uses the supplied search path nor would want to. For example: an ISP network engineer might want to use isp.net instead of the search path supplied by Joe's Pool Place. But the network engineer may find it useful to be able to go http://here./ click on a link, get to http://jukebox.here./ and change the music being played there.
WRT to the dns server "problem" - people can get it from DHCP, or DNS packets could be redirected (believe me that's done all the time in many Paid Internet places with very little problems). Tons of people ALREADY have dns servers in their home- just go look at an ADSL router sometime.
If DNS is such a big problem one could even reserve a special IP. Any area server would respond to requests to that IP that pass through it. So at worst you'd see an ISP advert page or a Linksys router's default page;).
If you don't get anything or much from http://here/ you could assume that the area doesn't have interesting local services intentionally available for the public.
But maybe in some sites after http://here/ I could go to http://whos.here/ and then see ernestoalvarez.whos.here (Mr Alvarez's wearable server) listed and be able to upload or share files, multimedia streams etc.
For security reasons one should use encryption, but for most browsers this means you'd want to click on a "secure mode" link that contains the https url using the FQDN for that area. e.g. https://jungle.alvarez.foobar.org/ which would offer similar pages to here.
Anyway, virtual telepathy and telekinesis are just steps away.
It's a matter of helping people create defacto standards.
You know the TLD at your parents place is.jungle. But what's the TLD at Starbucks or your generous neighbour's open WiFi? If you travel visit various spots around the world how would you find out whose network you are using?
What if you want to control the jukebox at a local bowling alley? If we could have defacto standard where http://here/ either shows stuff "here" or redirects to a site which does then more and more sites would have an incentive to offer free internet access, because they could offer other services that make sense in the local physical context.
People wouldn't then mostly "pass through" and go on to the internet "super highway". They might just stop by for fun.
Forcibly redirecting people is lame, plus after they click through how do they find your starting page again?
Where's the RFC or standard or announcement reserving the.local TLD though?
I'm actually not too particular what it's called, as long as it's reserved and everyone can use it freely in whatever way they choose just like RFC1918 IP addresses.
As for Bonjour. Personally I think bonjour/zeroconf could become redundant- it'll be more likely that you'd end up having "home/room servers" anyway.
After all you already need a hub/switch/WAP/router and maybe even an internet connection or two. People already run Linux on modern routers/WAPs, DHCP and DNS already runs on some of these.
So all you need is for devices to register their hostnames with the home/room "server" and optionally submit the type of services they offer. Then the central server can list the devices (wearable servers, printers etc) present. And people can then wirelessly fetch/send messages/multimedia from/to each other's wearable servers (virtual telepathy), or control stuff via Web UIs on the devices or room servers (virtual telekinesis).
What's all this silly talk about.xxx making it harder to get porn.
Anyone thinking straight will know that.xxx will make it easier to find and get porn (as if it's not easy enough already).
Note: I'm not arguing against the.xxx TLD though. I personally think ICANN sucks, but looking at the other alternatives it seems like anyone who's likely to take over from ICANN would suck even more.
You all should just let me take over from ICANN:).
Block.xxx? Huh, all along I thought a valid technical reason for having.xxx was so people looking for porn could do the following google searches:
site:.xxx;).
Seriously though IMO.xxx is more justifiable than.biz or.info.
However, I disagree with you and the GP on TLDs.
I have long been arguing for.here to be a reserved TLD for free use for everyone - like the private RFC1918 IP addresses (10.x.x.x 192.168.x.x etc).
Basically everybody can host their own airconditioner.here in their houses/offices/rooms, and control it with http://airconditioner.here/set?temp=25c. And _polite_ people trying to figure out whether they are explicitly allowed to use an open WAP can go to http://here/ to look for terms and conditions, more info etc.
Well, tuna is decent enough food (except for modern day pollutants - PCB, mercury - which don't kill quickly at usual tuna levels), and so I guess is tomato paste:).
May even have been better than a diet of beer, sausages and sauerkraut;).
There are air dried versions of ramen, and those have negligible amounts of oil.
These often don't come with any flavouring as well, so you may have to do a bit of cooking if you don't have extra flavour packets around.
You could fry in olive oil and black pepper + a bit of chopped parsley, then add a fried egg. Yes you're adding oil back again, but good olive oil is worth it:). With the egg the entire meal gives you a fair balance of carbo, protein and fat.
What if someone took a video/picture of a public place that happened to have a monitor showing that scene?
I doubt that guy who takes those gigapixel pics of cities etc gets permission from everyone in his pics. In one of those "panoramic city skyline" pics apparently you can zoom to an apartment window and be able to see a poster on the apartment wall!
I always had the impression that BGP routes based on company policies, and that will remain true even if something better than BGP comes along;).
The shortest and quickest path for a packet could be one way, but the ISPs may all decide that it's more profitable, or "easier" for it to take a different path. The ISPs don't care that much on what's best for the end points.
After the earthquake broke stuff, even though was faster for say a packet to travel from say Malaysia to West Coast, USA via Singapore and Australia, instead the packets could go via Amsterdam, London, NYC etc... all the way across USA to the West Coast...
Many users figured out it really is faster to get to the USA via Australia and used proxies for whatever traffic they could.
But the ISPs won't share their bandwidth freely, and I really doubt it's due to technical limitations. After all why should an Australian ISP help other ISPs for free and risk affecting their own customers (or take extra effort to do it without affecting their customers)?
Thus it's not surprising that Akamai can do better for most endpoints than the ISPs.
Maybe things will change if there was more work done in establishing decent _money_ routing protocols for the ISPs;).
Most people don't understand the big difference between making software and say making a skyscraper.
1) With software, the _prototype_ blueprint actually compiles and runs and is usually sold as is (version 1.0).
2) Too many people keep thinking that the programming part is like the construction phase of a skyscraper, and that it should be planned and managed that way.
But they are WRONG, programming is more like coming up with a detailed design for a new skyscraper - the complete blueprint.
The actual construction of the skyscraper is more like a compile of the final version of the source code.
Blueprint = code. Skyscraper in use = code being run.
3) The other important difference with software is: The design cost of a software project is much more than the compiling cost. The design cost of a skyscraper project is usually far less than the construction cost.
The owners of skyscraper projects are usually willing to spend a fair bit more on the design if _necessary_. No problem spending a million more to make sure the skyscraper "just works" - because it's still a fraction of the cost of building a new skyscraper.
But for the owners of software projects _each_ design and redesign adds significantly to the cost of the project (in time and money), and it doesn't matter whether it's the prototype, or the "final" version you are designing. Redesign = project costs 2 x more.
Constructing one thousand copies of the same "skyscraper" is trivial and cheap in software but not for civil engineering.
Designing one thousand _different_ "skyscrapers" is not trivial or cheap in software or in civil engineering.
And the advantage of designing skyscrapers is it is easier to _realize_ and explain to the bosses that they are asking you to break the laws of physics - "sorry you can't fit an extra 3 elevators there, this is how much space one takes, this is how much space there is".
Seems to me that most "Management", "Software Engineering" and "Project Management" people[1] don't understand these differences, or don't want to accept these differences.
And that's why most software sucks and will continue to suck.
[1] A project management trainer once asked a class I was in for example suggestions on how to speed up the building phase of a software project - he said for a civil eng project you'd add more machinery and people e.g. bulldozers, construction workers etc. So I suggested that the equivalent was more and faster CPUs, and he didn't like that.
I doubt the civil engineering bunch will keep adding fresh grads to speed up the design phase of a new building.
Even if I install Linux (like on my office PC), I'd still keep Windows XP - it's still useful. I can run it on vmware server on Linux.
What people should do if they ever want windows is INSIST on XP instead of Vista!
If almost everybody stays with XP and DirectX 9 and doesn't move on to Vista. Then Windows XP+DX9 could become a defacto standard that even Microsoft can't get rid of (just like Intel can't get rid of x86 - they tried and failed with their Itanic, and when IBM tried to switch to MCA).
Then the jobs of people doing Wine, Crossover office, Cedega and more become a lot easier - they then have a fixed target instead of multiple moving targets.
Be realistic and ignore the fanboys out there, there are many valid reasons for wanting Windows. XP makes a good substitute for Vista, (until more and more people start switching to Vista).
But there is no Linux substitute for Windows yet. My proposal will make it more likely for there to be one.
Just a look at Vista will tell you that Microsoft is no longer improving things significantly or meaningfully, so we might as well freeze Windows, and be able to spend more time and resources on innovating elsewhere.
If we hijack Windows from Microsoft, then Microsoft will be like a BIOS vendor when it comes to Windows. Anyone remember "IBM compatible PC"?
So everyone, start telling Dell, HP et all to preload and sell XP instead of Vista.
The average home user won't be able to choose correctly between a Windows option and a no Windows option, or deal with either choices without help.
Not having Windows = average home user yells at Dell and asks why it doesn't work like the PC in the office. The different permutations of hardware/software options would work. Upgrade from Windows XP Home to Professional = no yell at Dell.
With the business models, it's the job of the buyer's IT dept to figure them out- and most of them can handle the choices.
If the car was not moving, it should: 1) Warn the driver if the car thinks the driver is drunk. 2) Require the driver to press an override button to start using the car if the car thinks the driver is drunk - otherwise car stays off. 3) If the button is pressed, notify the cops of the ID/license plate of the car (and position if available).
If the car was moving at the time same as the above except 2) is out and the cops are notified if the car doesn't stop within a minute or so.
If it's such an emergency that you still need to drive when the car thinks you're drunk, you shouldn't mind having the cops around.
If the system has so many false positives that notifying the cops is unacceptable, then the system is unacceptable.
Could even be a failure being spun to look like a success- I mean look at this:
"And what of that old mainframe? It's still around, but Isiminger wouldn't say exactly what it was up to. It operates in a "reduced capacity," he said"
"reduced capacity", that smells like BS talk. Does that mean it's still doing some of its old tasks, or most of its old tasks?
Of course they might not want to get rid of the mainframe that they paid so much money for (though there is a market for 2nd hand mainframes), but basically you can't always believe these Press Releases.
I also find it hard to get impressed that a cluster of 49 _new_ machines with 118 processors is only 70% faster and 65% cheaper than a mainframe that was available in early 2002. Plus that mainframe might even still be used to help the cluster to achieve those rates!
Next thing - how much more reliable and available is that cluster?
From the bullshit that was printed, I'm actually less inclined to think it was that huge a success.
But more details would be needed before I'd be sure whether they actually screwed up majorly, or did passably, or it was actually a big success (yeah right).
How old is the mainframe? The first mention of that model I see is at least in early 2002.
How easy is it to get a new server that's 70% faster than a 2002 server and 65% cheaper (were they using 2002 prices?)?
Of course it's likely that IBM wants to charge people _high_prices_ every year for using a mainframe, and if that's still the case, then I wouldn't recommend using a mainframe - they aren't fast in processing (they never were- just usually had more IO) and aren't even that reliable compared to other technologies - you tend to need scheduled downtime to fix stuff. Luckily for IBM, HP bought up competing technologies (VMS and Tandem) and are busy burying them as fast as possible.
Last but not least: this sort of stuff is a PR release for CxOs and not a technical document. So have to take these with a pinch of salt. I've seen failures spun in so many ways for these sort of "releases".
Maybe they still have the mainframe because they still need it to do something the grid can't do so now they have dependencies on two really different things...
"Is it really so hard to strike some middle ground between no options and so damned many you can't find the one you're looking for?"
Yeah, it's like compression BUT with the restriction you have to try to compress things (options) in a pseudoconsistent way, consistent enough so that people can remember it, but at the same time allow things to be added later on without breaking stuff.
You don't list out all the possible options in a single list. You try to put the popular ones first, then less popular ones in sublists and so on. BUT also try to maintain some consistency amongst popular apps.
I prefer KDE because it seems like if the devs can't figure out where to put some option, they often still just leave it visible (causing clutter). Whereas I hear GNOME devs would rather just _hide_ it away and force you to use gconf or some key sequence to enable it.
Take the pronunciation of "one" for instance. Not an obscure word, a fairly important word in my opinion, and one of the first few words someone learning english has to learn.
Now add the sloppiness that seems to be getting back in fashion - your != you're, have != of, it gets hard to understand what people are saying.
Back in the old days when only a few people in the world spoke English and even fewer wrote it it's not such a big deal if people keep changing the meanings.
But now I think all that is starting to get a bit gay;).
Slashdot Mirror
Don't write in C or C++ unless you absolutely have to. If you absolutely have to, make decent plans for dealing with the eventual security problems - like write the crap anonymously or using the name of some politician you decided you didn't like ;).
;) ).
;).
There are plenty of other languages where a whole area of C/C++ problems just don't exist. Pick a decent language or two (Please NOT PHP! Forth is not much better than C). It'll still have problems, but at least you'd have a lot less "common programmer bug = attacker executes _arbitrary_ code of attackers choice" which should be totally unacceptable for modern code.
I personally use perl, and so far none of the bugs found in the programs I write for work would allow an attacker to "execute arbitrary code". It's more likely that any such bugs would likely be a bug in perl or the 3rd party libraries I use (nyah spread the blame
If you don't like perl, you could use python or similar.
Perl of course isn't fast (fast enough for a dhcpd[1]). But other languages like LISP, *ML are nowadays about as fast as C.
If you don't like the other fast languages, write the bulk of the stuff in a slow safe language that you like. Then write the performance critical programs in C and try to keep those really small and secure.
[1] Why a dhcpd in perl? We needed a fair number of features not in the isc-dhcpd and other dhcp servers. And go look at the source code of isc-dhcpd and a few other dhcp servers written in C sometime... Let me know what you think of the code
Why not Python? Existing Python libs don't allow us to make a dhcpd that does what we need.
If I were the Luxembourg Gov, I'd try to get my citizens to all have passports of other friendly countries.
;).
IMO the best defense for you guys in event of a "conventional war" (ala WWII) is to all leave your country, hide somewhere else and then when convenient take the fight to the attacking country (why screw up your own country?).
If that's not possible and a lot have to stay behind then it's Mr Sniper/Saboteur time...
Problem is you'd probably be called terrorists
I guess the other option is you could just leave and not come back. Bah.
King?
1) Just sail over with a big bunch of armed people
2) start a revolution (win of course)
3) execute dicators/kings if necessary
4) hold elections
5) establish democracy
They've pissed off the UK gov, so the UK gov is likely to ignore the extermination of Sealand.
Hasn't happened yet because someone else could easily do the same thing back to you and it's not worth the trouble for all that.
Gow would a random visitor using a free network know whether it's .lan or .local or .localnet or .localdomain? There's no standard at the moment. And that's my point.
How would one check the TOS of an intentionally free WAP, or see if there's a bulletin board, or maybe some cool service?
Heh no trouble finding porn. Spammers even send free samples to me, my mom, etc.
.localdomain, Apple (Bonjour) suggests .local, some suggest .localnet. I suggest that ICANN and the world stop mucking about and reserve one ala RFC1918. I don't really care what it is (as long as it's not too long or stupid), just frigging reserve one already.
;).
Anyway, my proposal is just a matter of helping people create defacto standards, a stepping stone to make further interesting stuff easier.
You suggest
You said: 'You enter "here" as an nntp server and you go to the local webserver, not the NNTP server. DNS does not care about ports, only addresses (DOH!).'
If you specified "here" as an NNTP server you would be trying to access the local server on port 119, I see no reason why you'd end up at the local server's webserver on port 80. They are different ports, capable of running different services without confusion. So what's the problem? Same goes for ssh and other stuff. If you'd prefer news.here that's fine.
What would airconditioner.here mean? That could be related to what http://here/ told you wouldn't it? I doubt it's polite to change settings of random airconditioners (nor would they necessarily allow you to do so - many ways to control it).
Believe me, I have had a lot of time to think about this and after all these years it still makes sense. I submitted a draft to the IETF in 2001 (and earlier to ICANN). If you think my proposal is convoluted/complex you should see the other proposals which create new proprietary "lock in" protocols to do fairly specialized things in similar areas.
My proposal uses _existing_ technology already available, and allows extensions. People can easily add stuff to it.
Sure DHCP sends a search path, but not everyone uses the supplied search path nor would want to. For example: an ISP network engineer might want to use isp.net instead of the search path supplied by Joe's Pool Place. But the network engineer may find it useful to be able to go http://here./ click on a link, get to http://jukebox.here./ and change the music being played there.
WRT to the dns server "problem" - people can get it from DHCP, or DNS packets could be redirected (believe me that's done all the time in many Paid Internet places with very little problems). Tons of people ALREADY have dns servers in their home- just go look at an ADSL router sometime.
If DNS is such a big problem one could even reserve a special IP. Any area server would respond to requests to that IP that pass through it. So at worst you'd see an ISP advert page or a Linksys router's default page
If you don't get anything or much from http://here/ you could assume that the area doesn't have interesting local services intentionally available for the public.
But maybe in some sites after http://here/ I could go to http://whos.here/ and then see ernestoalvarez.whos.here (Mr Alvarez's wearable server) listed and be able to upload or share files, multimedia streams etc.
For security reasons one should use encryption, but for most browsers this means you'd want to click on a "secure mode" link that contains the https url using the FQDN for that area. e.g. https://jungle.alvarez.foobar.org/ which would offer similar pages to here.
Anyway, virtual telepathy and telekinesis are just steps away.
It's a matter of helping people create defacto standards.
.jungle.
You know the TLD at your parents place is
But what's the TLD at Starbucks or your generous neighbour's open WiFi? If you travel visit various spots around the world how would you find out whose network you are using?
What if you want to control the jukebox at a local bowling alley? If we could have defacto standard where http://here/ either shows stuff "here" or redirects to a site which does then more and more sites would have an incentive to offer free internet access, because they could offer other services that make sense in the local physical context.
People wouldn't then mostly "pass through" and go on to the internet "super highway". They might just stop by for fun.
Forcibly redirecting people is lame, plus after they click through how do they find your starting page again?
Where's the RFC or standard or announcement reserving the .local TLD though?
I'm actually not too particular what it's called, as long as it's reserved and everyone can use it freely in whatever way they choose just like RFC1918 IP addresses.
As for Bonjour. Personally I think bonjour/zeroconf could become redundant- it'll be more likely that you'd end up having "home/room servers" anyway.
After all you already need a hub/switch/WAP/router and maybe even an internet connection or two. People already run Linux on modern routers/WAPs, DHCP and DNS already runs on some of these.
So all you need is for devices to register their hostnames with the home/room "server" and optionally submit the type of services they offer. Then the central server can list the devices (wearable servers, printers etc) present. And people can then wirelessly fetch/send messages/multimedia from/to each other's wearable servers (virtual telepathy), or control stuff via Web UIs on the devices or room servers (virtual telekinesis).
What's all this silly talk about .xxx making it harder to get porn.
.xxx will make it easier to find and get porn (as if it's not easy enough already).
.xxx TLD though. I personally think ICANN sucks, but looking at the other alternatives it seems like anyone who's likely to take over from ICANN would suck even more.
:).
Anyone thinking straight will know that
Note: I'm not arguing against the
You all should just let me take over from ICANN
Block .xxx? Huh, all along I thought a valid technical reason for having .xxx was so people looking for porn could do the following google searches:
;).
.xxx is more justifiable than .biz or .info.
.here to be a reserved TLD for free use for everyone - like the private RFC1918 IP addresses (10.x.x.x 192.168.x.x etc).
h ere-01.txt
.xxx, .biz, .info and maybe even .tel.
.here and then give .here to the whole world.
site:.xxx
Seriously though IMO
However, I disagree with you and the GP on TLDs.
I have long been arguing for
Basically everybody can host their own airconditioner.here in their houses/offices/rooms, and control it with http://airconditioner.here/set?temp=25c. And _polite_ people trying to figure out whether they are explicitly allowed to use an open WAP can go to http://here/ to look for terms and conditions, more info etc.
For more do a search on tldhere:
http://www.watersprings.org/pub/id/draft-yeoh-tld
I think my proposal had and has a lot more merit and utility than
But I just don't have USD100K to give to ICANN to try to get
Well, tuna is decent enough food (except for modern day pollutants - PCB, mercury - which don't kill quickly at usual tuna levels), and so I guess is tomato paste :).
;).
May even have been better than a diet of beer, sausages and sauerkraut
Yeah, that works well with the soup based ramen. The one I was talking about was for the those that don't include soup.
:).
The average student could probably survive long enough on a diet of just instant ramen and boiled fresh egg to graduate or drop out for other reasons.
Not recommended though
Currently searchmash doesn't give me any results. Maybe it's because I don't have javascript enabled.
If that's where Google is going, I hope there will be an alternative. A UI for search shouldn't require javascript.
There are air dried versions of ramen, and those have negligible amounts of oil.
:). With the egg the entire meal gives you a fair balance of carbo, protein and fat.
These often don't come with any flavouring as well, so you may have to do a bit of cooking if you don't have extra flavour packets around.
You could fry in olive oil and black pepper + a bit of chopped parsley, then add a fried egg. Yes you're adding oil back again, but good olive oil is worth it
What if someone took a video/picture of a public place that happened to have a monitor showing that scene?
I doubt that guy who takes those gigapixel pics of cities etc gets permission from everyone in his pics. In one of those "panoramic city skyline" pics apparently you can zoom to an apartment window and be able to see a poster on the apartment wall!
I always had the impression that BGP routes based on company policies, and that will remain true even if something better than BGP comes along ;).
;).
The shortest and quickest path for a packet could be one way, but the ISPs may all decide that it's more profitable, or "easier" for it to take a different path. The ISPs don't care that much on what's best for the end points.
After the earthquake broke stuff, even though was faster for say a packet to travel from say Malaysia to West Coast, USA via Singapore and Australia, instead the packets could go via Amsterdam, London, NYC etc... all the way across USA to the West Coast...
Many users figured out it really is faster to get to the USA via Australia and used proxies for whatever traffic they could.
But the ISPs won't share their bandwidth freely, and I really doubt it's due to technical limitations. After all why should an Australian ISP help other ISPs for free and risk affecting their own customers (or take extra effort to do it without affecting their customers)?
Thus it's not surprising that Akamai can do better for most endpoints than the ISPs.
Maybe things will change if there was more work done in establishing decent _money_ routing protocols for the ISPs
Linux is not a substitute for Vista. Windows XP is, for now unless more and more people switch to Vista.
As long as people keep switching to the next MS version of windows MS will continue to have control.
If people stick with XP, then Linux and everyone else will have a chance.
http://slashdot.org/~TheLink/journal/158520
Software still sucks because:
Most people don't understand the big difference between making software and say making a skyscraper.
1) With software, the _prototype_ blueprint actually compiles and runs and is usually sold as is (version 1.0).
2) Too many people keep thinking that the programming part is like the construction phase of a skyscraper, and that it should be planned and managed that way.
But they are WRONG, programming is more like coming up with a detailed design for a new skyscraper - the complete blueprint.
The actual construction of the skyscraper is more like a compile of the final version of the source code.
Blueprint = code. Skyscraper in use = code being run.
3) The other important difference with software is:
The design cost of a software project is much more than the compiling cost.
The design cost of a skyscraper project is usually far less than the construction cost.
The owners of skyscraper projects are usually willing to spend a fair bit more on the design if _necessary_. No problem spending a million more to make sure the skyscraper "just works" - because it's still a fraction of the cost of building a new skyscraper.
But for the owners of software projects _each_ design and redesign adds significantly to the cost of the project (in time and money), and it doesn't matter whether it's the prototype, or the "final" version you are designing. Redesign = project costs 2 x more.
Constructing one thousand copies of the same "skyscraper" is trivial and cheap in software but not for civil engineering.
Designing one thousand _different_ "skyscrapers" is not trivial or cheap in software or in civil engineering.
And the advantage of designing skyscrapers is it is easier to _realize_ and explain to the bosses that they are asking you to break the laws of physics - "sorry you can't fit an extra 3 elevators there, this is how much space one takes, this is how much space there is".
Seems to me that most "Management", "Software Engineering" and "Project Management" people[1] don't understand these differences, or don't want to accept these differences.
And that's why most software sucks and will continue to suck.
[1] A project management trainer once asked a class I was in for example suggestions on how to speed up the building phase of a software project - he said for a civil eng project you'd add more machinery and people e.g. bulldozers, construction workers etc. So I suggested that the equivalent was more and faster CPUs, and he didn't like that.
I doubt the civil engineering bunch will keep adding fresh grads to speed up the design phase of a new building.
Even if I install Linux (like on my office PC), I'd still keep Windows XP - it's still useful. I can run it on vmware server on Linux.
What people should do if they ever want windows is INSIST on XP instead of Vista!
If almost everybody stays with XP and DirectX 9 and doesn't move on to Vista. Then Windows XP+DX9 could become a defacto standard that even Microsoft can't get rid of (just like Intel can't get rid of x86 - they tried and failed with their Itanic, and when IBM tried to switch to MCA).
Then the jobs of people doing Wine, Crossover office, Cedega and more become a lot easier - they then have a fixed target instead of multiple moving targets.
Be realistic and ignore the fanboys out there, there are many valid reasons for wanting Windows. XP makes a good substitute for Vista, (until more and more people start switching to Vista).
But there is no Linux substitute for Windows yet. My proposal will make it more likely for there to be one.
Just a look at Vista will tell you that Microsoft is no longer improving things significantly or meaningfully, so we might as well freeze Windows, and be able to spend more time and resources on innovating elsewhere.
If we hijack Windows from Microsoft, then Microsoft will be like a BIOS vendor when it comes to Windows. Anyone remember "IBM compatible PC"?
So everyone, start telling Dell, HP et all to preload and sell XP instead of Vista.
Yeah, and that's what I'd do too.
The average home user won't be able to choose correctly between a Windows option and a no Windows option, or deal with either choices without help.
Not having Windows = average home user yells at Dell and asks why it doesn't work like the PC in the office.
The different permutations of hardware/software options would work. Upgrade from Windows XP Home to Professional = no yell at Dell.
With the business models, it's the job of the buyer's IT dept to figure them out- and most of them can handle the choices.
If the car was not moving, it should:
1) Warn the driver if the car thinks the driver is drunk.
2) Require the driver to press an override button to start using the car if the car thinks the driver is drunk - otherwise car stays off.
3) If the button is pressed, notify the cops of the ID/license plate of the car (and position if available).
If the car was moving at the time same as the above except 2) is out and the cops are notified if the car doesn't stop within a minute or so.
If it's such an emergency that you still need to drive when the car thinks you're drunk, you shouldn't mind having the cops around.
If the system has so many false positives that notifying the cops is unacceptable, then the system is unacceptable.
No real substance.
Could even be a failure being spun to look like a success- I mean look at this:
"And what of that old mainframe? It's still around, but Isiminger wouldn't say exactly what it was up to. It operates in a "reduced capacity," he said"
"reduced capacity", that smells like BS talk. Does that mean it's still doing some of its old tasks, or most of its old tasks?
Of course they might not want to get rid of the mainframe that they paid so much money for (though there is a market for 2nd hand mainframes), but basically you can't always believe these Press Releases.
I also find it hard to get impressed that a cluster of 49 _new_ machines with 118 processors is only 70% faster and 65% cheaper than a mainframe that was available in early 2002. Plus that mainframe might even still be used to help the cluster to achieve those rates!
Next thing - how much more reliable and available is that cluster?
From the bullshit that was printed, I'm actually less inclined to think it was that huge a success.
But more details would be needed before I'd be sure whether they actually screwed up majorly, or did passably, or it was actually a big success (yeah right).
How old is the mainframe? The first mention of that model I see is at least in early 2002.
How easy is it to get a new server that's 70% faster than a 2002 server and 65% cheaper (were they using 2002 prices?)?
Of course it's likely that IBM wants to charge people _high_prices_ every year for using a mainframe, and if that's still the case, then I wouldn't recommend using a mainframe - they aren't fast in processing (they never were- just usually had more IO) and aren't even that reliable compared to other technologies - you tend to need scheduled downtime to fix stuff. Luckily for IBM, HP bought up competing technologies (VMS and Tandem) and are busy burying them as fast as possible.
Last but not least: this sort of stuff is a PR release for CxOs and not a technical document. So have to take these with a pinch of salt. I've seen failures spun in so many ways for these sort of "releases".
Maybe they still have the mainframe because they still need it to do something the grid can't do so now they have dependencies on two really different things...
"Is it really so hard to strike some middle ground between no options and so damned many you can't find the one you're looking for?"
Yeah, it's like compression BUT with the restriction you have to try to compress things (options) in a pseudoconsistent way, consistent enough so that people can remember it, but at the same time allow things to be added later on without breaking stuff.
You don't list out all the possible options in a single list. You try to put the popular ones first, then less popular ones in sublists and so on. BUT also try to maintain some consistency amongst popular apps.
I prefer KDE because it seems like if the devs can't figure out where to put some option, they often still just leave it visible (causing clutter). Whereas I hear GNOME devs would rather just _hide_ it away and force you to use gconf or some key sequence to enable it.
Oh that sounds bad then. I doubt popular widgets will be that reliable/stable. Some might even be exploitable...
English is just a crazy language.
;).
Take the pronunciation of "one" for instance. Not an obscure word, a fairly important word in my opinion, and one of the first few words someone learning english has to learn.
Now add the sloppiness that seems to be getting back in fashion - your != you're, have != of, it gets hard to understand what people are saying.
Back in the old days when only a few people in the world spoke English and even fewer wrote it it's not such a big deal if people keep changing the meanings.
But now I think all that is starting to get a bit gay