But what if a large number of tree huggers totally stop buying cars and switch to something else (bicycles, public transport, walking etc)
While the car market would shrink, the car market for fuel efficient cars might shrink even more. So might car makers start focusing on non fuel efficient cars instead?;)
I dislike spammers, but that's rather long for spamming.
I'd rather they catch 11 spammers and give them 1 year each.
Catching just one and sentencing him to 11 years is just silly, and doesn't help much.
Most people believe they won't get caught. If you start catching and jailing 100s of spammers for even sentences of a few months AND fine them so they end up with a significant net loss, then spammers will stop spamming - because they start noticing that spammers ARE getting caught.
What's with these crazy sentences anyway? Like teacher threatened with 40 year jail sentence for showing porn to kids, whether it's voluntary or not who cares, 40 years is crazy. Or that poor kid in georgia who's serving out a 10 year sentence for having _consensual_ oral sex with a teenage girl - court calls it "child molestation" for some reason.
quote: "If we fully embrace nuclear power then efficiency could be less important then convenience."
Well I hope we don't completely forget about efficiency: if everyone has access to gigawatts of power and uses it all, it'll give global warming a new meaning...
Because of the resulting unbearable ambient temperatures (plentiful energy = lots of heat), everyone starts using more and more energy to pump heat skywards. The earth may start to glow as a result;).
Probably got thrown out by my mom with some other old stuff.
Anyway, it shouldn't be too hard to do again.
Basically if you press ctrl-apple-esc on the Apple IIGS that sends an interrupt that launches the "classic desk accessory" screen, there used to be docs on how to add programs to that.
On the Apple IIGS the old Apple II memory locations were in 0x000000 to 0x01ffff, So you write a "desk accessory" that copies the stuff in the memory to some other area (allocating the mem using the usual apple IIGS toolbox routines), exclude the IO stuff. The tricky bits are to do the bank switching stuff to access the stuff mapped to 0x00d000-0x00ffff.
Then you save the registers, and then you return (from the interrupt call), for restore you copy the stuff back, restore the saved registers and return. Of course the problem is not everything could be saveable- the exact machine state might not be entirely restoreable - stuff could have been written to the disk etc.
Reasonable trivial stuff. Anyway, nowadays with Apple IIGS emulators around, there's not much point - you can save the entire state, even if the game or application disables the "desk accessory" stuff.
Re:The first computer I owned
on
The Apple II At 30
·
· Score: 2, Interesting
I had an Apple II clone (called Viking-X), and later on a genuine Apple IIGS (and that was really expensive not even factoring in inflation!).
The early manuals were wonderful (the clone I had came with a manual that was a copy of 3 apple manuals combined;) ).
At about 8 years old, I learnt BASIC and 6502 machine code (cycle counting etc) from just that manual and the Apple II clone. I still keep that manual around, and I think people shouldn't underestimate what children can learn given decent sources of information.
Back then I "enhanced" a few games and even made a slightly modified DOS - and later on I wrote some disk caching software for the Apple IIGS - it cached some metadata (e.g. directory info reducing seeks), and cached a track (if the right sector wasn't around it still cached the "wrong" sectors just in case they'd be requested later). The whole idea was not to try to cache everything that passes by, but to increase and improve sequential reads - RAM was not that plentiful then. Worked pretty well if I do say so myself, but I never really released it (I think I passed it to a few friends and that's about it). Also made a utility for the Apple IIGS that allowed you to save/resume in old Apple II games - this was done by copying the entire "old Apple II" memory area, stack, CPU state etc, and restoring it if desired. Did lot of other stuff too.
Of course, now I've gone downhill and write stuff in Perl which on modern machines does loops a bit faster than 6502 code on a 1MHz 6502 or even a 2.5/2.8Mhz 65c816;).
Maybe it'll eventually be a bit like MP3 audio compression. You try to cut corners for the "resource expensive" stuff in a way that most people are unlikely to notice.
If I go to a restaurant and the waiter tells me what the cook isn't good at and recommends something that the cook is actually good at, that I'm likely to enjoy - given my stated constraints to the waiter (time, money, dietary preferences) I am more likely to return (and even recommend the restaurant to others).
Whereas, if the waiter recommends something that leaves a bad taste in my mouth, what are the odds I'd come back, and what are the odds that I'd tell others to NOT go to that place?
Of course, if I still insisted on ordering a valid menu item the waiter recommended against, the waiter should shut up and get it for me. I'd never do that of course, but hey if someone is like that, you probably want them to switch to eating at your competitors ASAP;).
On a vaguely related topic, I wonder who could have recommended any of those APS photo products with a clear conscience. APS was a format that was crap before it was even launched.
The problem is while analog prints degrade slowly if Joe Public leaves them wherever they are, digital stuff can effectively just vanish if Joe Public just leaves them wherever they are.
While HDDs are getting bigger and bigger, who is going to educate Joe Public to keep migrating old archival data or even _converting_ the data so that it can continue to be accessible?
Basically the "degradation curve" is still there, it just looks different. You lose stuff in chunks, rather than gradual fades of analog stuff (interspersed with lost chunks due to fire etc).
1) Nowadays it's practically all software. 2) A modern day 2GHz x86 core with Linux software RAID might be a lot faster at RAID than some expensive RAID controller with an el-cheapo 33MHz uP. An extra core+"software" RAID might be cheaper and more reliable than a RAID controller.
Examples of interesting questions people should be asking is:
1) How would you be notified if something has failed? It's pretty easy with Linux software raid, not so simple with hardware raid controllers+Linux (you need driver support etc). 2) How do you get stuff to regularly and automatically scan for "bad sectors"? Better to find them early. 3) What is the rebuild process like? 4) What happens if you have "bad sectors" in multiple disks, but not on the same logical block/sector. Some controllers/software can handle that gracefully - you can sometimes have read errors while rebuilding an array. You don't want the entire array to go offline just because the controller is too stupid to figure out the data is still there and good - it's just on different combinations of disks.
And there are more details like: say you have a RAID1 mirror, and the controller/software detects a read failure from one disk. There are a number of possible things it can do. It could log the error, and then retry the read from another disk, if that is successful, it could even try to write that block to the faulty disk to get the drive to remap that sector elsewhere - and only if that fails it marks the disk as offline.
From the perspective of a web application programmer and security consultant, I think it would be very useful to have HTML tags to mark HTML sections where active content should be disabled, possibly selected active content.
Right now the HTML environment with respect to potentially dangerous content is: In order to stop, you must make sure that none of the 1001 GO buttons were pressed before. There is no STOP button. No Big Red Emergency Stop button.
This seems to be a disaster prone situation. Like driving a car without brakes. Only experts can do it, and typically even they screw up too.
I think we need some form of brakes. Something like the following:
<activeoff lock="matchingrandomstring" allowed="java"/> Any active content disabled here. Even if slips past site's filters. <activeon lock="matchingrandomstring"/> The disabled active content reenabled. Does not mean everything enabled, just those disabled earlier.
I'm sure that, (for example) most people in Japan would have a rather different view on this "toilet stuff".
And in a number of places around the world there are squat toilets that have no seats or lids anyway, just be happy if it's fairly clean and there's clean water from tap etc.
On the topic of hygiene, what's so clean about partially wiping/smearing your shit off with toilet paper? How effective is that? Why not wash your shit off thoroughly with water and soap, or use whatever it is you consider good enough to wash your hands clean _enough_ (go check the CDC's recommendation - AFAIK 60-90% alcohol solutions work well at killing stuff, but are not so good for removing stuff ). There is no 100% clean in most houses/places, and your hands will get dirty the instant you touch something anyway, you're just going to have to live with it - or die trying;).
As for what Other People do, why be so bothered on whether they leave the lid up or down? How about whether they washed their hands properly before they touched other stuff, and whether they left traces of shit on the tap knobs or elsewhere. I'm not immune compromised so even if that thought may sometimes seem gross, I'm not going to let it gross me out a lot.
The last I heard, the US Gov will have access to X bits of the Lotus Notes keys (some of the keys bits are taken and encrypted to the US Gov key), so that they get a significant help to cracking stuff if they need to. Something like it's 40bit crypto for the US Gov, and 64 bit crypto for everyone else (other than the intended recipients).
You said: "Why on hell? For one, it wouldn't be www.microsoft.com but com/microsoft/www. For second, assuming equivalent to current expansions, it would be "www" the one to expand to microsoft/www or com/microsoft/www, exactly like now. I really don't see where the "prefixes for alternate searchs" in your example come from."
Are you for real? I gave the example that way because all the following URLs are expressed in the same way in the "New Berners" approach.
All these "conventional URLs": http://com:82/microsoft/www/foo/bar/com/baz/ http://microsoft.com:82/www/foo/bar/com/baz/ http://www.microsoft.com:82/foo/bar/com/baz/ http://foo.www.microsoft.com:82/bar/com/baz/ http://bar.foo.www.microsoft.com:82/com/baz/ http://com.bar.foo.www.microsoft.com:82/baz/ http://baz.com.bar.foo.www.microsoft.com:82/
Are all represented in the "New Berners" approach in a single URL as: http/82:com/microsoft/www/foo/bar/com/baz
Pasting 7 identical lines (ala new berners URL) to try to illustrate 7 possible different browser "attempts" (whether in parallel or not) would be silly. So that's why I stuck to the conventional (in fact I skipped some details and assumed basic knowledge of HTTP and TCP, as well as DNS, which may be the problem).
So even if it's just a single "New Berners" URL the browser still has to _potentially_ try ALL the possible _effective_ combinations in event of failures to fetch the "effective urls" - whether due to DNS failure, or failure to reach/connect to the webserver. Sure you might be able to avoid/workaround some of that by doing various stuff. But why do all that? What would people gain?
"What do you think is current DNS server topology? It already is a hierarchy of hosts, so no news here too"
While the DNS topology is a hierachy of hosts, the current WWW is NOT - it is a _Web_ of hosts. Right now once webmasters get their domain name and IP, they have a lot of freedom in the pathnames and filenames they can use. They will lose that freedom with the "new approach". Instead of just making a directory and putting a file somewhere, they'd have to do a few more dependency checks - e.g. does the new directory or file cause a clash with someone else's subdomain (past/present/planned). More work and cooperation required.
You're in a way proposing breaking the Web- forcing it into a hierachy (in fact tying it tightly to the DNS hierachy) and then using "SRV-like" stuff to glue the broken bits together. I repeat: what would people gain from that? Is it worth it?
Sorry, I'm giving up already, it's too much work explaining the simple stuff already (I'm a lazy person). If it makes you happy, go ahead and assume I'm wrong by default.
Would be efficient distributed locking and messaging at an O/S level that supports clustering too.
Just like the O/Ses of those good old days;)
Running stuff in parallel is trivial. Just get another core/processer/computer, run the process. The issue is when the processes need to exchange information or they need to be serialized.
Seems a lot of people nowadays delegate that stuff to some DB software (since DB software needs locking, serialization etc too, and provides it), but DB software has typically higher overheads and also a fair number of DBs have problems running across clusters too...
If you design the APIs well programmers will start using multiple cores, and even multiple computers. It's not necessarily an easy problem, but it's been done more than once already, I'm sure with the benefit of hindsight, some hardware support from say Intel, things should be even better nowadays.
I may be ignorant, but I'm not that ignorant. I know DNS and how it works. And I also know how lots of other stuff works too.
You're wrong because it's not just about DNS. It's about how things work in real life and equally importantly it's about how things _FAIL_ in real life.
TCP connections still take time to make, and there is a significant timeout if the other end is firewalled (ignores TCP/80).
Try this on a linux/*BSD box: time wget http://www.microsoft.com:82/foo/bar/com/baz
See how long that takes to time out. Sure interactive browsers might time out faster, but unlike DNS, they won't timeout after just a few seconds.
And remember after that times out, with the "New Berners" approach you will have to try to fetch: http://foo.www.microsoft.com:82/bar/com/baz http://bar.foo.www.microsoft.com:82/com/baz http://com.bar.foo.www.microsoft.com:82/baz http://baz.com.bar.foo.www.microsoft.com:82/ And only after all that should the browser give up.
There's also the scenario of trying to access a site that hosts lots of different people's stuff that uses a wildcarded DNS- say the dns works but the site is down - how long do you wait? All of the possible domains will work (you expect the admin to set up a system to put all the valid names of all the sites in the DNS? Esp when previously the customers got to have their own arbitrary sub domain names without any change to the configs).
OK lets say you try to do stuff in parallel, and display the first document that is successfully fetched. But what happens then if you get multiple documents? If a server higher up the hierachy (thus being more heavily loaded and more likely to be "slow") finally responds while the browser is _halfway_ displaying a different file should you suddenly tell the user "Ooops, pretend you didn't see that, here's what you should be seeing". What if you get tons of HTTP/404s? Which 404 should you show? There are pretty fancy 404s nowadays. You think that's bad? What if you get multiple HTTP/302s! Which 302 should you follow? All of them? And risk the problem getting even bigger?
And should the browser do negative caching for all failures? How long?
Sure you can put "don't recurse" stuff on the DNS servers, but in real life, the people who run the webservers often have little authority and control over the DNS servers.
Run the DNS server on the webserver? Despite what some people may like, not every web server be allowed to run an authoritative DNS server on it AND get the firewall administrators to pass DNS traffic to their pet server, nor is it likely that the DNS delegation be correctly done in enough cases for people to say "this system is viable".
Lastly, the "New Berners" approach is trying to merge file namespaces with host namespaces, sure that could work fine in scenarios where one entity controls everything. BUT with this approach you will no longer have an "every host is a peer" situation, it will be a hierachy of hosts. Some hosts WILL override other hosts so you can no longer put stuff in certain namespaces on those hosts. You will start to need cooperation amongst previously independent hosts/peers to avoid undesirable namespace clashes at what was previously a _file_ naming level.
If the cellphone companies don't want to pay they can stop providing cellular coverage for that entire area.
Sounds reasonable to me. Let's see how well their area does if nobody provides coverage to them.
Of course if they really want to be a sovereign nation and start making rules like that, then it's time they stopped collecting "pocket money" from "Step-mommy" just because "the evil great-great-great-grandstep mom" was bad to them, and started paying their own way.
If a culture/people takes so long to recover from such a set back then something is wrong somewhere.
A not so atypical scenario: First generation immigrant works hard and becomes rich, 2nd generation - maintains the wealth, 3rd generation blows it all away:). So in my opinion, the max a country should pay compensation to a peoples for past evils would be at _maximum_ 3 generations (of course if the country still doesn't think it did anything wrong in the first place, you better be careful).
With that, you can't tell where a host name begins and where the URI starts.
And he even thinks that is GOOD: quote: "This would mean the BCS could have one server for the whole site or have one specific to members and the URL wouldn't have to be different."
Doh.
Say you have a conventional URL of http://blah.example.com/sub/foo If we do things he proposed in that page how does he expect the browser to find the IP address for the server to go to?
With his suggestion the url will look like:
http:com/example/blah/sub/foo
Now that's very nice in "dreamland" where the speed of light is infinite and everything is perfect.
But in the real world, what domain name should the browser try in order to get the IP address to connect to?
Should the browser try to connect to "com" and fetch/example/blah/sub/foo Then if that fails connect to: example.com and try to fetch/blah/sub/foo Then if that fails connect to: blah.example.com and try to fetch/sub/foo
AND WORSE, even if that's the correct URL, say the server was temporarily broken/misconfigured, so now the browser is suppose to keep going? e.g.connect to: sub.blah.example.com and fetch/foo then try foo.sub.blah.example.com
The browser has to wait for the necessary failure timeouts on each try. Don't forget, the URL I used as an example isn't even a very long one. Imagine one with a greater "directory depth".
Make you wonder if he "stumbled" on his _original_ scheme by sheer luck. Or he actually thought long and hard on it, and has now unfortunately forgotten the original reasons why things were done that way.
Nowadays I find there are not very many people who understand how lots of different things work, the various limitations, and how certain choices/changes affect things. There's often so much you need to know AND keep in mind.
As for security: Apache isn't what you'd call secure software, but Lighttpd isn't either.
PHP is slower than Perl or Python for most stuff. This should be common knowledge amongst decent programmers- go find/make your own benchmarks and links.
lighttpd leaks memory, how secure and "lightweight" is that?
So far Apache is good enough for me, but if you're going to replace Apache with something that is fast and has features, you may wish try Zeus.
Anyway, I'd dump PHP first. PHP is slow and has tons of security problems - and judging from the way the devs do stuff, there'll be plenty more to come.
Slashdot Mirror
But what if a large number of tree huggers totally stop buying cars and switch to something else (bicycles, public transport, walking etc)
;)
While the car market would shrink, the car market for fuel efficient cars might shrink even more. So might car makers start focusing on non fuel efficient cars instead?
Just curious: does replacing EXE in urls with %45%58%45 work?
Heh it shouldn't work but who knows...
I dislike spammers, but that's rather long for spamming.
I'd rather they catch 11 spammers and give them 1 year each.
Catching just one and sentencing him to 11 years is just silly, and doesn't help much.
Most people believe they won't get caught. If you start catching and jailing 100s of spammers for even sentences of a few months AND fine them so they end up with a significant net loss, then spammers will stop spamming - because they start noticing that spammers ARE getting caught.
What's with these crazy sentences anyway? Like teacher threatened with 40 year jail sentence for showing porn to kids, whether it's voluntary or not who cares, 40 years is crazy. Or that poor kid in georgia who's serving out a 10 year sentence for having _consensual_ oral sex with a teenage girl - court calls it "child molestation" for some reason.
quote: "If we fully embrace nuclear power then efficiency could be less important then convenience."
;).
Well I hope we don't completely forget about efficiency: if everyone has access to gigawatts of power and uses it all, it'll give global warming a new meaning...
Because of the resulting unbearable ambient temperatures (plentiful energy = lots of heat), everyone starts using more and more energy to pump heat skywards. The earth may start to glow as a result
Uh, copyright laws have made public domain stuff go back into private domain.
There have been retroactive extensions at least for the USA.
Probably got thrown out by my mom with some other old stuff.
Anyway, it shouldn't be too hard to do again.
Basically if you press ctrl-apple-esc on the Apple IIGS that sends an interrupt that launches the "classic desk accessory" screen, there used to be docs on how to add programs to that.
On the Apple IIGS the old Apple II memory locations were in 0x000000 to 0x01ffff,
So you write a "desk accessory" that copies the stuff in the memory to some other area (allocating the mem using the usual apple IIGS toolbox routines), exclude the IO stuff. The tricky bits are to do the bank switching stuff to access the stuff mapped to 0x00d000-0x00ffff.
Then you save the registers, and then you return (from the interrupt call), for restore you copy the stuff back, restore the saved registers and return. Of course the problem is not everything could be saveable- the exact machine state might not be entirely restoreable - stuff could have been written to the disk etc.
Reasonable trivial stuff. Anyway, nowadays with Apple IIGS emulators around, there's not much point - you can save the entire state, even if the game or application disables the "desk accessory" stuff.
I had an Apple II clone (called Viking-X), and later on a genuine Apple IIGS (and that was really expensive not even factoring in inflation!).
;) ).
;).
The early manuals were wonderful (the clone I had came with a manual that was a copy of 3 apple manuals combined
At about 8 years old, I learnt BASIC and 6502 machine code (cycle counting etc) from just that manual and the Apple II clone. I still keep that manual around, and I think people shouldn't underestimate what children can learn given decent sources of information.
Back then I "enhanced" a few games and even made a slightly modified DOS - and later on I wrote some disk caching software for the Apple IIGS - it cached some metadata (e.g. directory info reducing seeks), and cached a track (if the right sector wasn't around it still cached the "wrong" sectors just in case they'd be requested later). The whole idea was not to try to cache everything that passes by, but to increase and improve sequential reads - RAM was not that plentiful then. Worked pretty well if I do say so myself, but I never really released it (I think I passed it to a few friends and that's about it). Also made a utility for the Apple IIGS that allowed you to save/resume in old Apple II games - this was done by copying the entire "old Apple II" memory area, stack, CPU state etc, and restoring it if desired. Did lot of other stuff too.
Of course, now I've gone downhill and write stuff in Perl which on modern machines does loops a bit faster than 6502 code on a 1MHz 6502 or even a 2.5/2.8Mhz 65c816
Times have sure changed.
Yeah.
Maybe it'll eventually be a bit like MP3 audio compression. You try to cut corners for the "resource expensive" stuff in a way that most people are unlikely to notice.
If I go to a restaurant and the waiter tells me what the cook isn't good at and recommends something that the cook is actually good at, that I'm likely to enjoy - given my stated constraints to the waiter (time, money, dietary preferences) I am more likely to return (and even recommend the restaurant to others).
;).
Whereas, if the waiter recommends something that leaves a bad taste in my mouth, what are the odds I'd come back, and what are the odds that I'd tell others to NOT go to that place?
Of course, if I still insisted on ordering a valid menu item the waiter recommended against, the waiter should shut up and get it for me. I'd never do that of course, but hey if someone is like that, you probably want them to switch to eating at your competitors ASAP
On a vaguely related topic, I wonder who could have recommended any of those APS photo products with a clear conscience. APS was a format that was crap before it was even launched.
The problem is while analog prints degrade slowly if Joe Public leaves them wherever they are, digital stuff can effectively just vanish if Joe Public just leaves them wherever they are.
While HDDs are getting bigger and bigger, who is going to educate Joe Public to keep migrating old archival data or even _converting_ the data so that it can continue to be accessible?
Basically the "degradation curve" is still there, it just looks different. You lose stuff in chunks, rather than gradual fades of analog stuff (interspersed with lost chunks due to fire etc).
Problem with digital is it is too likely to end up with stuff in a format or medium that can no longer be understood by any of the viewers available.
1) Nowadays it's practically all software.
2) A modern day 2GHz x86 core with Linux software RAID might be a lot faster at RAID than some expensive RAID controller with an el-cheapo 33MHz uP. An extra core+"software" RAID might be cheaper and more reliable than a RAID controller.
Examples of interesting questions people should be asking is:
1) How would you be notified if something has failed? It's pretty easy with Linux software raid, not so simple with hardware raid controllers+Linux (you need driver support etc).
2) How do you get stuff to regularly and automatically scan for "bad sectors"? Better to find them early.
3) What is the rebuild process like?
4) What happens if you have "bad sectors" in multiple disks, but not on the same logical block/sector. Some controllers/software can handle that gracefully - you can sometimes have read errors while rebuilding an array. You don't want the entire array to go offline just because the controller is too stupid to figure out the data is still there and good - it's just on different combinations of disks.
And there are more details like: say you have a RAID1 mirror, and the controller/software detects a read failure from one disk. There are a number of possible things it can do. It could log the error, and then retry the read from another disk, if that is successful, it could even try to write that block to the faulty disk to get the drive to remap that sector elsewhere - and only if that fails it marks the disk as offline.
From the perspective of a web application programmer and security consultant, I think it would be very useful to have HTML tags to mark HTML
/> />
. mozilla.security/browse_thread/thread/c02d6dfa7181 1d62/6d4cf22651a72812?lnk=st&q=&rnum=2&hl=en#6d4cf 22651a72812
sections where active content should be disabled, possibly selected active content.
Right now the HTML environment with respect to potentially dangerous
content is:
In order to stop, you must make sure that none of the 1001 GO buttons were
pressed before. There is no STOP button. No Big Red Emergency Stop button.
This seems to be a disaster prone situation. Like driving a car without
brakes. Only experts can do it, and typically even they screw up too.
I think we need some form of brakes. Something like the following:
<activeoff lock="matchingrandomstring" allowed="java"
Any active content disabled here. Even if slips past site's filters.
<activeon lock="matchingrandomstring"
The disabled active content reenabled. Does not mean everything enabled,
just those disabled earlier.
Rest here:
http://groups.google.com.my/group/netscape.public
Which cultures is this a big deal in?
;).
I'm sure that, (for example) most people in Japan would have a rather different view on this "toilet stuff".
And in a number of places around the world there are squat toilets that have no seats or lids anyway, just be happy if it's fairly clean and there's clean water from tap etc.
On the topic of hygiene, what's so clean about partially wiping/smearing your shit off with toilet paper? How effective is that? Why not wash your shit off thoroughly with water and soap, or use whatever it is you consider good enough to wash your hands clean _enough_ (go check the CDC's recommendation - AFAIK 60-90% alcohol solutions work well at killing stuff, but are not so good for removing stuff ). There is no 100% clean in most houses/places, and your hands will get dirty the instant you touch something anyway, you're just going to have to live with it - or die trying
As for what Other People do, why be so bothered on whether they leave the lid up or down? How about whether they washed their hands properly before they touched other stuff, and whether they left traces of shit on the tap knobs or elsewhere. I'm not immune compromised so even if that thought may sometimes seem gross, I'm not going to let it gross me out a lot.
That's only true for really foul-mouthed people.
Yeah, call 911. Or 112, or whatever.
;).
Also, the dog could choose to die in other different ways
The last I heard, the US Gov will have access to X bits of the Lotus Notes keys (some of the keys bits are taken and encrypted to the US Gov key), so that they get a significant help to cracking stuff if they need to. Something like it's 40bit crypto for the US Gov, and 64 bit crypto for everyone else (other than the intended recipients).
You said: "Why on hell? For one, it wouldn't be www.microsoft.com but com/microsoft/www. For second, assuming equivalent to current expansions, it would be "www" the one to expand to microsoft/www or com/microsoft/www, exactly like now. I really don't see where the "prefixes for alternate searchs" in your example come from."
Are you for real? I gave the example that way because all the following URLs are expressed in the same way in the "New Berners" approach.
All these "conventional URLs":
http://com:82/microsoft/www/foo/bar/com/baz/
http://microsoft.com:82/www/foo/bar/com/baz/
http://www.microsoft.com:82/foo/bar/com/baz/
http://foo.www.microsoft.com:82/bar/com/baz/
http://bar.foo.www.microsoft.com:82/com/baz/
http://com.bar.foo.www.microsoft.com:82/baz/
http://baz.com.bar.foo.www.microsoft.com:82/
Are all represented in the "New Berners" approach in a single URL as:
http/82:com/microsoft/www/foo/bar/com/baz
Pasting 7 identical lines (ala new berners URL) to try to illustrate 7 possible different browser "attempts" (whether in parallel or not) would be silly. So that's why I stuck to the conventional (in fact I skipped some details and assumed basic knowledge of HTTP and TCP, as well as DNS, which may be the problem).
So even if it's just a single "New Berners" URL the browser still has to _potentially_ try ALL the possible _effective_ combinations in event of failures to fetch the "effective urls" - whether due to DNS failure, or failure to reach/connect to the webserver. Sure you might be able to avoid/workaround some of that by doing various stuff. But why do all that? What would people gain?
"What do you think is current DNS server topology? It already is a hierarchy of hosts, so no news here too"
While the DNS topology is a hierachy of hosts, the current WWW is NOT - it is a _Web_ of hosts. Right now once webmasters get their domain name and IP, they have a lot of freedom in the pathnames and filenames they can use. They will lose that freedom with the "new approach". Instead of just making a directory and putting a file somewhere, they'd have to do a few more dependency checks - e.g. does the new directory or file cause a clash with someone else's subdomain (past/present/planned). More work and cooperation required.
You're in a way proposing breaking the Web- forcing it into a hierachy (in fact tying it tightly to the DNS hierachy) and then using "SRV-like" stuff to glue the broken bits together. I repeat: what would people gain from that? Is it worth it?
Sorry, I'm giving up already, it's too much work explaining the simple stuff already (I'm a lazy person). If it makes you happy, go ahead and assume I'm wrong by default.
Would be efficient distributed locking and messaging at an O/S level that supports clustering too.
;)
Just like the O/Ses of those good old days
Running stuff in parallel is trivial. Just get another core/processer/computer, run the process. The issue is when the processes need to exchange information or they need to be serialized.
Seems a lot of people nowadays delegate that stuff to some DB software (since DB software needs locking, serialization etc too, and provides it), but DB software has typically higher overheads and also a fair number of DBs have problems running across clusters too...
If you design the APIs well programmers will start using multiple cores, and even multiple computers. It's not necessarily an easy problem, but it's been done more than once already, I'm sure with the benefit of hindsight, some hardware support from say Intel, things should be even better nowadays.
I may be ignorant, but I'm not that ignorant. I know DNS and how it works. And I also know how lots of other stuff works too.
You're wrong because it's not just about DNS. It's about how things work in real life and equally importantly it's about how things _FAIL_ in real life.
TCP connections still take time to make, and there is a significant timeout if the other end is firewalled (ignores TCP/80).
Try this on a linux/*BSD box: time wget http://www.microsoft.com:82/foo/bar/com/baz
See how long that takes to time out. Sure interactive browsers might time out faster, but unlike DNS, they won't timeout after just a few seconds.
And remember after that times out, with the "New Berners" approach you will have to try to fetch:
http://foo.www.microsoft.com:82/bar/com/baz
http://bar.foo.www.microsoft.com:82/com/baz
http://com.bar.foo.www.microsoft.com:82/baz
http://baz.com.bar.foo.www.microsoft.com:82/
And only after all that should the browser give up.
There's also the scenario of trying to access a site that hosts lots of different people's stuff that uses a wildcarded DNS- say the dns works but the site is down - how long do you wait? All of the possible domains will work (you expect the admin to set up a system to put all the valid names of all the sites in the DNS? Esp when previously the customers got to have their own arbitrary sub domain names without any change to the configs).
OK lets say you try to do stuff in parallel, and display the first document that is successfully fetched. But what happens then if you get multiple documents? If a server higher up the hierachy (thus being more heavily loaded and more likely to be "slow") finally responds while the browser is _halfway_ displaying a different file should you suddenly tell the user "Ooops, pretend you didn't see that, here's what you should be seeing". What if you get tons of HTTP/404s? Which 404 should you show? There are pretty fancy 404s nowadays. You think that's bad? What if you get multiple HTTP/302s! Which 302 should you follow? All of them? And risk the problem getting even bigger?
And should the browser do negative caching for all failures? How long?
Sure you can put "don't recurse" stuff on the DNS servers, but in real life, the people who run the webservers often have little authority and control over the DNS servers.
Run the DNS server on the webserver? Despite what some people may like, not every web server be allowed to run an authoritative DNS server on it AND get the firewall administrators to pass DNS traffic to their pet server, nor is it likely that the DNS delegation be correctly done in enough cases for people to say "this system is viable".
Lastly, the "New Berners" approach is trying to merge file namespaces with host namespaces, sure that could work fine in scenarios where one entity controls everything. BUT with this approach you will no longer have an "every host is a peer" situation, it will be a hierachy of hosts. Some hosts WILL override other hosts so you can no longer put stuff in certain namespaces on those hosts. You will start to need cooperation amongst previously independent hosts/peers to avoid undesirable namespace clashes at what was previously a _file_ naming level.
"No more latencies than currently"
You still sure about that?
If the cellphone companies don't want to pay they can stop providing cellular coverage for that entire area.
:). So in my opinion, the max a country should pay compensation to a peoples for past evils would be at _maximum_ 3 generations (of course if the country still doesn't think it did anything wrong in the first place, you better be careful).
Sounds reasonable to me. Let's see how well their area does if nobody provides coverage to them.
Of course if they really want to be a sovereign nation and start making rules like that, then it's time they stopped collecting "pocket money" from "Step-mommy" just because "the evil great-great-great-grandstep mom" was bad to them, and started paying their own way.
If a culture/people takes so long to recover from such a set back then something is wrong somewhere.
A not so atypical scenario: First generation immigrant works hard and becomes rich, 2nd generation - maintains the wealth, 3rd generation blows it all away
That's a bad idea.
/example/blah/sub/foo /blah/sub/foo /sub/foo
/foo
With that, you can't tell where a host name begins and where the URI starts.
And he even thinks that is GOOD:
quote: "This would mean the BCS could have one server for the whole site or have one specific to members and the URL wouldn't have to be different."
Doh.
Say you have a conventional URL of http://blah.example.com/sub/foo
If we do things he proposed in that page how does he expect the browser to find the IP address for the server to go to?
With his suggestion the url will look like:
http:com/example/blah/sub/foo
Now that's very nice in "dreamland" where the speed of light is infinite and everything is perfect.
But in the real world, what domain name should the browser try in order to get the IP address to connect to?
Should the browser try to connect to "com" and fetch
Then if that fails connect to:
example.com and try to fetch
Then if that fails connect to:
blah.example.com and try to fetch
AND WORSE, even if that's the correct URL, say the server was temporarily broken/misconfigured, so now the browser is suppose to keep going?
e.g.connect to:
sub.blah.example.com and fetch
then try
foo.sub.blah.example.com
The browser has to wait for the necessary failure timeouts on each try. Don't forget, the URL I used as an example isn't even a very long one. Imagine one with a greater "directory depth".
Make you wonder if he "stumbled" on his _original_ scheme by sheer luck. Or he actually thought long and hard on it, and has now unfortunately forgotten the original reasons why things were done that way.
Nowadays I find there are not very many people who understand how lots of different things work, the various limitations, and how certain choices/changes affect things. There's often so much you need to know AND keep in mind.
I guess the bottled water people are making enough money without scientific studies. USD10 billion in sales?
;)
Maybe they're actually worried about people finding out about the genuine health hazards of drinking too much water...
Flamebait? Someone can't handle the truth? Tsk tsk :).
t e:trac.lighttpd.net+lighttpd+memory+leakr -a-small-vps
;). Need I say more?
"lighttpd leaks memory":
http://www.google.com/search?hl=en&safe=off&q=+si
Random complainer: http://hostingfu.com/article/nginx-vs-lighttpd-fo
As for security: Apache isn't what you'd call secure software, but Lighttpd isn't either.
PHP is slower than Perl or Python for most stuff. This should be common knowledge amongst decent programmers- go find/make your own benchmarks and links.
PHP security?
lighttpd leaks memory, how secure and "lightweight" is that?
So far Apache is good enough for me, but if you're going to replace Apache with something that is fast and has features, you may wish try Zeus.
Anyway, I'd dump PHP first. PHP is slow and has tons of security problems - and judging from the way the devs do stuff, there'll be plenty more to come.