Whether it's open source or not doesn't mean much with respect to security.
If you can't or don't want to do an audit of the source, it's usually safe to assume it's probably just as bad as whatever software the Mozilla programmers used to write.
A good tree produces good fruit. A bad tree produces bad fruit. Sure you can get a tree to change, but it often takes years (see BIND, Sendmail).
The fact that Mozilla crashes regularly (but not so predictably) on normal use (well at least my normal use;) ) means there are a fair number of not that obscure security bugs left to be found.
If it crashed predictably in only a few scenarios, then there are significantly fewer obvious bugs left to fix.
I've already said something similar about Mozilla before. (it's modded -1 for some reason;) ).
Anyway it looks like I've been vindicated. Within 2 weeks too;).
I used to work in the IT security line. I think I've learnt at least a thing or two...
Here's a free tip: Run your browser as a different user. You can do that in Linux/*BSD and Windows XP/2000. Use google and figure out the implementation details.
Your browser data will still be vulnerable to attackers, but if you set up the permissions correctly, the data and documents owned by your normal user account will be safe.
If you are paranoid you can also use a different browser user for different security levels/realms (e.g. mozbank for banking, mozprivate for other important stuff, moznormal for normal sites, mozother for unknown sites ).
If you do that try to have different browser schemes/colours too. That way it's easier to know when you're using the wrong browser for the wrong site;).
Also may wish to do stuff like: have the bank browser's homepage as the bank site, and have it only allow javascript etc to be on for the bank URLs - other urls just don't get javascript and other stuff active.
Sure it's a bit of work to set it up, but once you've got it setup it's not so hard - think of the different browsers as separate apps.
Also make your GUI preferences a slightly different colour from the default. So if a attacker plants a trick window, hopefully it won't match and you'll notice.
This won't protect you from GUI overflow bugs or "shatter" attacks in the case of Windows. But I believe hackers will find attacking the other users easier.
I personally use a virtual machine to browse unknown sites, or untrusted sites that require javascript/flash. Just a click of a button and the virtual machine is back to a known state.
That said I've managed to crash vmware. Just a DoS - (the vmware people that I emailed said they didn't think it's exploitable). When I've more time maybe I'll go check it out.
Artificial Intelligence is already plentiful
on
The Baby Bootstrap?
·
· Score: 1
Given the bias of the site if that's all the dirt they can dig up, Google must be a pretty good company, and/or the people at that site are just crap at digging up dirt.
Think about it, if someone really hated any of the Fortune 500 companies and bothered to dig up some dirt, there'd be tons more dirt.
I suppose Google is a young company. Give it a few more years and more parasites would have found their way into Google. Then you'd have a lot more dirt.
a) Don't use Google. b) Use a different anonymizing proxy for _each_ single search, preferably using SSL. c) Assume your searches AND non-encrypted web requests aren't anonymous and secure.
If I were running the NSA or some other spook agency, I'd tap the pipes leading to Google (and a few other sites too).
Same if I were a dubious org/agency.
Lots of finance institutions/orgs/ppl get the bulk of their info from just a few sources e.g. Bloomberg. So if Bloomberg gets/sends the bulk of their info down just a few pipes...;)
At least for my queries. So prefetching the first result is a bit silly in my case.
I do pretty specific queries.
Nowadays my searches seem to turn up tons of mailing lists with the same messages.
If I wanted mailing lists, I'd do the search in google groups...
That said, the mailing lists sometimes don't show up in google groups...
I think Page Rank is starting to fail. I'm not surprised actually. I'm actually surprised it worked for so long. Probably all that tweaking by the Google folk is keeping it moderately useful.
Teoma is producing less redundant results in comparison. Though it's less comprehensive compared to Google.
Of course, sometimes there just isn't any answer on the Web...
If you only have very short bursts of high dissipation/charging then it should be possible to combine capacitors with batteries. The batteries provide the long term "slow" capacity, and the capacitors absorb the spikes without wasting them. Wonder whether if it's really worth the complexity to combine them since capacitors don't really store very much...
If you're talking about cars, they're talking about hybrids, not pure electrics. I don't see the Joe Public charging his electric car from house mains for the reasons you state.
With hybrids (or fuel cell cars) charging to cruise for 600 miles is just loading up with fuel at the nearest petrol/gas/energy station. One litre of petrol = about 34Mjoules. 50 litres = 1.7GJ. One litre of diesel = 38MJ. (Same figures for hydrocarbon fuel cell cars, but different for hydrogen fuel cell cars).
Where I see the quick charging of the battery is important is for regenerative braking (and for charging the battery up between "sprints").
When you brake a 1 ton vehicle from 100kph to a stop, you dissipate the kinetic energy. e.g 0.5 x 1000kg x 27.7^2 m^2/s^2 joules= 380kJ in just a few seconds of braking time.
Assuming you only need the battery to supply 30kW extra for 2 minutes for "sprints" (or silent/battery only operation). The extra kW is on top of what the hybrid diesel or fuel cell banks can produce. That's 3.6MJ of battery capacity.
A one minute recharge rate to 80% (2.88MJ) = 48kJ a second. This means the fastest you can brake without any _extra_ wastage of energy is 380/48 seconds * some fudge factor (conversion, friction etc). If fudge factor = 1 it's about 8 seconds to go from 100 to zero. By extra wastage, I mean the dissipation of energy to brake disks or resistors because the battery can't absorb energy that fast.
1 battery cell is about 8kJ. 3.6MJ is about 450 of those batteries = 6-7kg? Assuming a battery cell is 14 to 15 grams.
If you are willing to carry around 13kg of batteries, you can brake from 100kph in 4 seconds and without throwing away "overflow" energy. However how much would 13kg of batteries cost? - assuming USD5 per battery that's about USD5000.
Also: 6-13kg of batteries gives you 10-20 decelerations from 100kph worth of storage - ( after that everything probably has to be burnt off as heat). Or potential energy of about 300-700 metres of height assuming a 1000kg car - which is not that bad.
The figures are of course quite inaccurate. Just to give an idea of things. At least my idea;).
We'd probably use hydrocarbons if someone figures out a suitable hydrocarbons fuel cell (that lasts and produces enough power etc) and a suitable catalyst+filter (that protects the fuel cell from nasties).
Problem is it may not be that efficient to store nuclear/solar energy in hydrocarbons. Might be more efficient to store it as hydrogen.
I had an Apple IIGS, and the 3.5 inch disk drive had an eject button, as far as I know, the eject button worked (unless the drive was plugged in to a Mac, I guess;) ).
So not all Apple systems were like that:).
That said, I'm not sure if the eject button worked when in GS/OS, can't really remember - it was a long time ago. Anyone remember?
Anyway it really is silly to not have it at least tell the O/S you want to eject the media, so it can help you do what you want.
Personally the one button mouse sucks too. Most people can figure out their index finger from their middle finger, but double-clicking is hard.
Without double clicking and keyboard keys, how are you going to allow selection and activation?
"Any modern OS that uses caching locks the drive when it's mounted anyway, so what's the point of having a button?"
To request the modern O/S to dismount the drive and eject the media, if possible?
Doh. I actually like those old CDROM drives with play, stop, FF buttons and volume controls on the front panel. You can actually select and play different tracks without having to wait for the O/S to launch the darn CD player app.
Also find it annoying to have to dismount USB drives before pulling them out. But I suppose that can't be helped given the design. (OK I know you don't always have to - it just depends on the situation, but it's still annoying).
Let me know how well not having any ports open in Ubuntu/Gentoo/etc Linux works the next time run FireFox as root and click on the wrong link with FireFox and you get an exploit installed on your machine that starts opening up connections from behind your firewalls.
Only the ignorant use IE as an administrator and browse unsafe sites. Sure it's Microsoft which made that as default, but I recall a Linux distro targetted at the masses which did a similar thing. Getting people to learn about sudo and su and normal user vs admin user was a bit too hard I suppose.
If you don't run IE as an administrator, it's not so easy to install a root level exploit.
On Windows XP you can also more easily (compared to W2K) do the equivalent of running IE as another user (which I used to do in my prev office- I don't use XP at home). In which case even if IE gets exploited it's hard for it to affect my important documents - which are owned by a different account from the one that is running IE.
If you want to be safe, I would recommend you do a similar thing on Linux, su to a different user from your normal user account, and then run FireFox. In fact it should be easier to do on Linux than on Windows (it's a bit kludgier and messier on Windows - though it works).
Of course you could also do what I do: I test sites/stuff which might be unsafe by using a vmware virtual machine. If anything happens I just revert to a known snapshot.
That said, there could be bugs in vmware that allow virtualization to break, so if you are really paranoid, use a totally separate machine, stick one of those hardware cards in it that make your harddisk appear to only be temporarily writeable - after a reboot everything reverts to a known state - this is done by some cybercafes. Then copy files over manually;).
Well so far this year most of the problems found with windows are application level problems - buffer overflows and application logic bugs (e.g. with Internet Explorer), and not with the windows "kernel".
In contrast there have been more problems found with the Linux kernel this year. I've had to update the SuSE 9.1 kernel a fair number of times just this year alone.
With the Windows XP SP2 the firewall should be up by default so it's harder to crack than a default install of say RedHat 9 or some other Linux Distro with sshd enabled and accessible.
Sure not all security issues are the same, but so far most of the reported problems with Windows are because Windows users are the sort who would actually try to open an encrypted zipfile from a stranger, enter the password in the email, and proceed to launch the program...
I said: "If you can get a C/C++ program to crash, an attacker can usually get it to run arbitrary code of the attacker's choice."
You said: "Nope. That's usually a sign of a "buffer overflow". "
If you actually had a "basic lesson in security" yourself, you would know that buffer overflows in C/C++ can allow an attacker to execute arbitrary code. Not all, but so far the obvious ones that cause crashes usually do. It's just most people have other things to do than to figure out how to exploit that particular problem, or announce the exploit if they do figure it out.
Smart people use/reuse the wheel when appropriate, because that gives them more time to invent new stuff (hopefully cool stuff for the rest of us to use/reuse).
If you want something like that, just outsource to India:).
The problem is not whether it's in English or something else.
The benefit of experienced programmers is that they know Marketing is going to change their minds and want such and such a feature months later, so they just get ready for it well in advance. Even if it's not in the spec written in English or whatever language.
You can already write code in English. If you write it well enough, cheap programmers can compile it to their programming language of choice. And the main benefit is the cheap programmers can continue maintaining the software, and you can move on to writing other stuff.
It could well be they had complete control and independence over the research and analysis.
But what also matters is the publishing part. Otherwise Microsoft could just sponsor 10 independent researchers. And only let the one favourable study get published.
That way, all the studies are independent, but you still pick the result you want.
In contrast picking the US president is the other way round... The US people get to independently pick the conclusion they want, from 2 pre-chosen results:).
That said, Linux Distros aren't really that secure - esp the desktop configurations - once all the typical desktop stuff is installed. I doubt Mozilla is secure - it's just not been as targetted. Mozilla regularly crashes and exits on me for no apparent reason. If you can get a C/C++ program to crash, an attacker can usually get it to run arbitrary code of the attacker's choice.
Same with OpenOffice. Not very stable even with just normal usage. Microsoft Word hardly crashes in comparison.
However for some reason, the latest fully patched IE seems to crash repeateably on some sites when I drag a link in a browser window and let go within the same window (needs javascript enabled - I only enable javascript for a few sites). I don't recall it doing that before.
However Microsoft has really dropped the ball with XP SP2 becoming vulnerable to the LAND attack. Sure speaks volumes that they allow such a vulnerability to be REINTRODUCED.
The Linux kernel has had a fair number of bugs just this year too.
I'm no AI expert, but it seems a lot of these AI stuff is about trying to find "meanings" of stuff - build from a set of axioms. The other ones try to automatically group stuff.
However I'm not sure those approaches would be good at dealing with "analogies/metaphors".
If I told someone from a few hundred years ago (but reasonably smart) who knows a bit about "cows" and "grass", but nothing about cars and then I tell them: "petrol/gasoline" is to "car" about the same way "grass" is to "cow".
And they'd understand a bit.
After a few more of these analogies, they'd have a more precise understanding of the connections between car and petrol and how similar the _"angles/vectors"_of_the_connections_ are compared to the connections between cow and grass.
They would then find connections with similar "angles" and point them out to me AND thus they might teach me a thing or two about other stuff and what they consume.
Later we might even chuckle mildly with the comparison of low octane and high octane fuel vs hay and grass. And we could spend some time debating how "close/far/different" the are.
Laughter sometimes is the result of realizing a more efficient "compression of data" or "connection/path shortcuts" between things that were not previously connected such.
I don't see groups scaling so well. Worse are the dumb keyword stuff.
It's like the most of the AI bunch are doing Newton suff - with absolute speeds and position (and meanings;) ). But most aren't doing the relativity stuff....
I suspect many overlook that a significant part of meanings can be the "vector" of the connection between stuff and not the stuff itself, or the existence of a connection.
I suppose if you take the absolute approach you could make more and more connections between things and assign types to each of these connections e.g. "eats/consumes". But as I mentioned I suspect this approach may not scale. Because you'd have tons of connections between these things and each of these connections would have many possible types.
Whereas if you have "vectors" linking the stuff and if you have a decent "map of the universe", adding stuff that makes sense could be easier the more accurate your "map" becomes/is.
(perhaps a flash of insight/humour is the realigning of a part of the map).
In contrast adding stuff that doesn't make sense is just memorization.
I don't see how you get the figure for about 100 neurons firing.
Even if the path is only 12 neurons deep (and the signal only goes that far in 1 second), a typical neuron has 1000 synaptic connections with other neurons. Assuming only 0.25 of the connections fire 250^12 is still quite a big number.
AFAIK the brain appears to have neurons for everything - down to like bunches of neurons that fire if you see lines at a particular angle. And probably bunches of neurons that fire if they detect particular bunches of those angle neurons firing (e.g. looking at a mesh).
It's probably simulating the whole perceivable world with neurons, and perhaps consciousness is when it recursively simulates itself over and over.
BTW there's buffering. Stuff goes into the buffer. That's pretty quick. You only recognize it later (that is if something bothers to bother to that bit that bothers you to recognize it at all;) ).
It's like when you hear something whilst paying attention to something else. You may not understand it, till you bother to examine your "audio buffer".
If it really matters that much, maybe you should be going pure digital all the way to each speaker.
And each speaker would have its own amplifier built-in or very close to built-in.
If you want to be extreme, you should be reading the position of the speaker cone optically (or some other way - using something like gray code) to 24 bit resolution and at >100K samples a second, and have a very high powered amplifier shove the cone (very very very fast) to the exact position which the digital signal says it should be. Of course you'd probably want to be able to control the volume but I don't think it is that hard.
Even though there'll be square waves and some overshooting due to all the shoving here and there, at > 100K samples a second (or higher) your _ears_ become the low pass filter for the "DAC".
At the amounts which some HiFi enthusiasts are willing to spend, I'm surprised no one has done this yet (not that I know of anyway - I suggested this more than a decade ago). I don't think it's impossible, just difficult. Should be even easier nowadays.
ROFL. 911 doesn't work here. Not in the UK. Nor in Finland. There the number to dial is 112 (even with locked keypad etc).
Over where I am, you're most likely screwed anyway if the police aren't nearby already. It's unlikely for the police to arrive on time to be useful for anything other than filling out police reports and doing the various redtape stuff. If the cops aren't already around the area, you might as well forget it.
Fortunately the odds of getting murdered where I live are about half that of the USA (and the odds of getting assaulted are way way lower).
See: http://www.nationmaster.com/cat/Crime
And given that the USA probably has MUCH better emergency/trauma medical care, the odds of "_attempted_ murder" occurring are by deduction be a lot higher in the USA.
If someone tries to murder you and the hospital brings you back from the dead, it's still just classified as attempted murder. So if the hospitals get better the murder rates improve.
I think the US should _concentrate_ on lowering their murder and _attempted_ murder rates. Sure 911 service is important, but the USA should get its priorities right.
0.04 murders per 1000 people. That's about the same as Uruguay. USA = 7.70 assaults per 1000 people. Uruguay = 1.34 per 1000 people.
Woohoo. Something is wrong, and your AG spends time suing VoIP providers for not providing a 911 that automatically works for stupid people.
Sure looks like the US hospitals are doing a decent job though;).
BTW either the UK hospitals are doing a FAR better job, or something else makes you 4 times more likely to be killed in the USA than in the UK. (assault rates are about the same, but murder rates are lower in the UK).
Alternative interpretation - being prodded with a little finger (or stuff like that) counts as being assaulted in the US and the UK, whereas the police in Uruguay will ignore or scold you if you try to report that as assault.
Once the standard hardware stops support their DOS stuff they can get/buy the cheapest hardware, and run their app using emulation/virtualization. VMware or something similar - see MAME32 for evidence of old hardware being emulated.
Could even be better = snapshots etc.
AFAIK you can also run many DOS apps on one of those DOS emulators on Linux. Not games. But I'm sure most business apps are OK.
I dare say many plain data entry stuff is fine with DOS.
"Refreshed to keep with the times".
LOL. This IT. Not the fashion industry. As long has they have backups and don't do crazy stuff - like improper power and cooling, they'll be fine.
Old hardware isn't a problem in itself. Crappy faulty hardware is. Whilst some old stuff is crap, lots of new stuff is crap too. In fact, if you have 4 year old hardware that still works within specs, it's likely to work for as long as brand new hardware. Most new stuff fails soon after the warranty;) - so what's left are the "golden oldies".
Slashdot Mirror
Whether it's open source or not doesn't mean much with respect to security.
;) ) means there are a fair number of not that obscure security bugs left to be found.
;) ).
;).
;).
If you can't or don't want to do an audit of the source, it's usually safe to assume it's probably just as bad as whatever software the Mozilla programmers used to write.
A good tree produces good fruit. A bad tree produces bad fruit. Sure you can get a tree to change, but it often takes years (see BIND, Sendmail).
The fact that Mozilla crashes regularly (but not so predictably) on normal use (well at least my normal use
If it crashed predictably in only a few scenarios, then there are significantly fewer obvious bugs left to fix.
I've already said something similar about Mozilla before. (it's modded -1 for some reason
Anyway it looks like I've been vindicated. Within 2 weeks too
I used to work in the IT security line. I think I've learnt at least a thing or two...
Here's a free tip: Run your browser as a different user. You can do that in Linux/*BSD and Windows XP/2000. Use google and figure out the implementation details.
Your browser data will still be vulnerable to attackers, but if you set up the permissions correctly, the data and documents owned by your normal user account will be safe.
If you are paranoid you can also use a different browser user for different security levels/realms (e.g. mozbank for banking, mozprivate for other important stuff, moznormal for normal sites, mozother for unknown sites ).
If you do that try to have different browser schemes/colours too. That way it's easier to know when you're using the wrong browser for the wrong site
Also may wish to do stuff like: have the bank browser's homepage as the bank site, and have it only allow javascript etc to be on for the bank URLs - other urls just don't get javascript and other stuff active.
Sure it's a bit of work to set it up, but once you've got it setup it's not so hard - think of the different browsers as separate apps.
Also make your GUI preferences a slightly different colour from the default. So if a attacker plants a trick window, hopefully it won't match and you'll notice.
This won't protect you from GUI overflow bugs or "shatter" attacks in the case of Windows. But I believe hackers will find attacking the other users easier.
I personally use a virtual machine to browse unknown sites, or untrusted sites that require javascript/flash. Just a click of a button and the virtual machine is back to a known state.
That said I've managed to crash vmware. Just a DoS - (the vmware people that I emailed said they didn't think it's exploitable). When I've more time maybe I'll go check it out.
Look at the typical Slashdot post. Definitely AI.
;)
_Actual_ intelligence would be more useful...
The problem with diesel in the USA is the lack of refuelling stations with diesel.
. htm
Otherwise there'd be diesel hybrids too. Think about it. Diesel engines are good with steady loads. Diesel-electrics are proven technology - e.g. trains see: http://travel.howstuffworks.com/diesel-locomotive
Given the bias of the site if that's all the dirt they can dig up, Google must be a pretty good company, and/or the people at that site are just crap at digging up dirt.
Think about it, if someone really hated any of the Fortune 500 companies and bothered to dig up some dirt, there'd be tons more dirt.
I suppose Google is a young company. Give it a few more years and more parasites would have found their way into Google. Then you'd have a lot more dirt.
a) Don't use Google.
;)
b) Use a different anonymizing proxy for _each_ single search, preferably using SSL.
c) Assume your searches AND non-encrypted web requests aren't anonymous and secure.
If I were running the NSA or some other spook agency, I'd tap the pipes leading to Google (and a few other sites too).
Same if I were a dubious org/agency.
Lots of finance institutions/orgs/ppl get the bulk of their info from just a few sources e.g. Bloomberg. So if Bloomberg gets/sends the bulk of their info down just a few pipes...
At least for my queries. So prefetching the first result is a bit silly in my case.
I do pretty specific queries.
Nowadays my searches seem to turn up tons of mailing lists with the same messages.
If I wanted mailing lists, I'd do the search in google groups...
That said, the mailing lists sometimes don't show up in google groups...
I think Page Rank is starting to fail. I'm not surprised actually. I'm actually surprised it worked for so long. Probably all that tweaking by the Google folk is keeping it moderately useful.
Teoma is producing less redundant results in comparison. Though it's less comprehensive compared to Google.
Of course, sometimes there just isn't any answer on the Web...
If you only have very short bursts of high dissipation/charging then it should be possible to combine capacitors with batteries. The batteries provide the long term "slow" capacity, and the capacitors absorb the spikes without wasting them. Wonder whether if it's really worth the complexity to combine them since capacitors don't really store very much...
If you're talking about cars, they're talking about hybrids, not pure electrics. I don't see the Joe Public charging his electric car from house mains for the reasons you state.
;).
With hybrids (or fuel cell cars) charging to cruise for 600 miles is just loading up with fuel at the nearest petrol/gas/energy station. One litre of petrol = about 34Mjoules. 50 litres = 1.7GJ. One litre of diesel = 38MJ. (Same figures for hydrocarbon fuel cell cars, but different for hydrogen fuel cell cars).
Where I see the quick charging of the battery is important is for regenerative braking (and for charging the battery up between "sprints").
When you brake a 1 ton vehicle from 100kph to a stop, you dissipate the kinetic energy. e.g 0.5 x 1000kg x 27.7^2 m^2/s^2 joules= 380kJ in just a few seconds of braking time.
Assuming you only need the battery to supply 30kW extra for 2 minutes for "sprints" (or silent/battery only operation). The extra kW is on top of what the hybrid diesel or fuel cell banks can produce. That's 3.6MJ of battery capacity.
A one minute recharge rate to 80% (2.88MJ) = 48kJ a second. This means the fastest you can brake without any _extra_ wastage of energy is 380/48 seconds * some fudge factor (conversion, friction etc). If fudge factor = 1 it's about 8 seconds to go from 100 to zero. By extra wastage, I mean the dissipation of energy to brake disks or resistors because the battery can't absorb energy that fast.
1 battery cell is about 8kJ. 3.6MJ is about 450 of those batteries = 6-7kg? Assuming a battery cell is 14 to 15 grams.
If you are willing to carry around 13kg of batteries, you can brake from 100kph in 4 seconds and without throwing away "overflow" energy. However how much would 13kg of batteries cost? - assuming USD5 per battery that's about USD5000.
Also: 6-13kg of batteries gives you 10-20 decelerations from 100kph worth of storage - ( after that everything probably has to be burnt off as heat). Or potential energy of about 300-700 metres of height assuming a 1000kg car - which is not that bad.
The figures are of course quite inaccurate. Just to give an idea of things. At least my idea
We'd probably use hydrocarbons if someone figures out a suitable hydrocarbons fuel cell (that lasts and produces enough power etc) and a suitable catalyst+filter (that protects the fuel cell from nasties).
Problem is it may not be that efficient to store nuclear/solar energy in hydrocarbons. Might be more efficient to store it as hydrogen.
We'll see.
I had an Apple IIGS, and the 3.5 inch disk drive had an eject button, as far as I know, the eject button worked (unless the drive was plugged in to a Mac, I guess ;) ).
:).
So not all Apple systems were like that
That said, I'm not sure if the eject button worked when in GS/OS, can't really remember - it was a long time ago. Anyone remember?
Anyway it really is silly to not have it at least tell the O/S you want to eject the media, so it can help you do what you want.
Personally the one button mouse sucks too. Most people can figure out their index finger from their middle finger, but double-clicking is hard.
Without double clicking and keyboard keys, how are you going to allow selection and activation?
"Any modern OS that uses caching locks the drive when it's mounted anyway, so what's the point of having a button?"
To request the modern O/S to dismount the drive and eject the media, if possible?
Doh. I actually like those old CDROM drives with play, stop, FF buttons and volume controls on the front panel. You can actually select and play different tracks without having to wait for the O/S to launch the darn CD player app.
Also find it annoying to have to dismount USB drives before pulling them out. But I suppose that can't be helped given the design. (OK I know you don't always have to - it just depends on the situation, but it's still annoying).
Let me know how well not having any ports open in Ubuntu/Gentoo/etc Linux works the next time run FireFox as root and click on the wrong link with FireFox and you get an exploit installed on your machine that starts opening up connections from behind your firewalls.
;).
Only the ignorant use IE as an administrator and browse unsafe sites. Sure it's Microsoft which made that as default, but I recall a Linux distro targetted at the masses which did a similar thing. Getting people to learn about sudo and su and normal user vs admin user was a bit too hard I suppose.
If you don't run IE as an administrator, it's not so easy to install a root level exploit.
On Windows XP you can also more easily (compared to W2K) do the equivalent of running IE as another user (which I used to do in my prev office- I don't use XP at home). In which case even if IE gets exploited it's hard for it to affect my important documents - which are owned by a different account from the one that is running IE.
If you want to be safe, I would recommend you do a similar thing on Linux, su to a different user from your normal user account, and then run FireFox. In fact it should be easier to do on Linux than on Windows (it's a bit kludgier and messier on Windows - though it works).
Of course you could also do what I do: I test sites/stuff which might be unsafe by using a vmware virtual machine. If anything happens I just revert to a known snapshot.
That said, there could be bugs in vmware that allow virtualization to break, so if you are really paranoid, use a totally separate machine, stick one of those hardware cards in it that make your harddisk appear to only be temporarily writeable - after a reboot everything reverts to a known state - this is done by some cybercafes. Then copy files over manually
Well so far this year most of the problems found with windows are application level problems - buffer overflows and application logic bugs (e.g. with Internet Explorer), and not with the windows "kernel".
In contrast there have been more problems found with the Linux kernel this year. I've had to update the SuSE 9.1 kernel a fair number of times just this year alone.
With the Windows XP SP2 the firewall should be up by default so it's harder to crack than a default install of say RedHat 9 or some other Linux Distro with sshd enabled and accessible.
Sure not all security issues are the same, but so far most of the reported problems with Windows are because Windows users are the sort who would actually try to open an encrypted zipfile from a stranger, enter the password in the email, and proceed to launch the program...
I said: "If you can get a C/C++ program to crash, an attacker can usually get it to run arbitrary code of the attacker's choice."
You said: "Nope. That's usually a sign of a "buffer overflow". "
If you actually had a "basic lesson in security" yourself, you would know that buffer overflows in C/C++ can allow an attacker to execute arbitrary code. Not all, but so far the obvious ones that cause crashes usually do. It's just most people have other things to do than to figure out how to exploit that particular problem, or announce the exploit if they do figure it out.
Heh, just took a look at it. Whoever is writing the site just sounds jealous to me. The background should probably be green and not red ;).
If that's the worst they can come up about Google then Google is not doing too badly.
Just compare it with the plentiful crappy companies for instance. Or even just an average company.
For laughs, compare it with HP in the days of Ms Fiorina...
Reinvent != reuse.
Using English != reinventing English
Smart people use/reuse the wheel when appropriate, because that gives them more time to invent new stuff (hopefully cool stuff for the rest of us to use/reuse).
If you want something like that, just outsource to India :).
The problem is not whether it's in English or something else.
The benefit of experienced programmers is that they know Marketing is going to change their minds and want such and such a feature months later, so they just get ready for it well in advance. Even if it's not in the spec written in English or whatever language.
You can already write code in English. If you write it well enough, cheap programmers can compile it to their programming language of choice. And the main benefit is the cheap programmers can continue maintaining the software, and you can move on to writing other stuff.
Also, sometimes the merchant pays 2% or something of the transaction for card service.
For some computer stuff, the margins are 5-10%, so 2% is significant.
It could well be they had complete control and independence over the research and analysis.
:).
;).
But what also matters is the publishing part. Otherwise Microsoft could just sponsor 10 independent researchers. And only let the one favourable study get published.
That way, all the studies are independent, but you still pick the result you want.
In contrast picking the US president is the other way round... The US people get to independently pick the conclusion they want, from 2 pre-chosen results
That said, Linux Distros aren't really that secure - esp the desktop configurations - once all the typical desktop stuff is installed. I doubt Mozilla is secure - it's just not been as targetted. Mozilla regularly crashes and exits on me for no apparent reason. If you can get a C/C++ program to crash, an attacker can usually get it to run arbitrary code of the attacker's choice.
Same with OpenOffice. Not very stable even with just normal usage. Microsoft Word hardly crashes in comparison.
However for some reason, the latest fully patched IE seems to crash repeateably on some sites when I drag a link in a browser window and let go within the same window (needs javascript enabled - I only enable javascript for a few sites). I don't recall it doing that before.
However Microsoft has really dropped the ball with XP SP2 becoming vulnerable to the LAND attack. Sure speaks volumes that they allow such a vulnerability to be REINTRODUCED.
The Linux kernel has had a fair number of bugs just this year too.
So they're all crap
I'm no AI expert, but it seems a lot of these AI stuff is about trying to find "meanings" of stuff - build from a set of axioms. The other ones try to automatically group stuff.
;) ). But most aren't doing the relativity stuff....
However I'm not sure those approaches would be good at dealing with "analogies/metaphors".
If I told someone from a few hundred years ago (but reasonably smart) who knows a bit about "cows" and "grass", but nothing about cars and then I tell them: "petrol/gasoline" is to "car" about the same way "grass" is to "cow".
And they'd understand a bit.
After a few more of these analogies, they'd have a more precise understanding of the connections between car and petrol and how similar the _"angles/vectors"_of_the_connections_ are compared to the connections between cow and grass.
They would then find connections with similar "angles" and point them out to me AND thus they might teach me a thing or two about other stuff and what they consume.
Later we might even chuckle mildly with the comparison of low octane and high octane fuel vs hay and grass. And we could spend some time debating how "close/far/different" the are.
Laughter sometimes is the result of realizing a more efficient "compression of data" or "connection/path shortcuts" between things that were not previously connected such.
I don't see groups scaling so well. Worse are the dumb keyword stuff.
It's like the most of the AI bunch are doing Newton suff - with absolute speeds and position (and meanings
I suspect many overlook that a significant part of meanings can be the "vector" of the connection between stuff and not the stuff itself, or the existence of a connection.
I suppose if you take the absolute approach you could make more and more connections between things and assign types to each of these connections e.g. "eats/consumes". But as I mentioned I suspect this approach may not scale. Because you'd have tons of connections between these things and each of these connections would have many possible types.
Whereas if you have "vectors" linking the stuff and if you have a decent "map of the universe", adding stuff that makes sense could be easier the more accurate your "map" becomes/is.
(perhaps a flash of insight/humour is the realigning of a part of the map).
In contrast adding stuff that doesn't make sense is just memorization.
"When I want a swimming buddy and no humans are available, I want a submarine that's been made to swim"
;).
You'd want a submarine? Really?
I'd prefer one of those Japanese fembots... Even if she can't actually swim, she could just wear a swimsuit.
I'm not that fussy
I don't see how you get the figure for about 100 neurons firing.
;) ).
Even if the path is only 12 neurons deep (and the signal only goes that far in 1 second), a typical neuron has 1000 synaptic connections with other neurons. Assuming only 0.25 of the connections fire 250^12 is still quite a big number.
AFAIK the brain appears to have neurons for everything - down to like bunches of neurons that fire if you see lines at a particular angle. And probably bunches of neurons that fire if they detect particular bunches of those angle neurons firing (e.g. looking at a mesh).
It's probably simulating the whole perceivable world with neurons, and perhaps consciousness is when it recursively simulates itself over and over.
BTW there's buffering. Stuff goes into the buffer. That's pretty quick. You only recognize it later (that is if something bothers to bother to that bit that bothers you to recognize it at all
It's like when you hear something whilst paying attention to something else. You may not understand it, till you bother to examine your "audio buffer".
If it really matters that much, maybe you should be going pure digital all the way to each speaker.
And each speaker would have its own amplifier built-in or very close to built-in.
If you want to be extreme, you should be reading the position of the speaker cone optically (or some other way - using something like gray code) to 24 bit resolution and at >100K samples a second, and have a very high powered amplifier shove the cone (very very very fast) to the exact position which the digital signal says it should be. Of course you'd probably want to be able to control the volume but I don't think it is that hard.
Even though there'll be square waves and some overshooting due to all the shoving here and there, at > 100K samples a second (or higher) your _ears_ become the low pass filter for the "DAC".
At the amounts which some HiFi enthusiasts are willing to spend, I'm surprised no one has done this yet (not that I know of anyway - I suggested this more than a decade ago). I don't think it's impossible, just difficult. Should be even easier nowadays.
But what do I know...
You mean here?
g e6 .asp
http://www.firingsquad.com/guides/ps2picture/pa
There seems to be a difference.
ROFL. 911 doesn't work here. Not in the UK. Nor in Finland. There the number to dial is 112 (even with locked keypad etc).
;).
Over where I am, you're most likely screwed anyway if the police aren't nearby already. It's unlikely for the police to arrive on time to be useful for anything other than filling out police reports and doing the various redtape stuff. If the cops aren't already around the area, you might as well forget it.
Fortunately the odds of getting murdered where I live are about half that of the USA (and the odds of getting assaulted are way way lower).
See: http://www.nationmaster.com/cat/Crime
And given that the USA probably has MUCH better emergency/trauma medical care, the odds of "_attempted_ murder" occurring are by deduction be a lot higher in the USA.
If someone tries to murder you and the hospital brings you back from the dead, it's still just classified as attempted murder. So if the hospitals get better the murder rates improve.
I think the US should _concentrate_ on lowering their murder and _attempted_ murder rates. Sure 911 service is important, but the USA should get its priorities right.
0.04 murders per 1000 people. That's about the same as Uruguay. USA = 7.70 assaults per 1000 people. Uruguay = 1.34 per 1000 people.
Woohoo. Something is wrong, and your AG spends time suing VoIP providers for not providing a 911 that automatically works for stupid people.
Sure looks like the US hospitals are doing a decent job though
BTW either the UK hospitals are doing a FAR better job, or something else makes you 4 times more likely to be killed in the USA than in the UK. (assault rates are about the same, but murder rates are lower in the UK).
Alternative interpretation - being prodded with a little finger (or stuff like that) counts as being assaulted in the US and the UK, whereas the police in Uruguay will ignore or scold you if you try to report that as assault.
Uh.
;) - so what's left are the "golden oldies".
Once the standard hardware stops support their DOS stuff they can get/buy the cheapest hardware, and run their app using emulation/virtualization. VMware or something similar - see MAME32 for evidence of old hardware being emulated.
Could even be better = snapshots etc.
AFAIK you can also run many DOS apps on one of those DOS emulators on Linux. Not games. But I'm sure most business apps are OK.
I dare say many plain data entry stuff is fine with DOS.
"Refreshed to keep with the times".
LOL. This IT. Not the fashion industry. As long has they have backups and don't do crazy stuff - like improper power and cooling, they'll be fine.
Old hardware isn't a problem in itself. Crappy faulty hardware is. Whilst some old stuff is crap, lots of new stuff is crap too. In fact, if you have 4 year old hardware that still works within specs, it's likely to work for as long as brand new hardware. Most new stuff fails soon after the warranty
The mind makes it real.