So, having used Symbian phones, I would suggest that what is really happening here is that Symbian are trying to patent a bugfix... The bug being that their phone O/S is painfully slow.
I very much doubt that they have invented something that will make all mobile phones regardless of O/S run faster, unless perhaps we're talking about little robotic legs? That would be a cool patent:)
I distinctly remember seeing a TV program way back (perhaps 10 years ago) where the Japanese carried out a study just like this. They wanted to know why a new car would not sell, and discovered entirely the opposite - The car was unpopular because of its not-smiley face - Which is why a large number of Japanese manufactured vehicles have a smiley face on them. I have noticed this quite regularly ever since.
It could be that drivers have changed in 10 years, it could be that Japanese drivers are just "nicer" and prefer a smiley car, or perhaps it is saying something about how the world is changing as a whole.
On the other hand, perhaps it is just so much statistical bullsh*t... You choose:)
You are entirely entitled to like what they are doing. At the same time, thousands of web admins are entirely entitled to complain about the un-needed extra cost that they are landed with for bandwidth.
What they are doing provides no benefit to the end user - it just slows down the PC by AV scanning links that you will never even think of clicking.
What AVG are doing protects my child from the Internet no more than scanning the link at the point it is clicked, rather than scanning a page full of links that will never be clicked.
AFAIK Every other AV manufacturer stores a pre-scanned list of good/bad URLS at a central location. Why can't AVG do the same rather than piss off web-admins un-necessarily?
Because 99% of AVG installers will not have the slightest clue that they are contributing in a harmful way to Internet traffic volumes - They are just installing the latest version of their free AV product, and is is largely harmless to them.
The user is freely choosing to install a "beneficial" application, one which in many respects is a very functional capable and respected product.
This can hardly be compared to the stealth-install used by trojans and viruses which create DDoS BotNets... Can it?;-)
OTOH, I would love to see a major ISP send AVG a bill for this traffic:)
...And then there is the "general public" definition of broadband, which is usually just "anything that is not dial-up".
This is largely due to a) The fact that they largely neither know nor care what a Mbit/s is b) Have been subject to TV advertising for so long that they just know that they need to "buy a broadband to make the Internet in their PC go fast"
A DNA test is a "closest match" test, and is only right about 99% of the time.
AFAIK, there are several levels of DNA profiling possible, the simplest, quickest and cheapest provide only a one in a few thousand chance of a duplication. The most expensive, complex and slowest can provide a one in several million chance of a false positive.
All of these tests are better than 99% certain because the labs will do more than just let a "closest match" search convict a criminal. They will provide a "decode" of the various samples to the court or the police, and there will be a statement of how certain (or not) the match is.
Whoever is in charge of the Internet will try to enforce their own version of "morality" upon us all.
Right now, the US is refusing to allow ".xxx" domains, and is banning online gambling, the alternatives would be equally unpaletable.
More usefully we would have a system where the "speech" is 100% free on the part of the author, but the system would be enclosed such that any part of the net could refuse or block certain data. China does not want it? Fine. The US doesn't want it? Fine. The UK wants to prosecute someone publishing something from their soil. Fine, the framework needs to develop sure, but banning everything down to the "least-common morality" is going to make the 'net a very quiet place.
Okay pie-in-the sky I agree, but while speech should be free, I reserve the right to wear earplugs. Personally I think that ".xxx" domains would more easily allow us to filter porn from minors, and how does online gambling differ from insuring your car on line:
1) "I bet that the next card is an Ace, here is my credit card number". vs.
2) "I bet I won't have a car crash in the next year, here is my credit card number."
faster booting thanks to upstart (sort of a replacement for init, among others)
I just had a look at "upstart" and some of its configuration documentation, and while I understand "traditional" rc script processes (such as sysvinit, and the variations on that) I cannot see how upstart will speed anything up. It still seems to be a serialised startup process, and the documentation does not make it clear how to specify startup dependencies ("IP before NTP", or "spamd before sendmail"), so there is no implied optimisation behind-the-scenes by using parallel startup.
Have I missed something, or is this just a move to an event-driven RC process "because I can" ?
So if a car is in traffic travelling at 3mph, I am jogging past at 5mph, and a motorbike rider drives past at 50mph, all parties have a mobile phone... What is the speed?
I can't see it working until lots of people subscribe, and lots of people will not subscribe to inaccurate data......will they?
What I actually meant by European was European:). I have only experienced US and UK/European systems, so I was intentionally limiting the scope of the comment.
You probably don't realise it but that is also a "local" feature. In Europe you can use either, so that "My birthday is on December [the] 20th" or "Christmas day is on the 25th of December" are both equally valid.
And why not? They both convey the same information perfectly clearly.
I look at it differently again, so perhaps we have to agree to disagree...
2006/8/2 is I agree then "best" way to represent a date - The majority of human-use number systems put the least-significant information to the right hand side. This has the bonus that sorting on a computer (mechanical or electrical) is simplified. Systems that include this are HH:MM:SS and good 'ol decimal numbering Hundreds-Tens-Units.
An alternative in the LSB/MSB world is 2/8/2006 - Computers can be (and are) architected to deal with this type of reversal. Humans can deal with the LSB being at the left hand end of the information. This is still "useful" as processing left-to-right and right-to-left are not really that different. I consider this to be "next best" as it retains a certain amount of logic.
The final option is to ignore whether your data has an order or significance, and just jumble it up. How is MM/DD/YYYY differnt to YYYY/DD/MM ? Would you consider YYYY/DD/MM valid or useful? I am afraid that I personally would not, and but the same token consider MM/DD/YYYY to be not-useful. Perhaps we should just switch to MM/YYYY/DD for fun?:)
I would be seriously interested to know where/why the different system in the USA originated, and by what measure you determine the USA system to be next-best and the European system to be "Fucked up"
Well, since you ask, personally I would grab a copy of [Favourite free O/S] and install it in place of [Monopolistic evil O/S]... *SLAP* sorry, wrong thread:)
Although they are improving dramatically with each release, I still believe that Windows installs go "mouldy" - They get slower, eat more memory and become generally less stable with time. The latest set of Windows versions does this quite slowly, unlike Windows 95/98/ME, where you could almost see it happen.
The reason for mentioning this is that if you buy a Windows XP upgrade CD, you will be able to "refresh" the mouldyness, and provide a new CD key in one go by upgrading your existing XP to the CD XP - This will probably then need a whole slew of updates downloading, but the end result is often noticably smoother running.
FYI In the USA it is possibly legal to track which customer is using which IP, but it is certainly not required.
It would be unconstitutional to require that type of violence against an American's privacy without a warrant. Not that the law has much affect on what actually happens in the USA anymore. But there are still ISPs that use 20 minute DHCP expires as a selling point for privacy reasons.
Interesting - While I understand that privacy is important, I am happy for information to be stored regarding my actions where those actions are "in plain sight". I do not want that information spread without due process, but it could be in my interest to have an alibi provided by my ISP. It may also be useful to be able to convict someone of spamming, phishing, child pornography or whatever through the use of this type of evidence.
The UK government also behaves differently in other respects towards privacy, and although there are some HORRIBLE exceptions where guilt is assumed (***GRRR***), I have no objections to city-centre video surveillance etc etc... Numerous criminals are caught through this type of "privacy invasion"
...Or are you doing something illegal???;-) If I were looking for criminals, the first place to go would be those 20 minute DHCP lease companies!!!
Sadly, some mail clients support signed and encrypted emails really badly (or not at-all). I have seen more than one installation of Outlook Express where, if a signed message is sent to them, you have to click extra buttons before it can be read, and you cannot reply-to or forward a message for some strange reason - I never did work out why.
Sadly Outlook Express still has a huge end-user following as it is familiar, and comes-with "that" operating system. Using POP3 mailboxes means that migrating between mailers is often painful, so we are stuck with incapable clients and Phishers are free to play.:-(
At worst it's simply the IP address and at best surely it could only be a list of IP addresses and what they uploaded - i.e, IP address xxx.xxx.xxx.xxx uploaded (file) on (date and time) to (server).
ISPs (at least in th UK, and I think in the USA and elsewhere) are required to log which customer was using which IP address at what time for tracability reasons. This means that the information above would quite easily identify a household, and I imagine it is a short jump from there with appropriate legal warrants, to access to the computers in that household.
This is now an old thread, but in case you are interested, my point is that i DO see LOTS of corporate websites spewing java stack traces up the screen, demonstrating an apparent screwup embedding java into webservers.:)
For a very long time I was in the "needs more speed and more stability" camp when it came to Java. Then a colleague of mine wrote a performance test suite in Java to load-test SMTP, HTTP, POP etc... And then wrote equivalents in Perl and Python. All 3 were equally maitainable, and we stuck with the Java version because it had the best performance running multiple threads/processes (memory settings DID need tweaking).
I _Believe_ that the real problems with Java occur when adding either a Graphical interface (it tries to be all things to all OSes, and fails miserably) or embedding it into a webserver (I assume we see crashes because the API is not sufficiently open/documented)
So an open source license may well help in some respects - Better API visibility and transparency, and the option to switch to an open-source graphical component that has been beaten on and fixed, and throw out the umpty-three existing graphical widget sets.
I used all 3 of these OS'es a while back in a datacentre. In those days (about 3 years ago) there was a concensus among many people that I worked with that:
OpenBSD (2.7) = More secure due to better code reviews - Good for firewalls and gateways FreeBSD (4.8) = Better more efficient network stack - Good for webservers etc. Linux (RH 6.2) = Good alrounder - Good choice for desktop and for a much wider choice of prebuilt applications. Also OS du jour at the time.
I would be very interested to see a good modern comparison of these OS'es, perhaps even with commercial *nix thrown into the analysis - HP/UX, AIX, Solaris and SCO for example.
I bet they still all have their strengths and weaknesses now, just like they did then.
Slashdot Mirror
So is an extension of this that fat people carry more weight, get tired more quickly, and therefore pant and generate more CO2.
Their additional weight also wears sidewalks more quickly...
Should we tax fat people under the same rules?
So, having used Symbian phones, I would suggest that what is really happening here is that Symbian are trying to patent a bugfix... The bug being that their phone O/S is painfully slow.
I very much doubt that they have invented something that will make all mobile phones regardless of O/S run faster, unless perhaps we're talking about little robotic legs? That would be a cool patent :)
I distinctly remember seeing a TV program way back (perhaps 10 years ago) where the Japanese carried out a study just like this. They wanted to know why a new car would not sell, and discovered entirely the opposite - The car was unpopular because of its not-smiley face - Which is why a large number of Japanese manufactured vehicles have a smiley face on them. I have noticed this quite regularly ever since.
It could be that drivers have changed in 10 years, it could be that Japanese drivers are just "nicer" and prefer a smiley car, or perhaps it is saying something about how the world is changing as a whole.
On the other hand, perhaps it is just so much statistical bullsh*t... You choose :)
How insightfully put :)
You are entirely entitled to like what they are doing. At the same time, thousands of web admins are entirely entitled to complain about the un-needed extra cost that they are landed with for bandwidth.
What they are doing provides no benefit to the end user - it just slows down the PC by AV scanning links that you will never even think of clicking.
Sounds pretty dumb to me.
What AVG are doing protects my child from the Internet no more than scanning the link at the point it is clicked, rather than scanning a page full of links that will never be clicked.
AFAIK Every other AV manufacturer stores a pre-scanned list of good/bad URLS at a central location. Why can't AVG do the same rather than piss off web-admins un-necessarily?
Just my 2p.
They knew they were installing AVG, unlike the malware! That make it a very big difference IMHO...
Because 99% of AVG installers will not have the slightest clue that they are contributing in a harmful way to Internet traffic volumes - They are just installing the latest version of their free AV product, and is is largely harmless to them.
The user is freely choosing to install a "beneficial" application, one which in many respects is a very functional capable and respected product.
This can hardly be compared to the stealth-install used by trojans and viruses which create DDoS BotNets... Can it? ;-)
OTOH, I would love to see a major ISP send AVG a bill for this traffic :)
Dumb is what they were BEFORE they were told about the problem. Slimy is what they are now that they are refusing to rectify the situation and behave.
I think they deserve everything they will inevitably get as a result of this.
...And then there is the "general public" definition of broadband, which is usually just "anything that is not dial-up".
This is largely due to
a) The fact that they largely neither know nor care what a Mbit/s is
b) Have been subject to TV advertising for so long that they just know that they need to "buy a broadband to make the Internet in their PC go fast"
Do MS buy up a bunch of Yahoos, which suits my opinion of their company already...
They both seem to fit quite well
...Don't encourage them by implying that DRM can be licensed and is a legitimate option.
:)
DRM is bad bad bad, and is broken whether licensed or not. Don't use it, that's the answer
A DNA test is a "closest match" test, and is only right about 99% of the time.
:)
AFAIK, there are several levels of DNA profiling possible, the simplest, quickest and cheapest provide only a one in a few thousand chance of a duplication. The most expensive, complex and slowest can provide a one in several million chance of a false positive.
All of these tests are better than 99% certain because the labs will do more than just let a "closest match" search convict a criminal. They will provide a "decode" of the various samples to the court or the police, and there will be a statement of how certain (or not) the match is.
For the record, IANA Forensic Scientist
Whoever is in charge of the Internet will try to enforce their own version of "morality" upon us all.
Right now, the US is refusing to allow ".xxx" domains, and is banning online gambling, the alternatives would be equally unpaletable.
More usefully we would have a system where the "speech" is 100% free on the part of the author, but the system would be enclosed such that any part of the net could refuse or block certain data. China does not want it? Fine. The US doesn't want it? Fine. The UK wants to prosecute someone publishing something from their soil. Fine, the framework needs to develop sure, but banning everything down to the "least-common morality" is going to make the 'net a very quiet place.
Okay pie-in-the sky I agree, but while speech should be free, I reserve the right to wear earplugs. Personally I think that ".xxx" domains would more easily allow us to filter porn from minors, and how does online gambling differ from insuring your car on line:
1) "I bet that the next card is an Ace, here is my credit card number".
vs.
2) "I bet I won't have a car crash in the next year, here is my credit card number."
Ho hum. Governments... Pah!
faster booting thanks to upstart (sort of a replacement for init, among others)
I just had a look at "upstart" and some of its configuration documentation, and while I understand "traditional" rc script processes (such as sysvinit, and the variations on that) I cannot see how upstart will speed anything up. It still seems to be a serialised startup process, and the documentation does not make it clear how to specify startup dependencies ("IP before NTP", or "spamd before sendmail"), so there is no implied optimisation behind-the-scenes by using parallel startup.
Have I missed something, or is this just a move to an event-driven RC process "because I can" ?
So if a car is in traffic travelling at 3mph, I am jogging past at 5mph, and a motorbike rider drives past at 50mph, all parties have a mobile phone... What is the speed?
...will they?
I can't see it working until lots of people subscribe, and lots of people will not subscribe to inaccurate data...
What I actually meant by European was European :). I have only experienced US and UK/European systems, so I was intentionally limiting the scope of the comment.
No harm meant...
You probably don't realise it but that is also a "local" feature. In Europe you can use either, so that "My birthday is on December [the] 20th" or "Christmas day is on the 25th of December" are both equally valid.
:)
And why not? They both convey the same information perfectly clearly.
Like I say, agree to disagree
I look at it differently again, so perhaps we have to agree to disagree...
:)
2006/8/2 is I agree then "best" way to represent a date - The majority of human-use number systems put the least-significant information to the right hand side. This has the bonus that sorting on a computer (mechanical or electrical) is simplified. Systems that include this are HH:MM:SS and good 'ol decimal numbering Hundreds-Tens-Units.
An alternative in the LSB/MSB world is 2/8/2006 - Computers can be (and are) architected to deal with this type of reversal. Humans can deal with the LSB being at the left hand end of the information. This is still "useful" as processing left-to-right and right-to-left are not really that different. I consider this to be "next best" as it retains a certain amount of logic.
The final option is to ignore whether your data has an order or significance, and just jumble it up. How is MM/DD/YYYY differnt to YYYY/DD/MM ? Would you consider YYYY/DD/MM valid or useful? I am afraid that I personally would not, and but the same token consider MM/DD/YYYY to be not-useful. Perhaps we should just switch to MM/YYYY/DD for fun?
I would be seriously interested to know where/why the different system in the USA originated, and by what measure you determine the USA system to be next-best and the European system to be "Fucked up"
Well, since you ask, personally I would grab a copy of [Favourite free O/S] and install it in place of [Monopolistic evil O/S]... *SLAP* sorry, wrong thread :)
Although they are improving dramatically with each release, I still believe that Windows installs go "mouldy" - They get slower, eat more memory and become generally less stable with time. The latest set of Windows versions does this quite slowly, unlike Windows 95/98/ME, where you could almost see it happen.
The reason for mentioning this is that if you buy a Windows XP upgrade CD, you will be able to "refresh" the mouldyness, and provide a new CD key in one go by upgrading your existing XP to the CD XP - This will probably then need a whole slew of updates downloading, but the end result is often noticably smoother running.
Just my 2p
FYI In the USA it is possibly legal to track which customer is using which IP, but it is certainly not required.
It would be unconstitutional to require that type of violence against an American's privacy without a warrant. Not that the law has much affect on what actually happens in the USA anymore. But there are still ISPs that use 20 minute DHCP expires as a selling point for privacy reasons.
Interesting - While I understand that privacy is important, I am happy for information to be stored regarding my actions where those actions are "in plain sight". I do not want that information spread without due process, but it could be in my interest to have an alibi provided by my ISP. It may also be useful to be able to convict someone of spamming, phishing, child pornography or whatever through the use of this type of evidence.
The UK government also behaves differently in other respects towards privacy, and although there are some HORRIBLE exceptions where guilt is assumed (***GRRR***), I have no objections to city-centre video surveillance etc etc... Numerous criminals are caught through this type of "privacy invasion"
Sadly, some mail clients support signed and encrypted emails really badly (or not at-all). I have seen more than one installation of Outlook Express where, if a signed message is sent to them, you have to click extra buttons before it can be read, and you cannot reply-to or forward a message for some strange reason - I never did work out why.
:-(
Sadly Outlook Express still has a huge end-user following as it is familiar, and comes-with "that" operating system. Using POP3 mailboxes means that migrating between mailers is often painful, so we are stuck with incapable clients and Phishers are free to play.
At worst it's simply the IP address and at best surely it could only be a list of IP addresses and what they uploaded - i.e, IP address xxx.xxx.xxx.xxx uploaded (file) on (date and time) to (server).
ISPs (at least in th UK, and I think in the USA and elsewhere) are required to log which customer was using which IP address at what time for tracability reasons. This means that the information above would quite easily identify a household, and I imagine it is a short jump from there with appropriate legal warrants, to access to the computers in that household.
This is now an old thread, but in case you are interested, my point is that i DO see LOTS of corporate websites spewing java stack traces up the screen, demonstrating an apparent screwup embedding java into webservers. :)
Perhaps I am just unlucky.
For a very long time I was in the "needs more speed and more stability" camp when it came to Java. Then a colleague of mine wrote a performance test suite in Java to load-test SMTP, HTTP, POP etc... And then wrote equivalents in Perl and Python. All 3 were equally maitainable, and we stuck with the Java version because it had the best performance running multiple threads/processes (memory settings DID need tweaking).
I _Believe_ that the real problems with Java occur when adding either a Graphical interface (it tries to be all things to all OSes, and fails miserably) or embedding it into a webserver (I assume we see crashes because the API is not sufficiently open/documented)
So an open source license may well help in some respects - Better API visibility and transparency, and the option to switch to an open-source graphical component that has been beaten on and fixed, and throw out the umpty-three existing graphical widget sets.
Just my 2c.
I used all 3 of these OS'es a while back in a datacentre. In those days (about 3 years ago) there was a concensus among many people that I worked with that:
OpenBSD (2.7) = More secure due to better code reviews - Good for firewalls and gateways
FreeBSD (4.8) = Better more efficient network stack - Good for webservers etc.
Linux (RH 6.2) = Good alrounder - Good choice for desktop and for a much wider choice of prebuilt applications. Also OS du jour at the time.
I would be very interested to see a good modern comparison of these OS'es, perhaps even with commercial *nix thrown into the analysis - HP/UX, AIX, Solaris and SCO for example.
I bet they still all have their strengths and weaknesses now, just like they did then.