"coverage is 24/7 across all time zones"
Good luck with Oracle 24/7 support in the Middle East. In Qatar, you'll have to wait several days for a guy to get on a plane. Oracle's Linux support offers nothing over Radhat's in this corner of the world.
Almost two years ago, Wikipedia banned users from the State of Qatar from editing anonymously, and from signing up for accounts. At the time, it was dismissed as a storm in a tea-cup by Jim Wales. Many people decided it wasn't a problem because of the small size of the country.
Well, now the rest of the world is catching up with the Middle East in the use of transparent proxies, maybe Wikipedia will try to find some solution to the problem. In Qatar, for the last two years, we have continued to be blocked for months at a time from editing.
I pointed out that this isn't just any omission and addition. When you omit Israel and add Palestine (which is not even recognized as a country by the United Nations) to a country drop-down menu, you seem to be making a very loaded political statement.
Israel was omitted because of fraud from that country, which seems like a good reason. Palestine was probably included in the list because it is recognised by the UN, and is included in ISO 3166-1. If you were to delete Palestine from the list, it would certainly be a very loaded political statement, but its inclusion is not.
> I'm interested to find out how to clean up "incorrect" Geolocation info too.
You can't.
I wrote IP::Country, the library used for country detection of relays within SpamAssassin. There is no manual tweaking of the country codes, I simply take the raw country data from the Regional Internet Registries (ARIN, RIPE, APNIC, LACNIC, AFRINIC) and encode them into a small, portable database.
That is also how all the other IP geolocation libraries work. If you started out from the premise that you'd accept manual changes to the database, it would produce more inaccuracies in the long run that simply following the WHOIS changes.
There is another way, which no one is using. Contraint-based geolocation pings the IP address from several known locations to estimate the location. There are academic papers saying it's more accurate than the WHOIS data, but it's never going to work on a webserver because it's doesn't produce a fast enough answer.
If you're interested in how the country data changes over time, check out the CVS of my data.
The only way to clear up a mistake in my country code is to correct the country code with the Regional Internet Registry. You won't be able to do that unless you own the block.
You should worry more about the disk cache. Previously opened files are cached in RAM in an unencrypted state.
Firewire ports and PCMCIA slots have direct memory access, so can be used to copy an image of your computer's RAM even if no one is logged in. This can recover useful forensic material even after a reboot cycle, as modern BIOS's don't clear RAM.
It looks like Vista's disk encryption is useless if you switch on the PC and access files.
... Qatar has a vastly different government than Saudi Arabia
You are correct in that there is more freedom of expression in Qatar than in Saudi, and the Ministry of Information was abolished in 1996. Unfortunately, QTel has not yet realised that there is no censorship in Qatar.
Before this story appeared on the front page of slashdot, Qataris were also prevented from creating new user accounts - thus preventing anyone from posting.
So, the problem appears to be fixed. Users can now register for accounts. Thank you slashdot front page (Kjkolb);)
The problem came from QTel censorship. Every connection passes through a QTel proxy server, which uses some simplistic rules to determine whether you should be protected from your own surfing habits. If you hit blocked pages too often, your phone rings and when you answer in English you get "I'm sorry, I must have a wrong number. CLUNK." Thus your voice has been recorded for posterity.
The shambles of Qatar's connection might be fixed soon. Q-CERT has just been set up and (hopefully), someone with a bit of influence will be in charge. It is obvious that a single point of failure for an entire population's internet connection is not sensible, but whether this means a better censorship system or the scrapping of censorship remains to be seen.
I'm sorry to interrupt your conspiracy theory with an unfortunate fact, but Lord Hutton isn't a Labour peer, but is a cross-bencher, meaning he's independent.
Rather than farting around with the feedback form (>/dev/null), start the formal complaints process rolling. The Programme Complaints Unit (PCU) deals with serious editorial complaints about all publicly funded BBC services (including the websites). Start by filling in the form.
I would imagine the BBC are currently quite touchy about editorial standards, so you might get a quick response.
Rather than farting around with the feedback form (>/dev/null), start the formal complaints process rolling. The Programme Complaints Unit (PCU) deals with serious editorial complaints about all publicly funded BBC services (including the websites). Start by filling in the form.
I would imagine the BBC are currently quite touchy about editorial standards, so you might get a quick response.
> See Geobutton
Or Geo::IP, IP::Country, javainetlocator or any other free, open source services. You shouldn't be paying $49 for a developers license to read the WHOIS data.
The city data are unreliable. I've posted elsewhere (link1, link2) the reasons why, but will repeat the main points here.
All IP geolocation techniques assume the user of an IP address lives close to the company which registered the address.
The above assumption is mostly true if you define 'close' as 'in the same country'.
In the USA, a lot of people live in (or close to), the same city as the company who registered their IP address. This is because lots of people live in cities and use ISPs which have a presence in that city.
Most of the world population (and US population) doesn't live in a city with major ISP presence. Their city locations aren't going to be accurate by any IP geolocation technique.
The commercial IP geolocation vendors (quova, digital envoy) have a business reason for making city geolocation sound accurate. If it wasn't accurate, why would anybody buy their products?
Conducting a survey that inflates the accuracy of city geolocation is easy - just ensure your survey participants live in a major US cities and you'll achieve high accuracy. One way would be to use server logs from a US financial industry website. Once you have a survey that shows high accuracy, you can sell your product to businesses whose customers don't live in major US cities
It's extremely difficult to measure accuracy of IP geolocation (even at the country level), so if you make bold claims, no one is going to be in a position to argue.
The 'others' are taking american jobs. Quick, let's protect our lifestyles by limiting free trade.
This is simply xenophobia or, worse, racism. Unfortuantely, it's going to increase as the US Presidential Elections draw near, and various politicians try to create enemies they can battle against.
It's not seen as particularly pleasant any more to target the Far Eastern worker for stealing American jobs, so why should it be OK to target Indian programmers. This is a problem with the American economy, not the Indian economy. American programmers are paid too much and limiting free trade is not the answer.
India's sufficiently far away that most Americans will already have a very strange view of the country. And of course, if America says it is saving the Indian savages from 'slave wages', then they should be pleased about the generosity of their great western lords.
Strange that we never hear Americans complaining that the UK is stealing all its financial jobs. But then, it's mostly white people working in the City of London, isn't it.
Suppose both don't agree: then what? Which one do you rely on? If you are an airliner on landing approach, it's not like you can stop in mid-air and ask for directions, you know.
Unfortunately, your logic is flawed. If both don't agree, then at least one is incorrect. This is a signal to the pilot that he should look out of the window rather than relying on his satellite location systems.
Having only one satellite location system does not make that system any more reliable.
The site is slashdotted, so I haven't been able to have a look at it. However, if I were building a geo-search engine, I'd use the WHOIS data for the bulk of the indexing work, and for providing a default location for visitors. The tweaking around the edges (changing the location of the website or page), is just icing on the cake.
No one really knows the accuracy of IP->Country lookup. There's an onlgoing thread on the london perl mongers list about this topic. Some geolocation companies state 98% accuracy, which is pure bullshit. It's more likely to be around 70%, with most of the error occuring in overestimation of US addresses.
By the way, if you want a fast IP locator, here's one that's just as accurate as any of the commercial products. I'm surprised more people don't use this sort of stuff for providing intelligent defaults for their users when filling in HTML forms.
Software written before Java emerged on the scene - software pre-1995 - was all pretty much a hack. Much of the software that is still being written is a hack. People change slowly. Programmers have to die, almost, for this to change. - Bill Joy
The problem with your argument is that both of our drivers follow the JDBC spec, and thus synchronize against the Connection object becuase the MySQL API is not thread safe.
That other drivers do not follow the specification is no relief when ours do, taking the resultant performance hit during multithreaded operation.
In other respects, I agree with you that not many developers use Statements concurrently, but any developer who has read the JDBC spec is entitled to assume that Statements are thread-safe. It would be polite for the database developers to ensure that the underlying libraries are re-entrant. Unfortunately, that is not the case at present.
I'd always thought MySQL was a fast, simple database until I built a Type-II JDBC driver for it.
Because the API does not allow more than one result (MYSQL_RES structure) per connection, and the client libraries are not thread safe by default, any Java classes must be synchronized on the connection. In addition, all rows in a result must be retrieved completely using mysql_store_result rather than the more network efficient mysql_use_result.
The JDBC specification insists (sensibly, in my opinion) that Statement objects be thread safe. The necessary synchronization and use of mysql_store_result severely limits the speed of any mutithreaded application sharing a connection, and probably discounts the speed benefits of MySQL over other free databases.
I would guess the same problems exist with other multithreaded languages using MySQL, and developers should consider these limitations before blindly agrreing with MySQL propaganda that the database if faster than its competitors for running web applications.
ITV Digital is currently in administration, meaning the administrators (Deloitte and Touche) can sell the business without its debts.
Here's the.sig file from their outgoing mail:
ITV Digital plc, 346 Queenstown Rd, London SW8 4DG. Reg No. 3302715:
In administration.
Joint Administrators
N.J. Dargan & N.G. Edwards
The Joint Administrators act as agents of the companies and contract without personal liability.
Slashdot Mirror
"coverage is 24/7 across all time zones" Good luck with Oracle 24/7 support in the Middle East. In Qatar, you'll have to wait several days for a guy to get on a plane. Oracle's Linux support offers nothing over Radhat's in this corner of the world.
Almost two years ago, Wikipedia banned users from the State of Qatar from editing anonymously, and from signing up for accounts. At the time, it was dismissed as a storm in a tea-cup by Jim Wales. Many people decided it wasn't a problem because of the small size of the country.
Well, now the rest of the world is catching up with the Middle East in the use of transparent proxies, maybe Wikipedia will try to find some solution to the problem. In Qatar, for the last two years, we have continued to be blocked for months at a time from editing.
Good guess, but no cigar.
From TFA:
Israel was omitted because of fraud from that country, which seems like a good reason. Palestine was probably included in the list because it is recognised by the UN, and is included in ISO 3166-1. If you were to delete Palestine from the list, it would certainly be a very loaded political statement, but its inclusion is not.
> I'm interested to find out how to clean up "incorrect" Geolocation info too.
You can't.
I wrote IP::Country, the library used for country detection of relays within SpamAssassin. There is no manual tweaking of the country codes, I simply take the raw country data from the Regional Internet Registries (ARIN, RIPE, APNIC, LACNIC, AFRINIC) and encode them into a small, portable database.
That is also how all the other IP geolocation libraries work. If you started out from the premise that you'd accept manual changes to the database, it would produce more inaccuracies in the long run that simply following the WHOIS changes.
There is another way, which no one is using. Contraint-based geolocation pings the IP address from several known locations to estimate the location. There are academic papers saying it's more accurate than the WHOIS data, but it's never going to work on a webserver because it's doesn't produce a fast enough answer.
If you're interested in how the country data changes over time, check out the CVS of my data.
The only way to clear up a mistake in my country code is to correct the country code with the Regional Internet Registry. You won't be able to do that unless you own the block.
You should worry more about the disk cache. Previously opened files are cached in RAM in an unencrypted state.
Firewire ports and PCMCIA slots have direct memory access, so can be used to copy an image of your computer's RAM even if no one is logged in. This can recover useful forensic material even after a reboot cycle, as modern BIOS's don't clear RAM.
It looks like Vista's disk encryption is useless if you switch on the PC and access files.
Baffling, but true.
... Qatar has a vastly different government than Saudi ArabiaYou are correct in that there is more freedom of expression in Qatar than in Saudi, and the Ministry of Information was abolished in 1996. Unfortunately, QTel has not yet realised that there is no censorship in Qatar.
Before this story appeared on the front page of slashdot, Qataris were also prevented from creating new user accounts - thus preventing anyone from posting.
So, the problem appears to be fixed. Users can now register for accounts. Thank you slashdot front page (Kjkolb) ;)
The problem came from QTel censorship. Every connection passes through a QTel proxy server, which uses some simplistic rules to determine whether you should be protected from your own surfing habits. If you hit blocked pages too often, your phone rings and when you answer in English you get "I'm sorry, I must have a wrong number. CLUNK." Thus your voice has been recorded for posterity.
The shambles of Qatar's connection might be fixed soon. Q-CERT has just been set up and (hopefully), someone with a bit of influence will be in charge. It is obvious that a single point of failure for an entire population's internet connection is not sensible, but whether this means a better censorship system or the scrapping of censorship remains to be seen.
I'm sorry to interrupt your conspiracy theory with an unfortunate fact, but Lord Hutton isn't a Labour peer, but is a cross-bencher, meaning he's independent.
Did they give Linux users a chance to respond to the allegation of writing this worm? If not, I guess they've broken their producers guidelines.
Rather than farting around with the feedback form (>/dev/null), start the formal complaints process rolling. The Programme Complaints Unit (PCU) deals with serious editorial complaints about all publicly funded BBC services (including the websites). Start by filling in the form.
I would imagine the BBC are currently quite touchy about editorial standards, so you might get a quick response.
Rather than farting around with the feedback form (>/dev/null), start the formal complaints process rolling. The Programme Complaints Unit (PCU) deals with serious editorial complaints about all publicly funded BBC services (including the websites). Start by filling in the form.
I would imagine the BBC are currently quite touchy about editorial standards, so you might get a quick response.
> See Geobutton Or Geo::IP, IP::Country, javainetlocator or any other free, open source services. You shouldn't be paying $49 for a developers license to read the WHOIS data.
<plug>javainetlocator and IP::Country</plug> are also available.
The city data are unreliable. I've posted elsewhere (link1, link2) the reasons why, but will repeat the main points here.
The 'others' are taking american jobs. Quick, let's protect our lifestyles by limiting free trade.
This is simply xenophobia or, worse, racism. Unfortuantely, it's going to increase as the US Presidential Elections draw near, and various politicians try to create enemies they can battle against.
It's not seen as particularly pleasant any more to target the Far Eastern worker for stealing American jobs, so why should it be OK to target Indian programmers. This is a problem with the American economy, not the Indian economy. American programmers are paid too much and limiting free trade is not the answer.
India's sufficiently far away that most Americans will already have a very strange view of the country. And of course, if America says it is saving the Indian savages from 'slave wages', then they should be pleased about the generosity of their great western lords.
Strange that we never hear Americans complaining that the UK is stealing all its financial jobs. But then, it's mostly white people working in the City of London, isn't it.
Unfortunately, your logic is flawed. If both don't agree, then at least one is incorrect. This is a signal to the pilot that he should look out of the window rather than relying on his satellite location systems.
Having only one satellite location system does not make that system any more reliable.
"We are not sure how the virus got into our network" said Rick Devenuti, chief information officer for Microsoft.
Duh. I suspect through port 1434, the same way it got into all the other networks.
The site is slashdotted, so I haven't been able to have a look at it. However, if I were building a geo-search engine, I'd use the WHOIS data for the bulk of the indexing work, and for providing a default location for visitors. The tweaking around the edges (changing the location of the website or page), is just icing on the cake.
No one really knows the accuracy of IP->Country lookup. There's an onlgoing thread on the london perl mongers list about this topic. Some geolocation companies state 98% accuracy, which is pure bullshit. It's more likely to be around 70%, with most of the error occuring in overestimation of US addresses.
By the way, if you want a fast IP locator, here's one that's just as accurate as any of the commercial products. I'm surprised more people don't use this sort of stuff for providing intelligent defaults for their users when filling in HTML forms.
If you'd like to be more specific about which countries you filter, there's a perl module here. Available from all good CPAN mirrors.
Interviewed today by the Financial Times. His solution for building elegant software seems a bit extreme ;)
I agree with a lot you say.
The problem with your argument is that both of our drivers follow the JDBC spec, and thus synchronize against the Connection object becuase the MySQL API is not thread safe.
That other drivers do not follow the specification is no relief when ours do, taking the resultant performance hit during multithreaded operation.
In other respects, I agree with you that not many developers use Statements concurrently, but any developer who has read the JDBC spec is entitled to assume that Statements are thread-safe. It would be polite for the database developers to ensure that the underlying libraries are re-entrant. Unfortunately, that is not the case at present.
I'd always thought MySQL was a fast, simple database until I built a Type-II JDBC driver for it.
Because the API does not allow more than one result (MYSQL_RES structure) per connection, and the client libraries are not thread safe by default, any Java classes must be synchronized on the connection. In addition, all rows in a result must be retrieved completely using mysql_store_result rather than the more network efficient mysql_use_result.
The JDBC specification insists (sensibly, in my opinion) that Statement objects be thread safe. The necessary synchronization and use of mysql_store_result severely limits the speed of any mutithreaded application sharing a connection, and probably discounts the speed benefits of MySQL over other free databases.
I would guess the same problems exist with other multithreaded languages using MySQL, and developers should consider these limitations before blindly agrreing with MySQL propaganda that the database if faster than its competitors for running web applications.
ITV Digital is currently in administration, meaning the administrators (Deloitte and Touche) can sell the business without its debts.
Here's the .sig file from their outgoing mail:
ITV Digital plc, 346 Queenstown Rd, London SW8 4DG. Reg No. 3302715: In administration. Joint Administrators N.J. Dargan & N.G. Edwards The Joint Administrators act as agents of the companies and contract without personal liability.