So wait if someone has access to your pc and can change things on it there may be a security problem?
This is not really different than someone being able to steal your cookie.
I would say that the problem is more that people can still access your data on the local pc at a latter stage (No worse than an old fashioned desktop application). This could and most probably will be mitigated by using some form of encryption of the offline storage using your login.
CORS used incorrectly will be a problem but then again you can say the same for all the current technology at the moment.
All new technology is a security problem till people work out best practices. Though there are many advantages to using them. (Transparent failover of your web app as one off the top of my head)
If you want to try something more underhanded. Scan for the visiable hot spots. Then add your own wifi router with one of those SSID's. You can then see where the students are going or if they are cheating. Though this is a bit underhanded but should be simpler than just blocking. Of course you could just tell the students that you've done this and do nothing which should work as well.
Not dead as were doing it at http://www.ptawug.co.za/. A community run and managed wifi network that spans Pretoria. I think the main thing stopping adoption in most places is there are easier and cheaper alternatives. With the Telkom monopoly in our country its cheaper and easier for us to build our own network.
I agree.
We hard drive reads are more of a bottle neck than about anything else. So the less data we have to pull off of the hard drive to perform a query. The quicker it is going to be.
Which is where normalisation comes in.
As for having multiple copies allows you to go back and find out which is correct. Sometimes they are both correct. Which then requires a manual intervention to merge the data:) Which could have been fixed by the scecond set of changes updating the first set of changes.
Though must say the find the fact that the the article say's randomly loosing data is ok:) rather scary. Sure there are certain transactions you don't care wether it works or not but there are others where is matters alot. To use his example of loosing once cent. Sure not a majour problem. But if that one cent was your paycheck going in at the end of the month you would care alot.
Slashdot Mirror
So wait if someone has access to your pc and can change things on it there may be a security problem? This is not really different than someone being able to steal your cookie. I would say that the problem is more that people can still access your data on the local pc at a latter stage (No worse than an old fashioned desktop application). This could and most probably will be mitigated by using some form of encryption of the offline storage using your login. CORS used incorrectly will be a problem but then again you can say the same for all the current technology at the moment. All new technology is a security problem till people work out best practices. Though there are many advantages to using them. (Transparent failover of your web app as one off the top of my head)
If you want to try something more underhanded. Scan for the visiable hot spots. Then add your own wifi router with one of those SSID's. You can then see where the students are going or if they are cheating. Though this is a bit underhanded but should be simpler than just blocking. Of course you could just tell the students that you've done this and do nothing which should work as well.
Not dead as were doing it at http://www.ptawug.co.za/. A community run and managed wifi network that spans Pretoria. I think the main thing stopping adoption in most places is there are easier and cheaper alternatives. With the Telkom monopoly in our country its cheaper and easier for us to build our own network.
I agree. We hard drive reads are more of a bottle neck than about anything else. So the less data we have to pull off of the hard drive to perform a query. The quicker it is going to be. Which is where normalisation comes in. As for having multiple copies allows you to go back and find out which is correct. Sometimes they are both correct. Which then requires a manual intervention to merge the data :) Which could have been fixed by the scecond set of changes updating the first set of changes.
Though must say the find the fact that the the article say's randomly loosing data is ok :) rather scary. Sure there are certain transactions you don't care wether it works or not but there are others where is matters alot. To use his example of loosing once cent. Sure not a majour problem. But if that one cent was your paycheck going in at the end of the month you would care alot.