Having worked for both USA Today and washingtonpost.com, I can tell you that know responsible news organization would tolerate this kind of behavior. Most have very explicit standing policies against digitally altering photos for publication with severe consequences (including termination) for violation of the policy.
While this seems a pretty clear cut violation, there is also some room for debate as to the proper role of Photoshop. Is cropping for presentation acceptable? Color correction? Graphical overlays (to point out characteristics of the photo or enhance the nformation value)? How about masking out someone who's permission you couldn't get for the photo?
Remember that the key asset of any news organization is the public's trust that they are reporting the "facts". While there is no real expectation of complete objectivity, altering the truth the fit your perspective will always be unacceptable. When you alter a photograph with the intent of changing it's meaning (even if it supports the other facts in the story), it erodes that trust.
Perhaps you should reconsider who is abusing privacy in this scenario. The STB will only know that you have just purchased a car if 1) you tell them or 2) your car dealer/leasing company/bank tells them. The fact that the advertisers are taking advantage of data about you sold by *other* companies does not make them the bad guys.
I am surprised....
on
The Chronoliths
·
· Score: 2, Interesting
...that no one has yet noted the parallel between the social unrest (and corresponding government reaction) in the book and the "war on terrorism(tm)" being prosecuted by the United States.
Clearly, the Chronoliths are a weapon of terror (a rather clever one, I might add). And while terror is not the central theme of the book, it clearly is a considerable factor in the economic, social and political upheaval that is the backdrop for the main story line.
I will let each reader decide for him/herself whether there is any conclusion to be drawn or not.
Actually, all three of you are wrong to some degree. Items are regularly bought, sold and traded which are the result of labor (many products in EQ are created by PCs using acquired materials and earned skills). Some spells require components that must be acquired before the spell is cast. The barter system for these items in EQ is quite evolved, as anyone who has spent more than a few hours playing the game will attest (try visiting Kelethin on any server on a Friday night and then tell me that there is no economy in EQ).
Further, items do go away regularly. Perhaps not in the no material existence sense, but people acquire items, retire characters or destroy items that they no longer need and can't sell or give away. Many items are not tradable (no-drop), and therefore are effectively "gone" once an individual has acquired that particular item.
I don't disagree that the economic rules in EQ are different than RW economies, but to claim that EQ has no economy *is* ridiculous.
Make sure that you are properly using the Cache-control and Expires directives. I managed to cut the bandwidth utilization of a major website by 15% by properly configuring the web server to allow AOL to cache images. While not allowing anyone to cache your pages sounds like a good idea at first, it drives your bandwidth way up. Also, using "Pragma: No-cache" is not a substitute. HTTP compliant proxy caches will ignore this as it is supposed to be a browser directive, not a proxy directive.
As another poster mentioned, these are all good suggestions. Although, a growing website has two other costs that rise rapidly as traffic increases - equipment and operations (people).
Equipment costs can be redcued substantially by taking an honest look at your site and deciding whether you really need every page to be PHP (in other words, are you parsing every page just so you can say "Hi, Allen"?). Static page delivery is extremely efficient. Can you, for example, frequently refresh pages instead of dynamically generating them? If a page gets more than one hit per minute, you can still reduce resource usage by regenerating it every minute.
People costs can be reduced by making the investment (some time, some money) to automate as many repetitive tasks as possible - log rolling, content updates, backups, purges, etc. There are quite a few free tools out there to automate these tasks - you just have to invest the time to learn them and install them properly. Also, install BigBrother or other monitoring tools to understand what things break frequently so you can fix them once instead of over and over again until you suddenly think to yourself, "Oh no, not again!".
Oh, and on the bandwidth issue - you might consider using "Content-encoding: gzip" where possible (there's an Apache module for this). I'm too lazy to look it up, but there was a discussion on/. not long ago regarding using this. It helps alot when the browser supports it. And finally, consider converting images to PNG's. They compress really well and are supported on all the v4 and v5 browsers.
Actually, it's not just the high-end Marriotts. I frequently travel to they bay area and all the Courtyard Marriotts (definitely NOT high-end) all have STSN.
However, my experience is that STSN provides mediocre service at best. I have used it in at least half a dozen different hotels over the past year and usually get 100-150K throughput. Not a lot of additional speed for $10/day.
Finally, on a more positive note, I recently stayed at the Four Seasons Austin (sometimes it pays to travel a lot). I do not remember the name of the service, but it wasn't STSN. I-something, I think. But it rocked! Averaged 300K+ on the few measurements I did.
The point being that caches are designed to maximize the efficiency of networks of computers, not individuals. The basic idea is that if I look at/., someone else on my network is more likely to be looking at/. than the Barney website (at least I hope thats true). By caching the pages most often requested by people on the network, bandwidth and latency are both reduced.
Caches and caching algorithms are not optimized for individual behavior. Yes, you could make them do what you want, but others have quite accurately pointed out that the cost of a little reduced latency is high in terms of consumed bandwidth. Or alternatively, the algorithm for look-ahead would have to be well-tuned to your preferences - something that has proved notoriously difficult to teach computers.
As someone working for one of the major political websites in 1996, I can tell you two things:
1) There were more than "localized" outages in 1996.
2) It is not going to happen to the major news sites this year.
Now for some explanation:
1) On Election Day 1996, about 6:30pm, UUNet had a major outage. What caused it, I don't remember. This caused a cascade effect, with everyone trying to route around the UUNet outage clogging the peering points (particularly the MAEs) causing widespread congestion.
The outage only lasted for about an hour. But once traffic started to flow, an hours worth of queued updates and requests flooded the servers causing additional congestion and/or failures.
Combine the above with the fact that no one had any idea how many people were going to look to the net for election coverage and we basically underestimated the impact that this would have.
2) I know most of the people that designed the systems used by the major news sites (WashingtonPost.com, NYTimes.com, CNN.com, ABCNews.com, USAToday.com) and they will not make those mistakes this year.
Clearly, the last four years has allowed us to gather a great deal more information about traffic patterns and ways to improve throughput (CDNs, lightweight pages, proxy caches, etc.). In addition, bandwidth is cheaper and more reliable than ever.
In fact, election night turns out to be anti-climactic in most cases. I designed and built the election returns system for Washingtonpost.com in 1998. I busted my ass for a month prior to the elections. On election night, I ate pizza and watched returns. There was simply nothing for me to do.
When considering the events of 1996 as compared to 2000, you have to remember that we were all flying by the seat of our pants in 1996 and had no idea what to expect both in terms of traffic and problems. We have learned a great deal since then and I would be incredibly surprised to hear of problems of this sort at any of the mainstream news sites on the net.
Slashdot Mirror
Having worked for both USA Today and washingtonpost.com, I can tell you that know responsible news organization would tolerate this kind of behavior. Most have very explicit standing policies against digitally altering photos for publication with severe consequences (including termination) for violation of the policy.
While this seems a pretty clear cut violation, there is also some room for debate as to the proper role of Photoshop. Is cropping for presentation acceptable? Color correction? Graphical overlays (to point out characteristics of the photo or enhance the nformation value)? How about masking out someone who's permission you couldn't get for the photo?
Remember that the key asset of any news organization is the public's trust that they are reporting the "facts". While there is no real expectation of complete objectivity, altering the truth the fit your perspective will always be unacceptable. When you alter a photograph with the intent of changing it's meaning (even if it supports the other facts in the story), it erodes that trust.
Perhaps you should reconsider who is abusing privacy in this scenario. The STB will only know that you have just purchased a car if 1) you tell them or 2) your car dealer/leasing company/bank tells them. The fact that the advertisers are taking advantage of data about you sold by *other* companies does not make them the bad guys.
...that no one has yet noted the parallel between the social unrest (and corresponding government reaction) in the book and the "war on terrorism(tm)" being prosecuted by the United States.
Clearly, the Chronoliths are a weapon of terror (a rather clever one, I might add). And while terror is not the central theme of the book, it clearly is a considerable factor in the economic, social and political upheaval that is the backdrop for the main story line.
I will let each reader decide for him/herself whether there is any conclusion to be drawn or not.
Actually, all three of you are wrong to some degree. Items are regularly bought, sold and traded which are the result of labor (many products in EQ are created by PCs using acquired materials and earned skills). Some spells require components that must be acquired before the spell is cast. The barter system for these items in EQ is quite evolved, as anyone who has spent more than a few hours playing the game will attest (try visiting Kelethin on any server on a Friday night and then tell me that there is no economy in EQ).
Further, items do go away regularly. Perhaps not in the no material existence sense, but people acquire items, retire characters or destroy items that they no longer need and can't sell or give away. Many items are not tradable (no-drop), and therefore are effectively "gone" once an individual has acquired that particular item.
I don't disagree that the economic rules in EQ are different than RW economies, but to claim that EQ has no economy *is* ridiculous.
Actually, theres a corillary to this:
Make sure that you are properly using the Cache-control and Expires directives. I managed to cut the bandwidth utilization of a major website by 15% by properly configuring the web server to allow AOL to cache images. While not allowing anyone to cache your pages sounds like a good idea at first, it drives your bandwidth way up. Also, using "Pragma: No-cache" is not a substitute. HTTP compliant proxy caches will ignore this as it is supposed to be a browser directive, not a proxy directive.
As another poster mentioned, these are all good suggestions. Although, a growing website has two other costs that rise rapidly as traffic increases - equipment and operations (people).
/. not long ago regarding using this. It helps alot when the browser supports it. And finally, consider converting images to PNG's. They compress really well and are supported on all the v4 and v5 browsers.
Equipment costs can be redcued substantially by taking an honest look at your site and deciding whether you really need every page to be PHP (in other words, are you parsing every page just so you can say "Hi, Allen"?). Static page delivery is extremely efficient. Can you, for example, frequently refresh pages instead of dynamically generating them? If a page gets more than one hit per minute, you can still reduce resource usage by regenerating it every minute.
People costs can be reduced by making the investment (some time, some money) to automate as many repetitive tasks as possible - log rolling, content updates, backups, purges, etc. There are quite a few free tools out there to automate these tasks - you just have to invest the time to learn them and install them properly. Also, install BigBrother or other monitoring tools to understand what things break frequently so you can fix them once instead of over and over again until you suddenly think to yourself, "Oh no, not again!".
Oh, and on the bandwidth issue - you might consider using "Content-encoding: gzip" where possible (there's an Apache module for this). I'm too lazy to look it up, but there was a discussion on
Actually, it's not just the high-end Marriotts. I frequently travel to they bay area and all the Courtyard Marriotts (definitely NOT high-end) all have STSN.
However, my experience is that STSN provides mediocre service at best. I have used it in at least half a dozen different hotels over the past year and usually get 100-150K throughput. Not a lot of additional speed for $10/day.
Finally, on a more positive note, I recently stayed at the Four Seasons Austin (sometimes it pays to travel a lot). I do not remember the name of the service, but it wasn't STSN. I-something, I think. But it rocked! Averaged 300K+ on the few measurements I did.
The point being that caches are designed to maximize the efficiency of networks of computers, not individuals. The basic idea is that if I look at /., someone else on my network is more likely to be looking at /. than the Barney website (at least I hope thats true). By caching the pages most often requested by people on the network, bandwidth and latency are both reduced.
Caches and caching algorithms are not optimized for individual behavior. Yes, you could make them do what you want, but others have quite accurately pointed out that the cost of a little reduced latency is high in terms of consumed bandwidth. Or alternatively, the algorithm for look-ahead would have to be well-tuned to your preferences - something that has proved notoriously difficult to teach computers.
As someone working for one of the major political websites in 1996, I can tell you two things:
1) There were more than "localized" outages in 1996.
2) It is not going to happen to the major news sites this year.
Now for some explanation:
1) On Election Day 1996, about 6:30pm, UUNet had a major outage. What caused it, I don't remember. This caused a cascade effect, with everyone trying to route around the UUNet outage clogging the peering points (particularly the MAEs) causing widespread congestion.
The outage only lasted for about an hour. But once traffic started to flow, an hours worth of queued updates and requests flooded the servers causing additional congestion and/or failures.
Combine the above with the fact that no one had any idea how many people were going to look to the net for election coverage and we basically underestimated the impact that this would have.
2) I know most of the people that designed the systems used by the major news sites (WashingtonPost.com, NYTimes.com, CNN.com, ABCNews.com, USAToday.com) and they will not make those mistakes this year.
Clearly, the last four years has allowed us to gather a great deal more information about traffic patterns and ways to improve throughput (CDNs, lightweight pages, proxy caches, etc.). In addition, bandwidth is cheaper and more reliable than ever.
In fact, election night turns out to be anti-climactic in most cases. I designed and built the election returns system for Washingtonpost.com in 1998. I busted my ass for a month prior to the elections. On election night, I ate pizza and watched returns. There was simply nothing for me to do.
When considering the events of 1996 as compared to 2000, you have to remember that we were all flying by the seat of our pants in 1996 and had no idea what to expect both in terms of traffic and problems. We have learned a great deal since then and I would be incredibly surprised to hear of problems of this sort at any of the mainstream news sites on the net.