They noticed the ATM overhead, but did not recognize it. Somewhere it gives the example of a 128 up 512 down ADSL connection, says these rates are in kbps and need to be entered in the configuration (in kbits, not kbytes per second) and then suggest you to subtract 50 from the figures as a first try.
This will cater for the ATM overhead (at these low rates). Usually, ADSL providers are cheating in that they specify the ATM rate, not the expected IP bitrate, in their advertisements. But when you set your shaper to 10-15% lower it should work OK on the average.
Of course the effect you mention exists as well, but as a first approximation a fixed cutback of the real linespeed will help.
I was prepared to consider feedback and was already thinking about offering an alternate stylesheet with larger fonts, but I am not doing this work to be offended by assholes like you so I suggest you go to hell with your eyesight problems.
This is not a problem of Philips or this specific technology, but of the western world as a whole. From a world where it was natural to learn from others, to build on other people's work (how could we have arrived where we are when every baby had to invent everything by itself?), we have moved into a society where knowledge is called "intellectual property" and is guarded as an economic valuable.
In other cultures, e.g. China, the situation is completely different. Copying what someone else has invented is natural behaviour there, and can even be seen as an award of the original inventor. But what we see now that China is evolving as an economic power is (in the picture of the western world) that China has to change to an intellectual property model, rather than that the western world would change, or the two would meet in the middle.
I think they either are compression artifacts (allmost everything you see on TV today is MPEG compressed at some point, even when you receive it analog!), or they are an indication that some digital processing inside the TV is underpowered and hits its processing performance limits when it has to work hard on motion compensation etc.
The article header describes the problem incorrectly. The problem is not the response time of the screen. By now that is good enough. The problem is that an LCD displays the same frame all the time until the next frame comes in, while a CRT displays only brief flashes.
See the response titled "Some information on the nature of the problem" for more details.
Unfortunately the Mozilla team prefers discussion about the names and version numbers of their products over the fixing of bugs.
I was disappointed when the Mozilla suite was split up in a browser and a mailer, dropping the editor part on the floor. I was more disappointed when they left the suite out in the cold and started fiddling with the name again.
But I did not believe my own eyes when they started to fiddle with version numbers, "skipping from 1.0 to 1.5 instead of 1.1 because it is a major upgrade".
I'm not so sure anymore that the project is in good hands of such a team.
Last time I ran into this (a couple of weeks back), by the time someone mentioned the problem with our mailserver (which was hosting the vanity domain in one of the addresses in the To: line that forwarded to an offsite pop box, from which the Microsoft product downloaded it) -- only 2-3 hours after it had started looping -- there were literally 60,000 messages waiting to be delivered to this one (remote) POP box.
At first I thought it was an unlikely coincidence that hit us. Googling for thinks like "mail loop" would find many causes for looping mail by misconfigured forwarding etc, but it took me a lot of effort to find this one in M$'s KB. Of course this is because the KB article does not even touch upon the meltdown consequences this bug may have, they just mention "there are more messages than expected in your outbound queue".
But apparently it happens all the time. You would expect this one to be on CNN in the section of "bad things that happen to the net" (adjacent to the virus-or-phish-of-the-day report), but no. You would also expect this to be listed in the critical updates on Windows Update, but no.
The reason I would like to see it configurable is that I would like to be able to keep the "Reply to All" button for most people, but to disable it for known illiterates that have shown carelessness before.
What we have seen many times in the company is people sending to large groups or even everyone in the company, asking for reply (e.g. "who is going to join at this week's drink"). Then, a certain group of users, and often the same every time, is going to hit "reply all" to send their confirmation. This results in many mails sent to everyone, and in frustration. Of course it could have been prevented when the original sender used BCC to send out the message, a practice we try to adopt as much as possible.
But of course you are right, in normal business communication it is a useful feature.
That is not simple. As those people are using a POP3 box at an ISP, they are often also sending out their mail via that same ISP's mailservers. So you see mail coming from the standard mailrelay of a big ISP, and only when looking at headers (after receiving the message) you can see it went via a Microsoft system. Even then, there is little difference between a typical Exchange mailserver and those broken SBS 2003 servers.
So, rejecting on the MAIL FROM:, at least temporarily until the loop has died, is the only practical method. (of course the original sender is only a victim)
It took me a week or two to find the reason this happened, of course not very actively looking for it because we were not the cause, only a victim.
It is not as bad as you think at first time, because in the loop there is a "look in POP box" action which occurs at an (apparently configurable) interval. When it would be direct SMTP looping it would melt down pretty quickly. However, there were about 4 of those broken SBS 2003 servers on the list that received the messages, so the behaviour was still pretty bad (we received thousands of messages in about 10 days, but it could have been millions).
These SBS systems were on ADSL and Cable, and when pinging them we got several-second reply times. They were of course very busy pumping out the junk.
So I thought, let's just lookup a couple of contact numbers and call them asking to at least pull the plug until they have resolved the problem (at that time I had not found the simple cause). As you would expect, they were all small businesses that had their IT outsourced to some other company, they had called about the problem, but nobody had gotten around to solving it yet and nobody was around with the "technical knowledge" do do something like pulling a cable.
What I sometimes find disturbing is that a company like Microsoft is allowed (by there customers) to deliver work of this quality and act this leisurely. This software is a potential timebomb under the Internet e-mail system, yet it is sold by "certified professionals" that are not alerted about issues like this.
Because, in my case, there were hundreds of addresses listed in the To: header. I don't consider that particularly clever. Not only did it triple the size of each message, it also disclosed the address of everyone in the mailinglist to all members. This, in general, is not very wise.
Besides that, I would guess that the Kerberos system uses the UTC time, not the local time, as a reference and thus is not affected by DST. But I have not verified that.
The looming catastrophe in America is on the horizon yet the political powers in Washington refuse to acknowledge that the current American way of life is not sustainable and must be significantly downsized if there is any chance of avoiding a tremendous disaster caused by a massive energy shortage.
Not only that, but there is a looming environmental disaster caused by massive CO2 output, and the US refuses to do anything because it would hurt their economy. Apparently the US government, unlike the rest of the planet, still believes that a growing economy is more important than a surviving planet. This does not bode well for the future...
Windows timestamps captured in the past will show local time based on the new rules
Actually, timestamps captured by Windows in the past will show weird behaviour when DST is involved no matter what. Try setting a filetime to a known value and looking at it after DST has started or ended. It does not look like Microsoft thinks of filetimes as important...
Cisco devices, both IOS and CatOS based, use the 'summertime' command to compensate for daylight saving time (example [cisco.com]). This means that a change in the DST setup would force you to upgrade code.
Or at least it would force you to study the command reference a bit better, and find the second optional form of the command that allows you to specify the beginning and end of summertime.
That would mean you require only a configuration change, and not a code upgrade. But of course you would need to read the manual...
There is an interesting feature in the Microsoft POP3 connector included with SBS 2003 that can also cause such a flurry of mails.
When the original sender is stupid enough to include all addresses a mail is sent to in the To: header, and two or more readers of mail have their mailbox at an ISP and copy it to their Exchange server using the abovementioned Microsoft POP3 connector, mail can really start bouncing around.
Why? Because of a bug in the Microsoft POP3 connector, mail that it retrieves from a POP3 box is sent to all addresses in the To: line. So the mailserver of every user of this crap will re-send a copy of the mail to all recepients, even those outside his or her own domain. When two or more users receive the message, they start sending more and more copies around.
A while ago we received the same message from someone several thousand times. It took me a while to figure out what was really happening (we are not using those MS products ourselves), and the only way to kill it off was to reject all mail from the original sender.
It seems that KB835734 offers a fix for this fatal bug, but MS does not consider it critical so I presume most admins have not applied it. Those SBS systems are a ticking bomb in the e-mail system.
There is an open bug in Bugzilla to have configurable toolbars, including options to remove items like "Reply to All". Unfortunately, it seems the coders don't understand the relevance and it remains open for several years now.
You get (more than) 3 7k250 250GB drives for the price of one of those 7k500 drives, so they are not very attractive for building a very large archive.
Same with workstations: when someone opens the case and/or resets the BIOS settings you will get an alert and can go after it. Usually the problem is not that someone can get access to the system. You only want to protect against doing that without being detected.
I run a similar setup but with sendmail. Indeed it works very well. For now.
But it relies on spammers making stupid mistakes, like running buggy smtp servers, sending messages that spamassassin easily recognizes, including urls that are blacklisted on SURBL lists, etc.
Once "everyone" starts running such a filter (or has their ISP do it on their behalf) the spammers will start working around it. That is certainly possible, just not worth the trouble to them for now.
Slashdot Mirror
They noticed the ATM overhead, but did not recognize it.
Somewhere it gives the example of a 128 up 512 down ADSL connection, says these rates are in kbps and need to be entered in the configuration (in kbits, not kbytes per second) and then suggest you to subtract 50 from the figures as a first try.
This will cater for the ATM overhead (at these low rates).
Usually, ADSL providers are cheating in that they specify the ATM rate, not the expected IP bitrate, in their advertisements. But when you set your shaper to 10-15% lower it should work OK on the average.
Of course the effect you mention exists as well, but as a first approximation a fixed cutback of the real linespeed will help.
I was prepared to consider feedback and was already thinking about offering an alternate stylesheet with larger fonts, but I am not doing this work to be offended by assholes like you so I suggest you go to hell with your eyesight problems.
Byebye.
Of course the fonts are absolute sized, but changing the text size in Mozilla works OK. It destroys the layout but the text becomes larger.
Does not work in IE, but that is M$'s problem.
A larger font is used in the print style sheet as well.
You can give it a try at http://www.uw.nl/
This is not a problem of Philips or this specific technology, but of the western world as a whole.
From a world where it was natural to learn from others, to build on other people's work (how could we have arrived where we are when every baby had to invent everything by itself?), we have moved into a society where knowledge is called "intellectual property" and is guarded as an economic valuable.
In other cultures, e.g. China, the situation is completely different. Copying what someone else has invented is natural behaviour there, and can even be seen as an award of the original inventor.
But what we see now that China is evolving as an economic power is (in the picture of the western world) that China has to change to an intellectual property model, rather than that the western world would change, or the two would meet in the middle.
I wonder how that is going to work out.
I think they either are compression artifacts (allmost everything you see on TV today is MPEG compressed at some point, even when you receive it analog!), or they are an indication that some digital processing inside the TV is underpowered and hits its processing performance limits when it has to work hard on motion compensation etc.
The article header describes the problem incorrectly.
The problem is not the response time of the screen. By now that is good enough.
The problem is that an LCD displays the same frame all the time until the next frame comes in, while a CRT displays only brief flashes.
See the response titled "Some information on the nature of the problem" for more details.
Unfortunately the Mozilla team prefers discussion about the names and version numbers of their products over the fixing of bugs.
I was disappointed when the Mozilla suite was split up in a browser and a mailer, dropping the editor part on the floor.
I was more disappointed when they left the suite out in the cold and started fiddling with the name again.
But I did not believe my own eyes when they started to fiddle with version numbers, "skipping from 1.0 to 1.5 instead of 1.1 because it is a major upgrade".
I'm not so sure anymore that the project is in good hands of such a team.
Last time I ran into this (a couple of weeks back), by the time someone mentioned the problem with our mailserver (which was hosting the vanity domain in one of the addresses in the To: line that forwarded to an offsite pop box, from which the Microsoft product downloaded it) -- only 2-3 hours after it had started looping -- there were literally 60,000 messages waiting to be delivered to this one (remote) POP box.
At first I thought it was an unlikely coincidence that hit us. Googling for thinks like "mail loop" would find many causes for looping mail by misconfigured forwarding etc, but it took me a lot of effort to find this one in M$'s KB.
Of course this is because the KB article does not even touch upon the meltdown consequences this bug may have, they just mention "there are more messages than expected in your outbound queue".
But apparently it happens all the time. You would expect this one to be on CNN in the section of "bad things that happen to the net" (adjacent to the virus-or-phish-of-the-day report), but no.
You would also expect this to be listed in the critical updates on Windows Update, but no.
Strange...
The bug is for the Mozilla suite, which allows removing some buttons but not all. It does not allow removal of Reply All.
The reason I would like to see it configurable is that I would like to be able to keep the "Reply to All" button for most people, but to disable it for known illiterates that have shown carelessness before.
What we have seen many times in the company is people sending to large groups or even everyone in the company, asking for reply (e.g. "who is going to join at this week's drink"). Then, a certain group of users, and often the same every time, is going to hit "reply all" to send their confirmation. This results in many mails sent to everyone, and in frustration.
Of course it could have been prevented when the original sender used BCC to send out the message, a practice we try to adopt as much as possible.
But of course you are right, in normal business communication it is a useful feature.
That is not simple. As those people are using a POP3 box at an ISP, they are often also sending out their mail via that same ISP's mailservers. So you see mail coming from the standard mailrelay of a big ISP, and only when looking at headers (after receiving the message) you can see it went via a Microsoft system. Even then, there is little difference between a typical Exchange mailserver and those broken SBS 2003 servers.
So, rejecting on the MAIL FROM:, at least temporarily until the loop has died, is the only practical method.
(of course the original sender is only a victim)
It took me a week or two to find the reason this happened, of course not very actively looking for it because we were not the cause, only a victim.
It is not as bad as you think at first time, because in the loop there is a "look in POP box" action which occurs at an (apparently configurable) interval. When it would be direct SMTP looping it would melt down pretty quickly.
However, there were about 4 of those broken SBS 2003 servers on the list that received the messages, so the behaviour was still pretty bad (we received thousands of messages in about 10 days, but it could have been millions).
These SBS systems were on ADSL and Cable, and when pinging them we got several-second reply times. They were of course very busy pumping out the junk.
So I thought, let's just lookup a couple of contact numbers and call them asking to at least pull the plug until they have resolved the problem (at that time I had not found the simple cause).
As you would expect, they were all small businesses that had their IT outsourced to some other company, they had called about the problem, but nobody had gotten around to solving it yet and nobody was around with the "technical knowledge" do do something like pulling a cable.
What I sometimes find disturbing is that a company like Microsoft is allowed (by there customers) to deliver work of this quality and act this leisurely. This software is a potential timebomb under the Internet e-mail system, yet it is sold by "certified professionals" that are not alerted about issues like this.
Because, in my case, there were hundreds of addresses listed in the To: header. I don't consider that particularly clever. Not only did it triple the size of each message, it also disclosed the address of everyone in the mailinglist to all members.
This, in general, is not very wise.
Besides that, I would guess that the Kerberos system uses the UTC time, not the local time, as a reference and thus is not affected by DST. But I have not verified that.
The looming catastrophe in America is on the horizon yet the political powers in Washington refuse to acknowledge that the current American way of life is not sustainable and must be significantly downsized if there is any chance of avoiding a tremendous disaster caused by a massive energy shortage.
Not only that, but there is a looming environmental disaster caused by massive CO2 output, and the US refuses to do anything because it would hurt their economy.
Apparently the US government, unlike the rest of the planet, still believes that a growing economy is more important than a surviving planet.
This does not bode well for the future...
Windows timestamps captured in the past will show local time based on the new rules
Actually, timestamps captured by Windows in the past will show weird behaviour when DST is involved no matter what.
Try setting a filetime to a known value and looking at it after DST has started or ended.
It does not look like Microsoft thinks of filetimes as important...
That would not work. We are that much ahead because of the irregular shape of timezones, but still we get an hour extra DST.
Cisco devices, both IOS and CatOS based, use the 'summertime' command to compensate for daylight saving time (example [cisco.com]). This means that a change in the DST setup would force you to upgrade code.
Or at least it would force you to study the command reference a bit better, and find the second optional form of the command that allows you to specify the beginning and end of summertime.
That would mean you require only a configuration change, and not a code upgrade.
But of course you would need to read the manual...
There is an interesting feature in the Microsoft POP3 connector included with SBS 2003 that can also cause such a flurry of mails.
When the original sender is stupid enough to include all addresses a mail is sent to in the To: header, and two or more readers of mail have their mailbox at an ISP and copy it to their Exchange server using the abovementioned Microsoft POP3 connector, mail can really start bouncing around.
Why? Because of a bug in the Microsoft POP3 connector, mail that it retrieves from a POP3 box is sent to all addresses in the To: line. So the mailserver of every user of this crap will re-send a copy of the mail to all recepients, even those outside his or her own domain.
When two or more users receive the message, they start sending more and more copies around.
A while ago we received the same message from someone several thousand times. It took me a while to figure out what was really happening (we are not using those MS products ourselves), and the only way to kill it off was to reject all mail from the original sender.
It seems that KB835734 offers a fix for this fatal bug, but MS does not consider it critical so I presume most admins have not applied it. Those SBS systems are a ticking bomb in the e-mail system.
There is an open bug in Bugzilla to have configurable toolbars, including options to remove items like "Reply to All". Unfortunately, it seems the coders don't understand the relevance and it remains open for several years now.
You get (more than) 3 7k250 250GB drives for the price of one of those 7k500 drives, so they are not very attractive for building a very large archive.
if you open a desktop machine that's been running for 10+ years it's probably a quantum
On the other hand, if the machine was ever powered down and back up and the disk started spinning again, it probably wasn't Quantum.
Same with workstations: when someone opens the case and/or resets the BIOS settings you will get an alert and can go after it.
Usually the problem is not that someone can get access to the system. You only want to protect against doing that without being detected.
When the system boots from a CD you don't need such complicated methods.
But of course a locked-down system won't boot from CD.
I run a similar setup but with sendmail.
Indeed it works very well. For now.
But it relies on spammers making stupid mistakes, like running buggy smtp servers, sending messages that spamassassin easily recognizes, including urls that are blacklisted on SURBL lists, etc.
Once "everyone" starts running such a filter (or has their ISP do it on their behalf) the spammers will start working around it. That is certainly possible, just not worth the trouble to them for now.