I'm not sure about the law where you live, but here if you are required to wait as a function of your job, you get paid for your waiting time. Period.
If that means you go over 8 hours a day or 40 hours a week you get overtime too.
Not only that, but with a couple exceptions (banking is one, IIRC), you cannot be penalized for refusing a shift change without 48 hours notice. In other words, you can walk out at the end of your shift and cannot be fired for doing so.
In theory, I agree. In practice, I'm be afraid of sticking too many devices on USB power simply because I don't trust cheap motherboard manufacturers to be able to handle the load. There is also a significant risk of a cheaply made device shorting out and taking out the motherboard as well as every other connected device.
That being said, it does drive me nuts how many USB devices have their own brick supply too. In fact, other then my cordless phone, I think every brick attached to a power bar on the floor beside my PC is connected to a device connected via USB to my system.
One other concern with having USB supplied power -- This isn't practical for all devices, many devices need power when the system is off too. My cordless mouse (rechargable), Palm, PPC and digital camera all would need power when the system is off. I'd be annoyed if my SB Extigy turned off with the system too, it's nice having it passthrough audio from line-in to my speakers even when my PC is off.
Also consider that even if power was available when the system is off, many devices really should turn themselves off when the system is off. It would be nice to cut power to these devices, but that would mean adding yet another pin (one for always-on power, one for switched power)
Now admittedly many systems (mine included), still provide USB power when the system is off. However, looking at my power supply's load limits, the power supplied when the power supply is off is minimal, and the hardware mentioned earlier would go over my power supply's standby maximums.
This could be resolved as part of the specification by requiring another connector from the power supply to the motherboard (to show the motherboard that the power supply is capable of supplying higher demand then the current ATX standby power). If this new connector wasn't connected then the motherboard would only supply power via USB when the system is running. As a workaround, some external USB hubs could have power connectors that would supply downstream devices with power even when the system is off.
All of this is a lot more hassle then standardizing the connections used by bricks connecting to external hardware. If each connector size corresponded to a specific voltage and minimum amp rating, bricks would become interchangable, and it would become feasible to buy a single power supply that would feed all of your power hungry gadgets without dinking around with sizing the power connector.
Sure IF the publisher is claiming that all IPs in the block are spammers.
Remember that truth is 100% defense against libel.
I don't see many, if any, DNSbls claiming they only contain spammers. The RBL includes a lot more then spam sources, it also includes any form of spam support, which is a lot easier to prove then if they said "This IP is a spammer." as long as they follow their own listing criteria.
A DNSbl being offline will always return negative results NEVER a positive result.
A DNSbl lookup just takes the IP in 1.2.3.4 format and does a query for 4.3.2.1.dnsbl.example.com
The only way you can get a positive result on a DNSbl lookup is if you get an IP returned. Typically 127.0.0.2 was used, although some DNSbls include feedback in the status either in bit or by response code.
If something goes wrong in the lookup (DNS resolved crashes, DNSbl server is down, DNSbl server is overloaded, port 53 is firewalled, or you've simply entered a wrong DNSbl host), you'll virtually always get a "not-listed" result back.
Going back to the RFC, a DNSbl error would tend to be a permanent error since retrying delivery attempts over and over will never get the message through. The only way the message will suddenly become deliverable is if the sender takes action to get delisted.
In general, anything that will resolve itself in time or needs resolution on the recipient's side should be a 4xx error, anything that needs sender corrective action should be a 5xx.
I'm not saying that people don't 4xx DNSbl listings all the time, but if they've bothered to read and understand rfc2821's definitions of temporary and permanent they will probably know better.
Either way though, it doesn't really address the issue: If your mail is important enough that you can't tolerate a delay in delivery, either get the recipient to whitelist you, or stick around for a confirmation (either a return receipt, if your recipient is willing, or wait for your recipient to hit reply and say "y0 d00d I got yer message" -- Email is not now and will not ever be a guaranteed delivery mechanism.
The vast majority of my bricks are between 5V and 12V, many being either 5V or 9V.
Some devices obviously couldn't use a central power supply, printers being a good example. Fine. None of my printers have a brick anyway, my laser printers both use standard PC power cables and my inkjet has an external power supply that takes a PC power cable input, so it doesn't block an entire power bar.
It's not the end of the world if a couple devices still have special needs, but why should I need a seperate brick for each device which fits into the common range?
As much as DC-DC power supplies might raise the price slightly, it would be offset by the back that the manufacturer doesn't need to supply a brick.
I'd be more then willing to spend the $150 I just spent on a computer power supply on a magic box that would provide everything from from 3v to 12v with standardized cables if it would remove the rats next of of power bars with 2-3 items connected to each.
Heck, if they'd just standardize on one size of plug for each voltage, those of us who want to invest in a central power supply could do so. Those that don't could stay with the manufacturer supplied bricks.
If the DNSbl claims to only list spammers, then you're correct.
However, if the listing criteria and spelled out and followed strictly, no libel has occured.
The RBL specifically mentions "multi-hop (multiple IP) open relay, a spam source, or a spam support service (e.g., a webserver or nameserver).", which is more then just spammers.
The Spamhaus Exploits Block List (XBL) is an example of one which doesn't even list spammers directly, XBL is "a realtime database of IP addresses of illegal 3rd party exploits, including open proxies (HTTP, socks, AnalogX, wingate, etc), worms/viruses with built-in spam engines, and other types of trojan-horse exploits."
The fact that spammres happen to abuse many of those resources is incidental, but it's one of the most effective lists out there.
What do you think will be more reliable, a thousand $50 power supplies (one in each server), or $50,000 spent on building a redundant power solution to power the entire data center?
You're already building around a number of potential single points of failure (Everything from the power in from the grid to the UPS(es) to the wire from the UPS to each servers' UPS)
Sure, but if they could settle on a couple standard voltages (5V and 12V anybody?), you could buy one power supply and connect it to all your third party junk.
It would even let manufacturers cheap out by not providing a supply at all, so you'd think they'd be all over this.
Yes and no. Redundancy is fine if the machines are actually redundant.
However, if you give me an option between having a server with a single power supply or a shared power supply between a dozen serves, all other factors being equal, I'll go for the shared one every time.
Why?
Well, if something goes wrong, they have 12x the incentive to fix it *now*
I believe that would be a "fire suppression system" or an "anti-fire system" -- You don't really need an anti-fire suppression system, you could save yourself a lot of money and not bother installing a fire suppression system in the first place.
If the receiving server responds with a 5xx (permanent error), the sending server should give up immediately it definitely should not try over and over.
How about if you wrote a letter and the postman ate it?
How about if the mail server (or mail client, for that matter) had a disk problem and lost the message?
Email is *NOT* has no guaranteed delivery mechanism, it is best-effort every step of the way.
More importantly, the sender would receive a bounce from their SMTP server, so they would know their message didn't get through. They'd call and tell the love of their life over the phone instead.
Then I'd say he has a good incentive to find an ISP who runs a clean shop.
You don't generally get listed on any of the larger lists (spamhaus, SPEWS, MAPS, etc) because a spammer got a hosting account and started spamming.
You get listed because you fail to terminate the spammer, or you fail to prevent the spammer from signing up again (which is substantially the same thing)
It's also worth noting that it's not as though any DNSbl has the ability to actually block mail.
My server refuses mail from IPs on a number of DNSbls because *I* choose to use those DNSbls. If I block an important email it is MY responsibility, no one else's responsbility.
I whitelist important senders so that they get through even if they manage to get listed.
My reject message is very simple: "<DNSBL NAME>: email whitelist@<servername> to be whitelisted" -- When an email is sent to the whitelist account, my server automatically whitelists the sender IP and forward the message to my support inbox. This means any sender who gets blacklisted can get themselves whitelisted without much difficulty and without requiring assistance from anyone.
Sure a spammer could abuse this mechansim, but so far, it hasn't happened. Even if they do, they'd only get to spam for a few hours, and their mail still goes through my other filtering systems.
For example: I don't care who you are if you HELO/EHLO as my server name or my IP, you don't get to send mail to me anymore. Period. This alone catches upwards of 50% of spam attempts.
Slashdot Mirror
Better yet, combine Adblock with Flashblock.
Flashblock automatically replaces all flash elements with an icon you can click on to start the flash.
This means I don't have to universally block flash, but I won't have any flash crap wasting my time unless I specifically request it.
Adblock is still useful to remove other offending items, but I don't end up blocking every flash item I see anymore.
Yeah, but my juristiction doesn't have an employment at will policy.
Either way though, if you're required to stay, it's not unreasonable to expect to be paid for your time.
Not if it only lasts 10 minutes...
Worse then if 4 50s were fakes?
The scary part is that if she's just type the amount into the register, it would do the math for her and she might even learn something.
Chances are as soon as they called the manager would say it's all good, he knows about the overage.
More importantly though, why didn't the employee count their cash at the beginning of their shift and notice they were over?
I'm not sure about the law where you live, but here if you are required to wait as a function of your job, you get paid for your waiting time. Period.
If that means you go over 8 hours a day or 40 hours a week you get overtime too.
Not only that, but with a couple exceptions (banking is one, IIRC), you cannot be penalized for refusing a shift change without 48 hours notice. In other words, you can walk out at the end of your shift and cannot be fired for doing so.
In theory, I agree. In practice, I'm be afraid of sticking too many devices on USB power simply because I don't trust cheap motherboard manufacturers to be able to handle the load. There is also a significant risk of a cheaply made device shorting out and taking out the motherboard as well as every other connected device.
That being said, it does drive me nuts how many USB devices have their own brick supply too. In fact, other then my cordless phone, I think every brick attached to a power bar on the floor beside my PC is connected to a device connected via USB to my system.
One other concern with having USB supplied power -- This isn't practical for all devices, many devices need power when the system is off too. My cordless mouse (rechargable), Palm, PPC and digital camera all would need power when the system is off. I'd be annoyed if my SB Extigy turned off with the system too, it's nice having it passthrough audio from line-in to my speakers even when my PC is off.
Also consider that even if power was available when the system is off, many devices really should turn themselves off when the system is off. It would be nice to cut power to these devices, but that would mean adding yet another pin (one for always-on power, one for switched power)
Now admittedly many systems (mine included), still provide USB power when the system is off. However, looking at my power supply's load limits, the power supplied when the power supply is off is minimal, and the hardware mentioned earlier would go over my power supply's standby maximums.
This could be resolved as part of the specification by requiring another connector from the power supply to the motherboard (to show the motherboard that the power supply is capable of supplying higher demand then the current ATX standby power). If this new connector wasn't connected then the motherboard would only supply power via USB when the system is running. As a workaround, some external USB hubs could have power connectors that would supply downstream devices with power even when the system is off.
All of this is a lot more hassle then standardizing the connections used by bricks connecting to external hardware. If each connector size corresponded to a specific voltage and minimum amp rating, bricks would become interchangable, and it would become feasible to buy a single power supply that would feed all of your power hungry gadgets without dinking around with sizing the power connector.
No -- Verio is listed, you're not.
If you have your own block of IPs (Assigned by ARIN), you won't get listed just because your traffic goes through Verio.
Well, not by SPEWS, anyway. Others may choose alternatives.
Sure IF the publisher is claiming that all IPs in the block are spammers.
Remember that truth is 100% defense against libel.
I don't see many, if any, DNSbls claiming they only contain spammers. The RBL includes a lot more then spam sources, it also includes any form of spam support, which is a lot easier to prove then if they said "This IP is a spammer." as long as they follow their own listing criteria.
A DNSbl being offline will always return negative results NEVER a positive result.
A DNSbl lookup just takes the IP in 1.2.3.4 format and does a query for 4.3.2.1.dnsbl.example.com
The only way you can get a positive result on a DNSbl lookup is if you get an IP returned. Typically 127.0.0.2 was used, although some DNSbls include feedback in the status either in bit or by response code.
If something goes wrong in the lookup (DNS resolved crashes, DNSbl server is down, DNSbl server is overloaded, port 53 is firewalled, or you've simply entered a wrong DNSbl host), you'll virtually always get a "not-listed" result back.
Going back to the RFC, a DNSbl error would tend to be a permanent error since retrying delivery attempts over and over will never get the message through. The only way the message will suddenly become deliverable is if the sender takes action to get delisted.
In general, anything that will resolve itself in time or needs resolution on the recipient's side should be a 4xx error, anything that needs sender corrective action should be a 5xx.
I'm not saying that people don't 4xx DNSbl listings all the time, but if they've bothered to read and understand rfc2821's definitions of temporary and permanent they will probably know better.
Either way though, it doesn't really address the issue: If your mail is important enough that you can't tolerate a delay in delivery, either get the recipient to whitelist you, or stick around for a confirmation (either a return receipt, if your recipient is willing, or wait for your recipient to hit reply and say "y0 d00d I got yer message" -- Email is not now and will not ever be a guaranteed delivery mechanism.
The vast majority of my bricks are between 5V and 12V, many being either 5V or 9V.
Some devices obviously couldn't use a central power supply, printers being a good example. Fine. None of my printers have a brick anyway, my laser printers both use standard PC power cables and my inkjet has an external power supply that takes a PC power cable input, so it doesn't block an entire power bar.
It's not the end of the world if a couple devices still have special needs, but why should I need a seperate brick for each device which fits into the common range?
As much as DC-DC power supplies might raise the price slightly, it would be offset by the back that the manufacturer doesn't need to supply a brick.
I'd be more then willing to spend the $150 I just spent on a computer power supply on a magic box that would provide everything from from 3v to 12v with standardized cables if it would remove the rats next of of power bars with 2-3 items connected to each.
Heck, if they'd just standardize on one size of plug for each voltage, those of us who want to invest in a central power supply could do so. Those that don't could stay with the manufacturer supplied bricks.
If the DNSbl claims to only list spammers, then you're correct.
However, if the listing criteria and spelled out and followed strictly, no libel has occured.
The RBL specifically mentions "multi-hop (multiple IP) open relay, a spam source, or a spam support service (e.g., a webserver or nameserver).", which is more then just spammers.
The Spamhaus Exploits Block List (XBL) is an example of one which doesn't even list spammers directly, XBL is "a realtime database of IP addresses of illegal 3rd party exploits, including open proxies (HTTP, socks, AnalogX, wingate, etc), worms/viruses with built-in spam engines, and other types of trojan-horse exploits."
The fact that spammres happen to abuse many of those resources is incidental, but it's one of the most effective lists out there.
Most mail servers are smart enough to reject a message based on a DNSbl (RBL) hit rather then just delete the message.
Most mail servers these days return a 5xx -- That's what I found when my IP range got listed on SPEWS.
A week later my server was on another network. Problem solved.
Sure but you can have multiple UPSes, plus a backup transformer that will convert AC from the grid to DC.
What do you think will be more reliable, a thousand $50 power supplies (one in each server), or $50,000 spent on building a redundant power solution to power the entire data center?
You're already building around a number of potential single points of failure (Everything from the power in from the grid to the UPS(es) to the wire from the UPS to each servers' UPS)
Sure, but if they could settle on a couple standard voltages (5V and 12V anybody?), you could buy one power supply and connect it to all your third party junk.
It would even let manufacturers cheap out by not providing a supply at all, so you'd think they'd be all over this.
Any idea if these panels are cost effective (vs the cost of the panel itself)?
Yes and no. Redundancy is fine if the machines are actually redundant.
However, if you give me an option between having a server with a single power supply or a shared power supply between a dozen serves, all other factors being equal, I'll go for the shared one every time.
Why?
Well, if something goes wrong, they have 12x the incentive to fix it *now*
Then there is the anti-fire supression problem
I believe that would be a "fire suppression system" or an "anti-fire system" -- You don't really need an anti-fire suppression system, you could save yourself a lot of money and not bother installing a fire suppression system in the first place.
If the receiving server responds with a 5xx (permanent error), the sending server should give up immediately it definitely should not try over and over.
Why would they wait 4 days to check their mail?
How about if you wrote a letter and the postman ate it?
How about if the mail server (or mail client, for that matter) had a disk problem and lost the message?
Email is *NOT* has no guaranteed delivery mechanism, it is best-effort every step of the way.
More importantly, the sender would receive a bounce from their SMTP server, so they would know their message didn't get through. They'd call and tell the love of their life over the phone instead.
Then I'd say he has a good incentive to find an ISP who runs a clean shop.
You don't generally get listed on any of the larger lists (spamhaus, SPEWS, MAPS, etc) because a spammer got a hosting account and started spamming.
You get listed because you fail to terminate the spammer, or you fail to prevent the spammer from signing up again (which is substantially the same thing)
It's also worth noting that it's not as though any DNSbl has the ability to actually block mail.
My server refuses mail from IPs on a number of DNSbls because *I* choose to use those DNSbls. If I block an important email it is MY responsibility, no one else's responsbility.
I whitelist important senders so that they get through even if they manage to get listed.
My reject message is very simple: "<DNSBL NAME>: email whitelist@<servername> to be whitelisted" -- When an email is sent to the whitelist account, my server automatically whitelists the sender IP and forward the message to my support inbox. This means any sender who gets blacklisted can get themselves whitelisted without much difficulty and without requiring assistance from anyone.
Sure a spammer could abuse this mechansim, but so far, it hasn't happened. Even if they do, they'd only get to spam for a few hours, and their mail still goes through my other filtering systems.
For example: I don't care who you are if you HELO/EHLO as my server name or my IP, you don't get to send mail to me anymore. Period. This alone catches upwards of 50% of spam attempts.