I remember seeing on a TV program about the Voyager project how serendipitous the timing of the launch was - where you could hop from planet to planet to planet using the gravity well of each planet to jump to the next one. Basically the alignment of the planets when Voyager launched made this possible, and such an alignment isn't going to come around again in our lifetime. So you'd need to build seperate probes to go to each planet, instead of being able to send one probe to many of them.
I've worked at a two different "small IT shops" where the owners offered out shares to hard working employees to keep them around. Not one of those employees is still on speaking terms with the owner of either company. Lawsuits were not involved with the departure of all of these people, but enough of them did end up involved in suits that I personally would never accept stock in a small IT shop where I was working. In fact, watching what said shops did to many close friends - I don't think I would even work as an employee at a small IT shop again!
But we don't throw them out, either. Where I work, all of the old equipment is sent to a company owned warehouse, because someone figured out the cost of just storing all of this equipment is lower than the cost of paying someone to recycle it (and then taking the risk that they pull confidential information off the machines.) And we have the desktops locked down, so there isn't even much interesting content on the drives.
I suppose it's possible that the white house destroys them because they have a way to do so. But if they were really archiving emails on the individual desktops, that's a huge problem in and of itself.
So, why do people still use work for private mail?
Where I work the proxy servers agressivley block any use of webmail. Not just the gmail/hotmail/yahoo variety, but many url's with "mail" in the name. Sometimes when trying to read archives from an internet mailing list to solve a problem it kicks in and you need to get a senior manager to approve an "exception" for that site so you can view it. So unless you want to avoid any personal communication at work, you're stuck using the company email.
and god forbid you have anything actually IN the glove compartment that might get it the way.
Not an issue. The glove box on the 3 series is small enough that if you have anything else in it (other than maybe a pair of gloves, but who keeps them in the glove box), there won't be room for the iPod.
Actually CitiBank has card readers at each teller and a Pin entry keypad next to the reader. Sometimes they ask me to swipe my ATM card and enter the pin there - in one case, when making a deposit(!). Beleive me, if someone wants to roll up $20 worth of lose change and deposit it in my account... why ask them for ID? But they do anyway.
Yes, that was a big deal with the 1571 when it first came out. I never thought much of it, until years later when I got my first PC as hand-me-down. I bought a modem, but had no software for it. It seemed quite lame to sign up for something like AOL (if it even existed at that point, most of my friends had Prodigy) just to download real terminal software. I ended up downloading it on the Commodore and writing it to a PC disk with the 1571.
I found it took a bit more than 2 days for me. At a certain point though I started sleeping a lot better and then I was more willing to put up with the exhaustion late in the day (which was quickly going away at that point).
One person I worked with quit a few months before me, was caffine free for months... had a cup of coffee one morning because it was a freezing day and he wanted to warm up... and was back on it on a regular basis. I decided to drink one caffinated beverage a week just to make sure I didn't become so sensitized to it that I could end up in that position. I have to do this before 1 PM now, though, or it keeps me up the following night. There's a brief desire for more about 24 hours later. But at least I know that I'm not going to end up going back to the volumes I was at.
I work a financial firm. All attachments are pulled out of the messages by the firewall software they use and put in a quarrantine. The receipient receives the message, sans attachments, along with an explanatory note that if they need the attachments they need to send a message to the keeper of the mail system with a business reason why they need those attachments. Those people will then send them the actual attachments.
Even doing it with automated systems isn't new. At one point MCI had a system (this is when answering machines were fairly rare) that would take the beep from the answering machine as a yes. No modification to the OGM required - just find a number with a machine and hope they don't wonder what these messages that say "Thank you for using MCI" are.
As someone who has used several versions of both Sybase and SQL Server... the code bases hae diverged a lot since they split up. I've found this kind of annoying, because it means Sybase works almost, but not exactly like you'd expect if you've been using SQL server for a while. For instance, the following syntax will not work in Sybase, but has worked in SQL server since at least 6.5:
select * from (select * from pubs..books) q
Admittedly, a contrived example of the syntax, but quite annoying one the less. There are also important diversions in the way they handle tempdb, identity columns, and statistics.
At one of my jobs the normal practice is to have the middle tier developers provide short acronym codes for the error conditions, and let the UI designers provide the error message the user will see based on these codes. The UI designers can then code several lines of help text explaining how the user might fix the problem.
Part of the process involves the middle tier developer sending the UI developer a list of the codes and what they mean. It can provide a scary look at what the error messages would say if developer wrote them. The best one I can remember seeing:
A few jobs ago, I had a client (who was a DJ) that wanted exactly this written. Except he wanted to use CDJ (program that comes with the Slink-E), and have the MP3's go into CDJ.
We had a sound card with a TossLink input that did an outstanding job of recording from the Sony changers.
CDJ made it easier to get the track names since it handled the CDDB stuff, and it had an automation interface so we could control it (and the CD player) from the program.
We built a version in about a week that would rip two sample CD's we'd made for the purpose.
When we went to production we had various problems - mainly, with the Slink-E, there are enough delays that you can't guarantee you are starting the recording at the right place. The gap isn't always the same, so you can't add a set amount to the beginning.
The end is even worse - For a track reported as 2:30 by the player when it started, sometimes 2:30 was the last time indication we'd see. Other itmes, 2:29 was the last indication. This made it very hard to stop recording at the right place.
I don't know if they ever used this or not, because I don't work there anymore. I think if I were doing it I'd use a computer, since you can rip faster than 1x.
The bottom line is, I think TiVo is doing a very good job of making both the viewer and the network happy. The sonicBlue box can share over the internet out of the box - on the TiVo, mods are necessary that put this out of the reach of Joe Blow on the street.
And TiVo underground has "secret" remote sequences.. once of which gets you a 30 second skip button on your TiVo. I've actually gone back to using the Fast Forward button instead. The automatic backup when you exit makes it easier.
The signal strength has to be very high. I remember hearing stories about WGY AM in Schenectady, NY doing an experiment where they substantially increased their already high power levels at their transmitter. (They're one of the original stations, back then selling radio spectrum wasn't a big money business for the government so they issued licenses for much higher power levels. They'd never do that now since lower power means that same frequency can be sold again to another station in another are many more times.) Anyhow, on the even higher power, people were picking up the radio on fillings, braces, electric fences, and all kinds of other things... people who were happy this was just a temporary experiment.
I once did sort of the opposite. I had robots.txt setup so that it would log the IP and return a bunch of directories with confidential sounding names. (Since it was on a public site for my company I obviously didn't want to lock out the real search engines.) The files in the robots.txt file were there, as a sort of honey pot. Password files? Sure, in standard unix format. Except the passwords don't go to anything and if you run crack on them, you'll find it takes a long, long time - since they aren't real passwords we encrypted strings of letters,numbers, and symbols that aren't in any known crack dictionary. One IP downloaded the file three times! But we never saw the fake passwords being used at the honeypot linux box. I guess he got tired of waiting.
Must be an Albany graduate. For some reason (at least when I was there), csh was your default shell. If you were lucky you got it changed to tcsh before they disabled ypchsh for "security reasons". Bash? Nobody used it. I remember setting up Linux for the first time back then, trying to figure out how the heck environment variables worked in Bash. I knew this was a goddam simple question, but nobody at school knew because everyone there had only seen C Shells. I also knew that asking a question like "How do I set environment variables in Bash" on the net would not get me the answer. But at the time, a file that said "FOO=bar ; export FOO" or "export FOO=bar" was not that easy to find. I don't know why they continued to use CSH scripts for everything even when it appears everyone else in the world uses bash.
I used to work at a company that sold computers. Two major UPS experiences come to mind.
The first was a computer shipped from Albany, NY to florida. The box arrived with some minor bumps, but it was packed well and there was no damage to the interior case. But there was this odd ratling sound and it wouldn't boot up.
Turns out that the machine had received such an impact that it had knocked the CPU and heat sink/fan out of the socket. This was a pentium system with a ZIF socket... consider the amount of force required to get a chip out of once of those sockets when the lever is set to the LOCK position. Of course, diagnosing this problem over the phone with someone who had no idea what was inside the computer...
Another machine was shipped downstate. It arrived dented so badly the case didn't stay together. The client filed a claim, and UPS inspected it, then sent it back to us so we could get them an estimate to repair the damage. (The fact that the client was the state of NY may have impacted the way they handled this.) The way UPS packed it to ship to us was interesting - a box nearly five times the size usually used to ship a mini tower. Inside were these form fitting foam pieces. Not just corner blocks - the entire thing. It was some kind of expanding foam that expanded to a certain point and froze in place.
Just today, I had a different problem with UPS which makes me less likely to deal with them in the future. Namely, the driver doesn't feel like coming to my address today so he marks "Nobody Home" and doesn't even bother to show up.
Things I've learned over the years:
- Insure the package. The more fragile it is, the more insurance. As my boss used to say "Insure it enough to scare them." It works, too.
- Just because the tracking info says "Out For Delivery", don't take the day off and wait for the package. The driver may decide he doesn't feel like delivering it.
- You can never over-pack something
We're developing Java applications. They claimed that they couldn't support the deployed application if we all had different machines because we might do machine specific stuff. Never mind that we developed on NT and they deployed on Unix and we already had case-sensitivity issues popping up all over the place.
They never got around to actually locking the dev machines down, because there was too much backfighting when they tried. Most of the developers said that they would fix whatever problems they created. IT tried to use a management tool to push new software onto peoples machines because "people usually mess up their machines when they install their own software." The second time they used it, they ended up wiping all of the apps and registry entries off several machines. We kind of put the brakes on them using that software after that.
Since then, they've been heavilty hit by layoffs and they no longer complain when we want access to things that will only reduce their workload.
But to answer the original question, no, you can't develop anything that way.
So, what you're saying is that you don't mind a core dump of your computer being sent to Microsoft every time it crashes? That you don't mind Microsoft saying "That GNU program you're developing... it crashes too much. You may never run it again."
Then the developer should include a version with the JRE inside it. There are lots of apps out there that need MSVBVM60.DLL to work, and nobody ever complains about having to download that. (Probably since a lot of apps come with it anyway). If enough Java apps include the JRE then the machine will eventuall have it anyway.
It's not so much stopping NAT as detecting NAT and then demanding that you stop using it. How do they detect it?
- The port numbers that the connections come from are much, much higher (over 10000)
- If you have mixed machines, and they have a transparent proxy server - for instance, you normally wouldn't see a Linux version and a Windows version of Netscape running on the same IP within seconds of each other.
Thats how you would think it would work. I've actually seen this method described several times over the years and it amazes me that people haven't fixed it yet.
Personally, I prefer to send the data to the client.. so that I can compare the results coming back with those on the server. Then you can detect people playing with the hidden values in advance.
CA does the same stupid thing. Setting their software up on a machine that has no net access requires reading 26 digit numbers with no hypenation to someone over the phone without making a mistake and then having them read you back a huge number. In short, a complete pain. Change a piece of hardware that is used to id the machine and the license becomes void, another phone call. The problem is, if MS does this, the call volume will be incredible. How do they prove that you didn't get a new PC? They'll just have to give you a new license. Ok, it cuts down some, but It'll be a logistical nightmare for them.
At my last job one of our big clients was a Competitive Local Exchange Carrier (CLEC) that provided telephone and internet service to office buildings in a package deal. We built a system to generate phone bills of their switches. After that was done they wanted to meter internet usage. We figured this was the traditional dollars/byte type of thing that some high bandwith customers get. Then we found out they wanted to have the price for meg be different depending on weather the content was Data, Voice, or Video! Now, ignoring for the moment the fact that their cost to provide a T1 is fixed regardless of weather data or video goes over it... What we basically discovered after brining several experts in on it was that there just wasn't any reliable way to decide what category a packet fell into. Several commercial solutions were demoed to us but we were always able to bypass the video/voice metering rate and pump video through at the data rate. I'd say the same applies here. On an aggregation device of any meaningful size, it is simply not practical to examine every single packet for say, H.323. And that's just one encoding!
Slashdot Mirror
I remember seeing on a TV program about the Voyager project how serendipitous the timing of the launch was - where you could hop from planet to planet to planet using the gravity well of each planet to jump to the next one. Basically the alignment of the planets when Voyager launched made this possible, and such an alignment isn't going to come around again in our lifetime. So you'd need to build seperate probes to go to each planet, instead of being able to send one probe to many of them.
I've worked at a two different "small IT shops" where the owners offered out shares to hard working employees to keep them around. Not one of those employees is still on speaking terms with the owner of either company. Lawsuits were not involved with the departure of all of these people, but enough of them did end up involved in suits that I personally would never accept stock in a small IT shop where I was working. In fact, watching what said shops did to many close friends - I don't think I would even work as an employee at a small IT shop again!
But we don't throw them out, either. Where I work, all of the old equipment is sent to a company owned warehouse, because someone figured out the cost of just storing all of this equipment is lower than the cost of paying someone to recycle it (and then taking the risk that they pull confidential information off the machines.) And we have the desktops locked down, so there isn't even much interesting content on the drives.
I suppose it's possible that the white house destroys them because they have a way to do so. But if they were really archiving emails on the individual desktops, that's a huge problem in and of itself.
So, why do people still use work for private mail?
Where I work the proxy servers agressivley block any use of webmail. Not just the gmail/hotmail/yahoo variety, but many url's with "mail" in the name. Sometimes when trying to read archives from an internet mailing list to solve a problem it kicks in and you need to get a senior manager to approve an "exception" for that site so you can view it. So unless you want to avoid any personal communication at work, you're stuck using the company email.
and god forbid you have anything actually IN the glove compartment that might get it the way.
Not an issue. The glove box on the 3 series is small enough that if you have anything else in it (other than maybe a pair of gloves, but who keeps them in the glove box), there won't be room for the iPod.
Actually CitiBank has card readers at each teller and a Pin entry keypad next to the reader. Sometimes they ask me to swipe my ATM card and enter the pin there - in one case, when making a deposit(!). Beleive me, if someone wants to roll up $20 worth of lose change and deposit it in my account... why ask them for ID? But they do anyway.
Yes, that was a big deal with the 1571 when it first came out. I never thought much of it, until years later when I got my first PC as hand-me-down. I bought a modem, but had no software for it. It seemed quite lame to sign up for something like AOL (if it even existed at that point, most of my friends had Prodigy) just to download real terminal software. I ended up downloading it on the Commodore and writing it to a PC disk with the 1571.
I found it took a bit more than 2 days for me. At a certain point though I started sleeping a lot better and then I was more willing to put up with the exhaustion late in the day (which was quickly going away at that point).
One person I worked with quit a few months before me, was caffine free for months... had a cup of coffee one morning because it was a freezing day and he wanted to warm up... and was back on it on a regular basis. I decided to drink one caffinated beverage a week just to make sure I didn't become so sensitized to it that I could end up in that position. I have to do this before 1 PM now, though, or it keeps me up the following night. There's a brief desire for more about 24 hours later. But at least I know that I'm not going to end up going back to the volumes I was at.
I work a financial firm. All attachments are pulled out of the messages by the firewall software they use and put in a quarrantine. The receipient receives the message, sans attachments, along with an explanatory note that if they need the attachments they need to send a message to the keeper of the mail system with a business reason why they need those attachments. Those people will then send them the actual attachments.
Even doing it with automated systems isn't new. At one point MCI had a system (this is when answering machines were fairly rare) that would take the beep from the answering machine as a yes. No modification to the OGM required - just find a number with a machine and hope they don't wonder what these messages that say "Thank you for using MCI" are.
Admittedly, a contrived example of the syntax, but quite annoying one the less. There are also important diversions in the way they handle tempdb, identity columns, and statistics.
At one of my jobs the normal practice is to have the middle tier developers provide short acronym codes for the error conditions, and let the UI designers provide the error message the user will see based on these codes. The UI designers can then code several lines of help text explaining how the user might fix the problem.
Part of the process involves the middle tier developer sending the UI developer a list of the codes and what they mean. It can provide a scary look at what the error messages would say if developer wrote them. The best one I can remember seeing:
NDFAIL - Operation Succesfully failed.
A few jobs ago, I had a client (who was a DJ) that wanted exactly this written. Except he wanted to use CDJ (program that comes with the Slink-E), and have the MP3's go into CDJ.
We had a sound card with a TossLink input that did an outstanding job of recording from the Sony changers.
CDJ made it easier to get the track names since it handled the CDDB stuff, and it had an automation interface so we could control it (and the CD player) from the program.
We built a version in about a week that would rip two sample CD's we'd made for the purpose.
When we went to production we had various problems - mainly, with the Slink-E, there are enough delays that you can't guarantee you are starting the recording at the right place. The gap isn't always the same, so you can't add a set amount to the beginning.
The end is even worse - For a track reported as 2:30 by the player when it started, sometimes 2:30 was the last time indication we'd see. Other itmes, 2:29 was the last indication. This made it very hard to stop recording at the right place.
I don't know if they ever used this or not, because I don't work there anymore. I think if I were doing it I'd use a computer, since you can rip faster than 1x.
The bottom line is, I think TiVo is doing a very good job of making both the viewer and the network happy. The sonicBlue box can share over the internet out of the box - on the TiVo, mods are necessary that put this out of the reach of Joe Blow on the street. And TiVo underground has "secret" remote sequences.. once of which gets you a 30 second skip button on your TiVo. I've actually gone back to using the Fast Forward button instead. The automatic backup when you exit makes it easier.
The signal strength has to be very high. I remember hearing stories about WGY AM in Schenectady, NY doing an experiment where they substantially increased their already high power levels at their transmitter. (They're one of the original stations, back then selling radio spectrum wasn't a big money business for the government so they issued licenses for much higher power levels. They'd never do that now since lower power means that same frequency can be sold again to another station in another are many more times.) Anyhow, on the even higher power, people were picking up the radio on fillings, braces, electric fences, and all kinds of other things... people who were happy this was just a temporary experiment.
I once did sort of the opposite. I had robots.txt setup so that it would log the IP and return a bunch of directories with confidential sounding names. (Since it was on a public site for my company I obviously didn't want to lock out the real search engines.) The files in the robots.txt file were there, as a sort of honey pot. Password files? Sure, in standard unix format. Except the passwords don't go to anything and if you run crack on them, you'll find it takes a long, long time - since they aren't real passwords we encrypted strings of letters,numbers, and symbols that aren't in any known crack dictionary. One IP downloaded the file three times! But we never saw the fake passwords being used at the honeypot linux box. I guess he got tired of waiting.
Must be an Albany graduate. For some reason (at least when I was there), csh was your default shell. If you were lucky you got it changed to tcsh before they disabled ypchsh for "security reasons". Bash? Nobody used it. I remember setting up Linux for the first time back then, trying to figure out how the heck environment variables worked in Bash. I knew this was a goddam simple question, but nobody at school knew because everyone there had only seen C Shells. I also knew that asking a question like "How do I set environment variables in Bash" on the net would not get me the answer. But at the time, a file that said "FOO=bar ; export FOO" or "export FOO=bar" was not that easy to find. I don't know why they continued to use CSH scripts for everything even when it appears everyone else in the world uses bash.
I used to work at a company that sold computers. Two major UPS experiences come to mind.
The first was a computer shipped from Albany, NY to florida. The box arrived with some minor bumps, but it was packed well and there was no damage to the interior case. But there was this odd ratling sound and it wouldn't boot up.
Turns out that the machine had received such an impact that it had knocked the CPU and heat sink/fan out of the socket. This was a pentium system with a ZIF socket... consider the amount of force required to get a chip out of once of those sockets when the lever is set to the LOCK position. Of course, diagnosing this problem over the phone with someone who had no idea what was inside the computer...
Another machine was shipped downstate. It arrived dented so badly the case didn't stay together. The client filed a claim, and UPS inspected it, then sent it back to us so we could get them an estimate to repair the damage. (The fact that the client was the state of NY may have impacted the way they handled this.) The way UPS packed it to ship to us was interesting - a box nearly five times the size usually used to ship a mini tower. Inside were these form fitting foam pieces. Not just corner blocks - the entire thing. It was some kind of expanding foam that expanded to a certain point and froze in place.
Just today, I had a different problem with UPS which makes me less likely to deal with them in the future. Namely, the driver doesn't feel like coming to my address today so he marks "Nobody Home" and doesn't even bother to show up.
Things I've learned over the years:
- Insure the package. The more fragile it is, the more insurance. As my boss used to say "Insure it enough to scare them." It works, too.
- Just because the tracking info says "Out For Delivery", don't take the day off and wait for the package. The driver may decide he doesn't feel like delivering it.
- You can never over-pack something
We're developing Java applications. They claimed that they couldn't support the deployed application if we all had different machines because we might do machine specific stuff. Never mind that we developed on NT and they deployed on Unix and we already had case-sensitivity issues popping up all over the place.
They never got around to actually locking the dev machines down, because there was too much backfighting when they tried. Most of the developers said that they would fix whatever problems they created. IT tried to use a management tool to push new software onto peoples machines because "people usually mess up their machines when they install their own software." The second time they used it, they ended up wiping all of the apps and registry entries off several machines. We kind of put the brakes on them using that software after that.
Since then, they've been heavilty hit by layoffs and they no longer complain when we want access to things that will only reduce their workload.
But to answer the original question, no, you can't develop anything that way.
So, what you're saying is that you don't mind a core dump of your computer being sent to Microsoft every time it crashes? That you don't mind Microsoft saying "That GNU program you're developing... it crashes too much. You may never run it again."
Then the developer should include a version with the JRE inside it. There are lots of apps out there that need MSVBVM60.DLL to work, and nobody ever complains about having to download that. (Probably since a lot of apps come with it anyway). If enough Java apps include the JRE then the machine will eventuall have it anyway.
It's not so much stopping NAT as detecting NAT and then demanding that you stop using it. How do they detect it?
- The port numbers that the connections come from are much, much higher (over 10000)
- If you have mixed machines, and they have a transparent proxy server - for instance, you normally wouldn't see a Linux version and a Windows version of Netscape running on the same IP within seconds of each other.
Thats how you would think it would work. I've actually seen this method described several times over the years and it amazes me that people haven't fixed it yet.
Personally, I prefer to send the data to the client.. so that I can compare the results coming back with those on the server. Then you can detect people playing with the hidden values in advance.
CA does the same stupid thing. Setting their software up on a machine that has no net access requires reading 26 digit numbers with no hypenation to someone over the phone without making a mistake and then having them read you back a huge number. In short, a complete pain. Change a piece of hardware that is used to id the machine and the license becomes void, another phone call. The problem is, if MS does this, the call volume will be incredible. How do they prove that you didn't get a new PC? They'll just have to give you a new license. Ok, it cuts down some, but It'll be a logistical nightmare for them.
At my last job one of our big clients was a Competitive Local Exchange Carrier (CLEC) that provided telephone and internet service to office buildings in a package deal. We built a system to generate phone bills of their switches. After that was done they wanted to meter internet usage. We figured this was the traditional dollars/byte type of thing that some high bandwith customers get. Then we found out they wanted to have the price for meg be different depending on weather the content was Data, Voice, or Video! Now, ignoring for the moment the fact that their cost to provide a T1 is fixed regardless of weather data or video goes over it... What we basically discovered after brining several experts in on it was that there just wasn't any reliable way to decide what category a packet fell into. Several commercial solutions were demoed to us but we were always able to bypass the video/voice metering rate and pump video through at the data rate. I'd say the same applies here. On an aggregation device of any meaningful size, it is simply not practical to examine every single packet for say, H.323. And that's just one encoding!