$10K? I could use that! Since I have no career to trash (I write software), this would be free money. I've read the book, State of Fear - so this should be a cinch... I know what to say and how to say it!
I don't want radio at all, much less by subscription. Podcasts let me listen on my schedule, to exactly what i want to listen to. Currently, that's books, short SciFi stories, and science centering on astronomy. In my queue is an astronomy course.
And, i don't want to listen to these just in my car. So all i want in my car is an amp with an Aux-in jack. Not even AM. I don't want a hard disk or anything in the car. That's just more to manage. When i listen to something on my mp3 player, i want to be able to delete it from there if that's what i want. I don't want two places to do that.
My car has developed an intermittant problem. It seems to manifest in cold weather. The car's idle speed will hang out at 2,000 RPM, even when the engine is hot (i drive an hour to work) (Normal idle for this car is 600 RPM). The solution is to turn the car off, and turn it back on. Sound familiar?
Perhaps it isn't so much an analogy as the fact that my car's idle is governed by a computer.
A new technique, involving looking for eclipsing binary stars in the target galaxy, has been used to measure the distance to m33. m31 is next. This technique is still based on the brightness of stars, but the intrinsic brightness is supposed to be better known.
Also of interest is that proper motion has been measured for m33. No proper motion has yet been detected for m31.
Coffee, with caffeine, can have anti-oxidant effects, can keep you awake and alert, and give you a little rush
but
caffeine can screw up your sleep cycle, is addictive - complete with withdrawl symptoms, can give you arthritis and osteoporosis, stays in your body for a long time, has cumulatative effects.
---
In the US, food containing caffeine has to say it has it, but does not have to say how much. This makes it more difficult to self medicate.
A colleague complained that she needed to stay awake, but didn't want to take too much, for fear of not getting to sleep at night. I computed it down to 2 oz of Coke. I suggested she by a 12 oz, drink 2 oz, and toss the rest. People don't take my advice very often.
Wouldn't take that much aspartame to kill me. I'm alergic to it. It's a royal pain. Aspartame is getting into everything. I have to check all the time.
You are looking for a slight dimming. The atmosphere of the Earth can provide slight dimming. Two cameras looking through different bits to of the Earth's atmosphere can provide extra confidence that the dimming you detect is not here at home.
By using a command like
mke2fs -j -b 1024 -m 0 -N 2000000/dev/hdb1
one can have a 1K sized blocks. I argue that with mostly continous filesystems, the block size is largely irrelevant to performance. A small block size reduces the wasted space at the ends of files. For me, going from 4K to 1K gave me 4 GB more free space on my 160 GB drive for my current files - 2.5%.
System supported file compression would yield more, of course. Sure, the mp3 & jpg files won't compress, but I have lots of text files. A good compression system will know this and not bother compressing files that are incompressible. It might even use a file like magic cookie mechanism so it doesn't have to attempt compression to find out it's pointless.
If the improvement in space due to consolodated ECC codes and other overhead saves more total space than the end-of-file wastage, it would still be a win. The article doesn't say how much better the new standard will be.
It may be that Linux will provide 1K filesystem blocks on top of the 4K phsyical blocks. Performance will be worse. However, the original ext filesystem provided half K blocks, and that option is now moribund.
For new systems, this standard is fine. However, I run my machines into the ground, and the better ones have lasted fifteen years. Given that my current machine could last another ten years, it would be a shame to have to toss it into the landfill in five years because the disk drives can't be replaced. Progress is good. Forced upgrades are not.
I run multiple tasks at a time frequently, under Linux. My currently machine is three years old, and wasn't high end at the time. It's plenty fast. The bottleneck is disk. While a disk to disk copy is in progress, such as during backup, the user interface grinds to a crawl. The IDE ATA disks suck down CPU, true enough, but also, the controllers are so tied up doing the copy that other tasks have a hard time getting through. SCSI performs better, but currently, SCSI drives are priced out of the market. Multiple CPUs, same core or not, don't solve this problem. How are USB or Firewire drives on cost and performance?
Generally, i'm happy with current performance. The system can display full screen real time video. I have no requirements past this. My system is also upgradable.
I tend to mean more than one thing at a time. At one time, The Beatles (Apple Records) sued Apple Computer. In the settlement, Apple Computer agreed not to get into the music business.
I'd like Apple to get a clue. They're a nice company to deal with. I want a single device in my home entertainment center. It does audio and video time shifting, displays content, allows editing, burns media, everything. And, any compentent system will not produce any audible sound. I do not want a media hub, a TiVo, a VHS VCR, a DVD player, etc. I'd be willing to compromise and say that the display, speakers and amp can be external components. Though, i'll want the remote to include volume control.
Perhaps in their zeal to provide content, they neglected to provide functionality. Many assumed that the new Mac mini would, among other things, provide Tivo functionality. It doesn't. Yet, it connects to your TV, beams content around your house, and otherwise has everything it needs to be a Tivo, except software.
It is starting to look as if the Beatles were right after all. Apple should not be in the content business. Perhaps it is time for someone with a clue to get into the appliance business.
My mother-in-law had one of her heart valves repaired in open heart surgery a few years ago. I was amazed. They said that if they couldn't sew it back together, that they would put in a pig's valve, with somewhat higher risk factors.
Makes you want to live near a hospital. See? City living is healthier after all!
My last laptop was a Compaq Aero - a 3.5 lb 486/25 with 16 MB RAM, 170 MB disk, 640x480 with 64 greys. The only way to install on it was by floppy, so a stack of disks for an old Slackware distro was cut. After a few months, extraneous stuff was deleted, and the result was a system with Postgres, Apache, perl, gcc, emacs running under X (fvwm2) with 70 mb of disk free for my development (so, 100 MB for the system included a 16 MB swap file). I ran PLIP for networking (ssh'ed xterms or NFS) to a desktop, and used CVS to coordinate my development projects between systems.
I still have my resulting system image... If I could spec out a replacement, it'd have flash disk, use AA batteries running at least 30 hrs per charge, and would be fanless.
My current oldest systems are 486/33's with 16 MB ram, 200 MB disk. However, even 100 of them in a cluster does not match the computes of a low end Athlon. They can be OK when dedicated to some task, like a fax server. 486's do not have CPU fans.
It takes less energy to leave the flourescent light on than to turn it off, then back on later.
I imaging this has to do with turing off the gym lights for five minutes between classes. The claim is that starting them back up takes alot of juice.
My application is household 40 watt tube flourescent lights, and incandescent replacement screw-in lightbulbs. Same idea - turn it off for five minutes vs. leaving it on.
In the '60s, probably centered on the CDC 6600, an idea was promoted that a balanced' computer would have capabilities in a ratio. The 3 M's was one with 1 MIPS, 1 mbps, and 1 megabit of memory. So, it could executed 1,000,000 instructions per second, communicate to disk at 1,000,000 bits per second (100,000 bytes per second) and had 1,000,000 bits of RAM (one system had 130,000 bytes of RAM, for example).
The box i'm using to edit this note executes on the order of 1 GIPS, with 100 mbps, and 10 gigabits of memory. That is 1,000,000 instructions per second, 100,000,000 bits per second (10,000,000 bytes per second) to disk, and has on the order of 10,000,000,000 bits of RAM (1 GB). (These numbers are rounded, and, no, i'm not terribly interested in my-box-is-faster-than-yours pissing matches - its just an example).
So, if communications speeds will be 1,000,000,000,000 bits per second anywhere by 2010, that implies a computer with at lest 10 GIPS and 10 GB RAM - which doesn't seem that unlikely in five years.
Oddly enough, I'm hoping to still be running this box in five years. Its only two years old, and I don't really want to get a new one. That is, i don't want to spend the money to replace it. More importantly, i don't want to do the administration involved to get a new machine up and running with my current set of capabilities. I ran my 1987 Machintosh II as my primary machine for over ten years and the hardware lasted an additional five years (and counting) to allow for transfer of data. It pisses me off that my most long-lived x86 based PC has lasted only five years. So, i've just finished migrating from the Mac to Linux, and the Mac (with OS/x) now appears to be the better choice (low administrative maintenance) again.
With the recent announcement of low power PPC chips, perhaps Apple will abandon its move to the x86 hardware platform. Still, i've been pretty happy so far with my low-end Athlon's performance and reliability. Who knows? Perhaps i'd be happy with OS/x on AMD.
Much of the energy used is for air conditioning. One might think that this would be easy for data centers in Michigan, but I've worked for places that heat the building and air condition the data center in January. One place had the data center a/c die, and a box fan in a window allowed everything to run. A box fan has to be cheaper to run than a/c. So what we might see is smarter environmental control. At least in the winter, it makes sense to run outside air in, and use the waste heat to heat the rest of the building.
Mac OS 6.4 is the OS the let me cut the Windows cord. Oh, that's right, in 1987, PCs came with DOS. I never did dual boot my Mac, though I considered running 68k BSD.
At first, my PC, a 386/33 running Linux kernel 0.97 - Slackware, was used for browsing the internet. When Open Office became available, I slowly moved everything I cared about from the Mac. I hardly ever boot the Mac anymore. There still is the project to convert all my old Word docs to RTF or some other open format.
Since OS/x, it looked to me that Apple finally had a product worth owning. I couldn't wait, which is why I run Linux. But when 'normal people' tell me they want a new computer, or tell me the the latest Windows worm/virus horror story, I tell them that they should get a Mac. If Linux required less admin, I'd tell them to use it. So, as a Linux bigot, I still tell people to use a Mac.
If Dvorak would say interesting things, I'd be tempted to read them.
The weather here near Detroit, MI, was clear - an unusual event by itself. If Detroit were much further west, then the eclipse would have happened after the moon set. As it was, it set while the heart of the eclipse was just getting started.
It was a partial lunar eclipse, not the full monty. The max size of the shadow of the Earth on the Moon was predicted to happen around 8 AM EST, and really just a small bite at that. Moon set was about 7:45 AM EST, local time. Oddly, though one might expect the orange hue due to being on the horizon, and due to light filtering through the Earth's atmosphere onto the Moon like a million sunsets at once. These effects were quite apparent at the last full lunar eclipse visible in these parts. However, none of that was apparent this time. Perhaps the rising Sun washed it out.
Hacking is programming, which is computer science, which isn't science at all, but rather engineering. So the question is:
Is engineering science?
As an engineer who hacks, no. However, I remain scientifically literate.
Since the scientifically literate form a small minority, perhaps I could get special deals as an oppressed minority. The handicapped got parking spaces, and what did they do to get that?
About 15 years ago, i performed POSIX testing for Interactive Unix - an x86 System Vr3 system. I haven't done anything with LSB, but basically nothing surprized me in Ulrich's article. Mostly, I'd go further. And yet, my conclusions differ somewhat.
The testing process was to run a test, and when it failed, try to figure out if the problem was in the test suite or the tested code. Simple enough.
The tests certainly at some point worked.
No. That wasn't the case. I found myself fixing obvious bugs in the test suite, then attempting to use the fixed version against the target. It was often clear that the test suite could never have worked.
Some distributions still somehow manage to pass the test suits of a new version of the spec. And all this without the people reporting any problems and requesting waiving the test.
We'd report the bugs, with suggested fixes, but we could not wait for fixes to come back and retest. We had to plow forward. We claimed compliance when we had a test we thought tested the assertions and passed it. We never asked for a waiver.
Another nice things we came across during the LSBv3 testing are numerous timing problems.
Been there. Done that, though I didn't have to find some slow machine.
What is the value of such a certification? What assurance does this give you? Is don't use fast SMP machines an acceptable answer in any universe, especially when it comes to thread tests?
If you have need of slow machines, I can provide approximately 25 working 486/33's. I'd put this on his blog, but he doesn't allow comments. I thought this was strange, because I use livejournal primarily as a place where people can comment. However, he talks about his choice there, too. To each their own.
It is not possible to achieve the goal of 100% binary compatibility...
All good points. And its worse than that. Yet, the exercise was valuable. For us, it uncovered many bugs in SVr3. Many. This was ultimately a good thing for our customers.
We were also a Unix porting house. We fixed lots of bugs in our prior ports of Unix. We offered our fixes to AT&T for free. They declined. We had to apply our fixes to each port - without the benefit of CVS. And, we had thousands of patches. And all this for a basically stable system. It was around then that I was convinced of the incredible inefficiency of propietary software. This would never happen to gcc.
My advise: but the losses.
I read this as "My advice, cut the losses." Oddly, many versions of this mispelling pass my
spell checker. Ulrich needs an editor. Perhaps
I'll volunteer. Perhaps he can check my work.
Will you be a swap editor for me? I'll check
your work, you check mine.
So, i agree that the test suite was a horrible idea from the idea that one might assure customers that their old software will still run, or will run on compatible platforms. I agree that the last bug will not be found. However, that is not an excuse to give up the search.
Slashdot Mirror
$10K? I could use that! Since I have no career to trash (I write software), this would be free money. I've read the book, State of Fear - so this should be a cinch... I know what to say and how to say it!
Considering that the plan is to make the next landing in 2019 (50 years after Buzz and Niel were there), the 2010 date is an error.
I don't want radio at all, much less by subscription. Podcasts let me listen on my schedule, to exactly what i want to listen to. Currently, that's books, short SciFi stories, and science centering on astronomy. In my queue is an astronomy course.
And, i don't want to listen to these just in my car. So all i want in my car is an amp with an Aux-in jack. Not even AM. I don't want a hard disk or anything in the car. That's just more to manage. When i listen to something on my mp3 player, i want to be able to delete it from there if that's what i want. I don't want two places to do that.
My current favorite computer/car analogy is this:
My car has developed an intermittant problem. It seems to manifest in cold weather. The car's idle speed will hang out at 2,000 RPM, even when the engine is hot (i drive an hour to work) (Normal idle for this car is 600 RPM). The solution is to turn the car off, and turn it back on. Sound familiar?
Perhaps it isn't so much an analogy as the fact that my car's idle is governed by a computer.
A new technique, involving looking for eclipsing binary stars in the target galaxy, has been used to measure the distance to m33. m31 is next. This technique is still based on the brightness of stars, but the intrinsic brightness is supposed to be better known.
Also of interest is that proper motion has been measured for m33. No proper motion has yet been detected for m31.
I was out birdwatching the other day. I thought i saw a thrush, but it was just a Robin.
I have some. Do you want 360KB or 1.2 MB? Do you want dual 3.5" & 5.12" drives? I gots 'em all.
Coffee, with caffeine, can have anti-oxidant effects, can keep you awake and alert, and give you a little rush
but
caffeine can screw up your sleep cycle, is addictive - complete with withdrawl symptoms, can give you arthritis and osteoporosis, stays in your body for a long time, has cumulatative effects.
---
In the US, food containing caffeine has to say it has it, but does not have to say how much. This makes it more difficult to self medicate.
A colleague complained that she needed to stay awake, but didn't want to take too much, for fear of not getting to sleep at night. I computed it down to 2 oz of Coke. I suggested she by a 12 oz, drink 2 oz, and toss the rest. People don't take my advice very often.
Wouldn't take that much aspartame to kill me. I'm alergic to it.
It's a royal pain. Aspartame is getting into everything. I have
to check all the time.
My diet drink of choice: water.
You are looking for a slight dimming. The atmosphere of the Earth can provide slight dimming. Two cameras looking through different bits to of the Earth's atmosphere can provide extra confidence that the dimming you detect is not here at home.
This is true by default, but is not a requirement. I just got a new 160 GB drive. See http://predelusional.blogspot.com/2006/03/using-pr icewatch-effectively.html
By using a command like /dev/hdb1
mke2fs -j -b 1024 -m 0 -N 2000000
one can have a 1K sized blocks. I argue that with mostly continous filesystems, the block size is largely irrelevant to performance. A small block size reduces the wasted space at the ends of files. For me, going from 4K to 1K gave me 4 GB more free space on my 160 GB drive for my current files - 2.5%.
System supported file compression would yield more, of course. Sure, the mp3 & jpg files won't compress, but I have lots of text files. A good compression system will know this and not bother compressing files that are incompressible. It might even use a file like magic cookie mechanism so it doesn't have to attempt compression to find out it's pointless.
If the improvement in space due to consolodated ECC codes and other overhead saves more total space than the end-of-file wastage, it would still be a win. The article doesn't say how much better the new standard will be.
It may be that Linux will provide 1K filesystem blocks on top of the 4K phsyical blocks. Performance will be worse. However, the original ext filesystem provided half K blocks, and that option is now moribund.
For new systems, this standard is fine. However, I run my machines into the ground, and the better ones have lasted fifteen years. Given that my current machine could last another ten years, it would be a shame to have to toss it into the landfill in five years because the disk drives can't be replaced. Progress is good. Forced upgrades are not.
I run multiple tasks at a time frequently, under Linux. My currently machine is three years old, and wasn't high end at the time. It's plenty fast. The bottleneck is disk. While a disk to disk copy is in progress, such as during backup, the user interface grinds to a crawl. The IDE ATA disks suck down CPU, true enough, but also, the controllers are so tied up doing the copy that other tasks have a hard time getting through. SCSI performs better, but currently, SCSI drives are priced out of the market. Multiple CPUs, same core or not, don't solve this problem. How are USB or Firewire drives on cost and performance?
r icewatch-effectively.html
Generally, i'm happy with current performance. The system can display full screen real time video. I have no requirements past this. My system is also upgradable.
http://predelusional.blogspot.com/2006/03/using-p
I tend to mean more than one thing at a time. At one time, The Beatles (Apple Records) sued Apple Computer. In the settlement, Apple Computer agreed not to get into the music business.
I'd like Apple to get a clue. They're a nice company to deal with. I want a single device in my home entertainment center. It does audio and video time shifting, displays content, allows editing, burns media, everything. And, any compentent system will not produce any audible sound. I do not want a media hub, a TiVo, a VHS VCR, a DVD player, etc. I'd be willing to compromise and say that the display, speakers and amp can be external components. Though, i'll want the remote to include volume control.
Perhaps in their zeal to provide content, they neglected to provide functionality. Many assumed that the new Mac mini would, among other things, provide Tivo functionality. It doesn't. Yet, it connects to your TV, beams content around your house, and otherwise has everything it needs to be a Tivo, except software.
It is starting to look as if the Beatles were right after all. Apple should not be in the content business. Perhaps it is time for someone with a clue to get into the appliance business.
H bombs predate any tokomaks. Though brief, they produce more energy than they consume.
My mother-in-law had one of her heart valves repaired in open heart surgery a few years ago. I was amazed. They said that if they couldn't sew it back together, that they would put in a pig's valve, with somewhat higher risk factors.
Makes you want to live near a hospital. See? City living is healthier after all!
My last laptop was a Compaq Aero - a 3.5 lb 486/25 with 16 MB RAM, 170 MB disk, 640x480 with 64 greys. The only way to install on it was by floppy, so a stack of disks for an old Slackware distro was cut. After a few months, extraneous stuff was deleted, and the result was a system with Postgres, Apache, perl, gcc, emacs running under X (fvwm2) with 70 mb of disk free for my development (so, 100 MB for the system included a 16 MB swap file). I ran PLIP for networking (ssh'ed xterms or NFS) to a desktop, and used CVS to coordinate my development projects between systems.
I still have my resulting system image... If I could spec out a replacement, it'd have flash disk, use AA batteries running at least 30 hrs per charge, and would be fanless.
My current oldest systems are 486/33's with 16 MB ram, 200 MB disk. However, even 100 of them in a cluster does not match the computes of a low end Athlon. They can be OK when dedicated to some task, like a fax server. 486's do not have CPU fans.
It takes less energy to leave the flourescent light on than to turn it off, then back on later.
I imaging this has to do with turing off the gym lights for five minutes between classes. The claim is that starting them back up takes alot of juice.
My application is household 40 watt tube flourescent lights, and incandescent replacement screw-in lightbulbs. Same idea - turn it off for five minutes vs. leaving it on.
In the '60s, probably centered on the CDC 6600, an idea was promoted that a balanced' computer would have capabilities in a ratio. The 3 M's was one with 1 MIPS, 1 mbps, and 1 megabit of memory. So, it could executed 1,000,000 instructions per second, communicate to disk at 1,000,000 bits per second (100,000 bytes per second) and had 1,000,000 bits of RAM (one system had 130,000 bytes of RAM, for example).
The box i'm using to edit this note executes on the order of 1 GIPS, with 100 mbps, and 10 gigabits of memory. That is 1,000,000 instructions per second, 100,000,000 bits per second (10,000,000 bytes per second) to disk, and has on the order of 10,000,000,000 bits of RAM (1 GB). (These numbers are rounded, and, no, i'm not terribly interested in my-box-is-faster-than-yours pissing matches - its just an example).
So, if communications speeds will be 1,000,000,000,000 bits per second anywhere by 2010, that implies a computer with at lest 10 GIPS and 10 GB RAM - which doesn't seem that unlikely in five years.
Oddly enough, I'm hoping to still be running this box in five years. Its only two years old, and I don't really want to get a new one. That is, i don't want to spend the money to replace it. More importantly, i don't want to do the administration involved to get a new machine up and running with my current set of capabilities. I ran my 1987 Machintosh II as my primary machine for over ten years and the hardware lasted an additional five years (and counting) to allow for transfer of data. It pisses me off that my most long-lived x86 based PC has lasted only five years. So, i've just finished migrating from the Mac to Linux, and the Mac (with OS/x) now appears to be the better choice (low administrative maintenance) again.
With the recent announcement of low power PPC chips, perhaps Apple will abandon its move to the x86 hardware platform. Still, i've been pretty happy so far with my low-end Athlon's performance and reliability. Who knows? Perhaps i'd be happy with OS/x on AMD.
Much of the energy used is for air conditioning. One might think that this would be easy for data centers in Michigan, but I've worked for places that heat the building and air condition the data center in January. One place had the data center a/c die, and a box fan in a window allowed everything to run. A box fan has to be cheaper to run than a/c. So what we might see is smarter environmental control. At least in the winter, it makes sense to run outside air in, and use the waste heat to heat the rest of the building.
Perhaps data centers could be moved to Canada.
Mac OS 6.4 is the OS the let me cut the Windows cord. Oh, that's right, in 1987, PCs came with DOS. I never did dual boot my Mac, though I considered running 68k BSD.
At first, my PC, a 386/33 running Linux kernel 0.97 - Slackware, was used for browsing the internet. When Open Office became available, I slowly moved everything I cared about from the Mac. I hardly ever boot the Mac anymore. There still is the project to convert all my old Word docs to RTF or some other open format.
If Dvorak would say interesting things, I'd be tempted to read them.
It was a partial lunar eclipse, not the full monty. The max size of the shadow of the Earth on the Moon was predicted to happen around 8 AM EST, and really just a small bite at that. Moon set was about 7:45 AM EST, local time. Oddly, though one might expect the orange hue due to being on the horizon, and due to light filtering through the Earth's atmosphere onto the Moon like a million sunsets at once. These effects were quite apparent at the last full lunar eclipse visible in these parts. However, none of that was apparent this time. Perhaps the rising Sun washed it out.
Hacking is programming, which is computer science, which isn't science at all, but rather engineering. So the question is:
Is engineering science?
As an engineer who hacks, no. However, I remain scientifically literate.
Since the scientifically literate form a small minority, perhaps I could get special deals as an oppressed minority. The handicapped got parking spaces, and what did they do to get that?
Jeez.
The answer is clearly 42.
The testing process was to run a test, and when it failed, try to figure out if the problem was in the test suite or the tested code. Simple enough.
The tests certainly at some point worked.
No. That wasn't the case. I found myself fixing obvious bugs in the test suite, then attempting to use the fixed version against the target. It was often clear that the test suite could never have worked.
Some distributions still somehow manage to pass the test suits of a new version of the spec. And all this without the people reporting any problems and requesting waiving the test.
We'd report the bugs, with suggested fixes, but we could not wait for fixes to come back and retest. We had to plow forward. We claimed compliance when we had a test we thought tested the assertions and passed it. We never asked for a waiver. Another nice things we came across during the LSBv3 testing are numerous timing problems.
Been there. Done that, though I didn't have to find some slow machine. What is the value of such a certification? What assurance does this give you? Is don't use fast SMP machines an acceptable answer in any universe, especially when it comes to thread tests?
If you have need of slow machines, I can provide approximately 25 working 486/33's. I'd put this on his blog, but he doesn't allow comments. I thought this was strange, because I use livejournal primarily as a place where people can comment. However, he talks about his choice there, too. To each their own.
It is not possible to achieve the goal of 100% binary compatibility...
All good points. And its worse than that. Yet, the exercise was valuable. For us, it uncovered many bugs in SVr3. Many. This was ultimately a good thing for our customers.
We were also a Unix porting house. We fixed lots of bugs in our prior ports of Unix. We offered our fixes to AT&T for free. They declined. We had to apply our fixes to each port - without the benefit of CVS. And, we had thousands of patches. And all this for a basically stable system. It was around then that I was convinced of the incredible inefficiency of propietary software. This would never happen to gcc.
My advise: but the losses.
I read this as "My advice, cut the losses." Oddly, many versions of this mispelling pass my spell checker. Ulrich needs an editor. Perhaps I'll volunteer. Perhaps he can check my work. Will you be a swap editor for me? I'll check your work, you check mine.
So, i agree that the test suite was a horrible idea from the idea that one might assure customers that their old software will still run, or will run on compatible platforms. I agree that the last bug will not be found. However, that is not an excuse to give up the search.