| Date: | November 17, 2005 / year-entry #352 |
| Tags: | history |
| Orig Link: | https://blogs.msdn.microsoft.com/oldnewthing/20051117-11/?p=33303 |
| Comments: | 27 |
| Summary: | Windows 95 contained some new networking features, and since I was one of those crazy people who sampled every flavor of dogfood in the store, I actually tried out all of them. One of the features, a network protocol, I thought was interesting enough that I decided to help them out by forcing everybody else on... |
Windows 95 contained some new networking features, and since I was one of those crazy people who sampled every flavor of dogfood in the store, I actually tried out all of them. One of the features, a network protocol, I thought was interesting enough that I decided to help them out by forcing everybody else on the team to dogfood it. Here's how I did it. I had a bunch of debugging documents and other materials that people generally found useful. I put them on my machine, which acted as a file server. Anybody who wanted these files had to install the network protocol in order to get them. And the files were valuable enough that people were willing to take a chance on a new network driver just to get them. Not only did the client side of the driver get a lot of dogfood test coverage, so too did the server side, since my computer would be servicing a lot of simultaneous connections from people reading my documents. I remember finding a variety of interesting bugs this way. (And of course I ran stress over this protocol.) I later became friends with the lead tester for the protocol, and he told me that my simple act of force-feeding the dogfood to every other member of the Windows 95 team was a key element of making their feature a success. |
Comments (27)
Comments are closed. |
Which protocol? TCP/IP?
My guess would have been IPX, but then IPX was reasonably known in WFW, so maybe it was AppleTalk?
Was it NetBEUI?
Banyan Vines?
Was it UDP?
Ahhhh!!!! I have to know!
I got it. You didn’t backup your work on external SCSI drives because you backed it up on other machines over a network. Yes that kind of dogfooding is respectable.
Still, please note that that kind of dogfooding served corporate customers not home customers. Perhaps Windows 95 should have been marketed only to corporate customers and shouldn’t have been sold with PCs to home customers. If no one in your store had a SCSI drive to dogfood with for their backups, then your store shouldn’t have published documents instructing home customers to use defective tools, and device vendors should have been informed to continue delivering their own tools. Yes networking was more reliable.
Norman, I don’t know where you get these wild topic changes from. For the record, my primary drive on my Windows 95 machine was an external SCSI drive, with backups to a tape drive. What this has to do with forcing people to dogfood network drivers I’m not sure.
Raymond, you should know by now, no technical post on your blog is complete without Norman being able to somehow relate it to data he lost in Win95 due to some random hardware he had attached to an external SCSI card, or (more recently) the fact that his Vista checked build install has been chugging along for 54+ days now :)
Norman,
You constantly post comments moaning about some issue you had with FDISK on Windows 95. Even if Raymond’s post has nothing to do with Windows 95 or SCSI, you still bring this up.
You even admitted that Raymond had little to do with this area of Windows 95, but you still insist on telling him about your issue.
What are you trying to prove?
What does repeating it over and over again accomplish?
For that matter, just how did you conclude that it was a bug in Windows 95’s FDISK?
I’m looking forward to the outcome of Norman’s vista install – Norman, please let us know. In decreasing order of likelihood, my guesses are:
He gets bored and kills it.
The power fails.
The hardware fails.
It actually finishes (ha!)
Friday, November 18, 2005 12:50 AM by oldnewthing
> Norman, I don’t know where you get these
> wild topic changes from.
I agree that your dogfooding of network connectivity had results it was supposed to have. If I had known at the time to do backups over a network instead of to an external SCSI drive, and if a second PC didn’t cost 4 times as much as an external SCSI drive, I might not have discovered half of my reasons for bitching either.
> For the record, my primary drive on my
> Windows 95 machine was an external SCSI
> drive,
Interesting. Partitions created by using Windows 95 FDISK on a blank disk, or partitions created by some other tool before you installed Windows 95? Geometry translation computed by a BIOS or computed by Windows 95?
Friday, November 18, 2005 8:46 AM by msemack
> You constantly post comments moaning about
> some issue you had with FDISK on Windows 95.
That’s part of it. Associated with it are Microsoft’s denial of warranty service and subsequent assertion that it’s a non-issue.
I suppose you don’t mind if every file in two of your external hard disk partitions gets lost because Windows created overlapping partitions. It took months to track down, with numerous train trips (also not free) back to vendors and the store where I bought the stuff. After knowing where the problem is, it takes 10 minutes to reproduce it. But Microsoft couldn’t ever afford those 10 minutes. Microsoft really shows their attitude towards customers. I didn’t echo this attitude to Microsoft until they demonstrated it.
Friday, November 18, 2005 10:22 AM by Carlos
> I’m looking forward to the outcome of
> Norman’s vista install
My guesses in order of likelihood are:
(1) Hardware failure
(2) Power failure (the battery is completely dead, it was bought as a crash box).
(3) It actually finishes but I won’t be able to log in to activate it.
One day it wrote enough disk files to reduce the free space by 13 megabytes instead of the usual roughly 1 megabyte per day. Also the long-running mscorsvw.exe process has grown enough that paging is starting to become an issue, and when it spawns a second temporary mscorsvw.exe process, CPU utilization drops to around 10%. But still most of the time it’s running around 90%.
Sunday, November 20, 2005 10:00 PM by Norman Diamond
"I suppose you don’t mind if every file in two of your external hard disk partitions gets lost because Windows created overlapping partitions."
You have yet to demonstrate that there really was an "overlapping partitions" bug in Win95’s FDISK. How exactly did you come to this conclusion?
"But Microsoft couldn’t ever afford those 10 minutes. Microsoft really shows their attitude towards customers. I didn’t echo this attitude to Microsoft until they demonstrated it. "
Did you ever contact Microsoft about your issue? Did a Microsoft employee ever tell you that they didn’t care about their customers?
I don’t want to be adversarial here, but I still don’t understand what you’re trying to accomplish by repeating your story over and over again.
Monday, November 21, 2005 4:54 PM by msemack
> You have yet to demonstrate that there
> really was an "overlapping partitions" bug
> in Win95’s FDISK.
I’ve demonstrated it live to a number of people. At least one card vendor updated their documentation and added updated partitioning tools to the software accompanying their product.
One time Microsoft accidentally allowed correspondence between a product support manager and myself. I wrote details of how to reproduce it (which took longer to write than to do the actual reproducing). He asserted that very few people were affected by it, Windows 95 was old, and Microsoft would not act unless I paid for support. The assertion about the number of victims is refuted by observing how quickly external SCSI drives sold in the marketplace (except apparently in the US). Sure Windows 95 was old but it was still on sale and it why wasn’t it old enough for Microsoft to have discovered the bug yet? Oops it was old enough for Microsoft to discover it because Windows 98 had it fixed, so why didn’t Microsoft let Windows 95 victims get the fix?
> How exactly did you come to this conclusion?
If you write a few dozen files to the last partition then the second-to-last partition becomes completely corrupted, Windows Explorer displays garbage, Scandisk can’t recover any files, etc. If you write a few dozen files to the second-to-last partition then the last partition becomes corrupted.
Duplicated in Windows 95 original in English, Windows 95 original in Japanese, Windows 95 OSR2 in English, Windows 95 OSR2 in Japanese, Windows 95 original plus two downloadable IDE driver patches (which indeed fixed IDE drivers but not FDISK for SCSI) in both languages. Duplicated on several vendors’ laptops, several vendors’ PCMCIA SCSI adapters, several vendors’ external SCSI drives.
Workarounds duplicated by using several vendors’ partitioning tools that were intended for Windows 3.1, after which Windows 95 used the existing partitions and did not add corruption.
> I still don’t understand what you’re trying
> to accomplish by repeating your story over
> and over again.
Sure it won’t accomplish much, but it gets my goat when I see assertions about how much testing or how much dogfooding was done. More than that, what really gets my goat is the repudiation of warranties. Microsoft still asserts that OEMs are responsible for support but for some reason OEMs still aren’t allowed to deliver to customers hotfixes that Microsoft has developed. Sure Mr. Chen isn’t in charge of warranties, and I’ve actually bit my tongue off a few times instead of repeating the facts when warranties were discussed before in Mr. Chen’s blog. But sometimes my tongue gets tired of being bitten off.
"I’ve demonstrated it live to a number of people."
But how do you know that the problem WAS overlapping partitions? Did you take a dump of the drive’s partition table an analyze it?
For that matter, this doesn’t prove that it was an FDISk problem. It could very well have been a bug in the SCSI BIOS.
"He asserted that very few people were affected by it, Windows 95 was old, and Microsoft would not act unless I paid for support."
Considering that I find no one on Usenet decribing this issue (besides you), I find it unlikely that anyone else was affected by it.
http://groups.google.com/groups?q=windows+95+SCSI+%22overlapping+partitions%22&start=0&hl=en&lr=&
If you were the only person affected by this issue, it is easy to see why it was missed in testing.
I will also point out that I used Windows 95 A and B on machines with SCSI drives (internal and external), and I never saw the problem you describe.
"Sure it won’t accomplish much"
It doesn’t accomplish ANYTHING. If you had issues with ONE product support manager, then maybe you should take it up with him. Making pithy comments in Raymond’s blog (especially on topics that are totally unrelated) won’t solve your issue.
Tuesday, November 22, 2005 12:38 PM by msemack
> But how do you know that the problem WAS
> overlapping partitions?
Because writing to one partition caused destruction of the contents of the other.
> Did you take a dump of the drive’s partition
> table an analyze it?
I must confess I did not. At the time PowerQuest[*1] offered a free downloadable tool to analyse partitions, and the best that it could come up with was utter garbage. Even when I took my PC[*2] and adapter card and hard drive to vendors and we experimented with those and with their PCs and their adapter cards and hard drives, they also didn’t dump the partition chain. It was clear enough to them too what was happening.
But you are right, I ought to try to get a dump of a partition chain. The procedure would be to attach the drive to a Linux machine and use the dd command to read each logical block number, figuring out in turn which are the interesting block numbers to read next. Maybe I’ll have a day to spend on that during the new years’ holidays.
> For that matter, this doesn’t prove that it
> was an FDISk problem. It could very well
> have been a bug in the SCSI BIOS.
That makes no sense. It occurs on machines that have no SCSI BIOS. Do you even know any PCMCIA SCSI adapters that have SCSI BIOSes?
And conversely the problem can be avoided on desktop machines if there is a SCSI BIOS, because then FDISK obeys the SCSI BIOS’s geometry translation instead of doing its own.
> Considering that I find no one on Usenet
> decribing this issue (besides you),
Yeah I noticed.
> I find it unlikely that anyone else was
> affected by it.
Your finding is utter garbage. I seem to be the only one who spent months tracking down the cause and discussing it with vendors, but there were millions[*3] who didn’t know that they had time bombs sitting in corrupt partition chains, and when they lost the contents of their hard drives they didn’t know what the reason was.
Plus, if you have a notebook PC with Windows 95 on it, and a PCMCIA SCSI adapter and external SCSI drive, it will take all of 10 minutes for you to find yourself affected by it too. Of course now it might be a bit difficult to find a hard drive smaller than 30 gigabytes (Windows 95’s limit) so you might need to get a used one, and be sure to zero out at least the MBR before starting.
> If you had issues with ONE product support
> manager
Oh he was just the icing on the cake. Microsoft’s entire operation reneges on warranties, both before and after that. Microsoft’s assertion that OEMs provide support isn’t just the policy of one product support manager, and the inability of OEMs to provide Microsoft hotfixes to customers isn’t just the policy of one product support manager.
[*1 Maker of Partition Magic, which I did not own or use, but I did use the analyser while trying to figure out what was happening]
[*2 At the time I only owned one, but duplicated the problems using other PCs that I didn’t own]
[*3 PCMCIA SCSI adapters and external SCSI drives were huge sellers, even if they weren’t in the USA]
There is another way to get a hint, which I forgot yesterday while writing. This one takes maybe 15 minutes to reproduce instead of 10.
After zeroing the MBR and using Windows 95 FDISK to create the partitions (and optionally copy some files to them, it doesn’t matter), get permission to disconnect the adapter and hard drive. Then connect them to a Windows 2000 machine. In Windows 2000, open up Disk Manager. Windows 2000 shows assertions of the status of each partition on the drive, healthy or damaged. It asserts that only one partition is damaged. But when you tell it to delete the one damaged partition, it really deletes both of the overlapping partitions.
"I must confess I did not."
Then you have no idea what the real culprit was. It sounds like you saw some data corruption and decided to blame FDISK without fully researching the problem.
"That makes no sense. It occurs on machines that have no SCSI BIOS. Do you even know any PCMCIA SCSI adapters that have SCSI BIOSes?"
Up to this point, I was anaware that this happened on PCMCIA SCSI controllers. However, that is besides the point. Crapping on Mr. Chen’s blog
"Your finding is utter garbage. I seem to be the only one who spent months tracking down the cause and discussing it with vendors, but there were millions[*3] who didn’t know that they had time bombs sitting in corrupt partition chains, and when they lost the contents of their hard drives they didn’t know what the reason was. "
You claim there were "millions". Do you have ANY evidence to support that claim? Can you find a reference of how many people bought PCMCIA SCSI controllers and used them under Windows 95?
Can you find me a web page or Usenet posting from someone else that was affected by the same problem? If you can’t, it’s looking like the problem affected FAR less people than you suggest. I suspect something was very unique in your configuration.
"Oh he was just the icing on the cake. Microsoft’s entire operation reneges on warranties, both before and after that."
Welcome the world of software development. Nobody offers warranties on software, execpt in the extreme case of mission-critical embedded systems.
"Microsoft’s assertion that OEMs provide support isn’t just the policy of one product support manager."
Then don’t buy an OEM version of the software. There’s a reason it’s cheaper, you know.
"OEMs to provide Microsoft hotfixes to customers isn’t just the policy of one product support manager."
I think this is a perfectly reasonable policy. If the patch hasn’t been fully regression-tested, you don’t want it distributed willy-nilly.
Norman,
I’m not trying to be adversarial here. However, I don’t get why you think posting pithy comments in Mr. Chen’s blog is a worthwhile execrcise.
Sunday, November 27, 2005 1:58 PM by msemack
> Then you have no idea what the real culprit
> was.
Wrong, because it’s 100% reproducible, and because it’s 100% solved by using any of the following:
Windows 98 FDISK (how in the world did that get fixed before release, eh?)
Windows NT4/2000/XP/2003 disk manager
Partitioning tools that vendors made for use under Windows 3.1
> Up to this point, I was anaware that this
> happened on PCMCIA SCSI controllers.
Then up to this point, you didn’t even read as much of the "crapping" as you thought you read.
> You claim there were "millions". Do you have
> ANY evidence to support that claim?
Yes, because PCMCIA SCSI controllers and external SCSI hard disks were flying off store shelves and there were dozens of vendors. Wouldn’t have happened with a marketplace of less than millions. That market is almost gone now because USB2 is fast enough and cheaper.
> Nobody offers warranties on software, execpt
> in the extreme case of mission-critical
> embedded systems.
So you really haven’t done much reading. There’s at least one software company that PRETENDS to offer warranties even when they don’t.
> Then don’t buy an OEM version of the
> software.
Bingo. Again you haven’t done as much reading as you thought. I’ve posted several times about my wish not to buy an OEM version of the software when the software was Windows 95, I wanted to buy NT4, but the only way to buy a computer and buy NT4 was to buy an unwanted OEM version of 95 along with it.
I will appreciate very much your assistance in buying computers without buying OEM versions of the software.
It’s still impossible to buy a laptop other than a crippled one without also buying an OEM version of XP. In most countries it’s even still impossible to buy a brand name desktop without also buying an OEM version of XP. Of course XP isn’t as bad as 95 most of the time, but still, I will appreciate very much your assistance in buying computers without buying OEM versions of the software. Thank you very much for providing that instruction, I just need a bit of help in carrying it out.
> I don’t get why you think posting pithy
> comments in Mr. Chen’s blog is a worthwhile
> execrcise.
Agreed, it just gets my goat when I see assertions of how much testing was done, or that support exists, etc.
"Wrong, because it’s 100% reproducible, and because it’s 100% solved by using any of the following:
Windows 98 FDISK (how in the world did that get fixed before release, eh?)
Windows NT4/2000/XP/2003 disk manager
Partitioning tools that vendors made for use under Windows 3.1 "
But you still haven’t demonstrated that the problem is OVERLAPPING PARTITIONS. This was my point in the first place (which you seem to have missed).
To determine that partitions are overlapping, you must look at the partiton table for the drive and see where the various partitions start and end. To fully understand the problem, you need to do this. Otherwise, you are just speculating.
"Yes, because PCMCIA SCSI controllers and external SCSI hard disks were flying off store shelves and there were dozens of vendors."
How many people WERE AFFECTED BY THE ISSUE? Not how many people bought PCMCIA SCSI controllers. There is a difference.
"I’ve posted several times about my wish not to buy an OEM version of the software when the software was Windows 95, I wanted to buy NT4, but the only way to buy a computer and buy NT4 was to buy an unwanted OEM version of 95 along with it."
Blame the PC vendor for not offering an NT4.0 OEM version. I bought PC’s with NT4.0 pre-loaded (no Windows 95).
"There’s at least one software company that PRETENDS to offer warranties even when they don’t."
Who offers warranties on software? Name one software vendor that guarantees their software will be free of defects.
BTW, you keep claiming that I "haven’t done as much reading as I thought". You don’t know how much reading I have done. You don’t know how much reading I THINK I have done. Please do not make claims about me which you can not verify.
"Agreed, it just gets my goat when I see assertions of how much testing was done, or that support exists, etc. "
And just WHAT gets your goat about this?
Obviously, a lot of testing was done on Windows 95. Raymond’s stories are evidence of this.
Did bugs make it into the release of Windows 95? Absolutely. I don’t think anyone will claim otherwise.
However, even if your issues WAS caused by Windows 95’s FDISK (which you STILL haven’t convinced me of that), I don’t see how this undermines the accuracy of Raymond’s stories.
Furthermore, I don’t see how your issues undermines the fact that extensive testing was done.
While your issue was clearly very upsetting to you, it obviously didn’t hurt a lot of people. If it had, I would have found far more Usenet postings about the issue.
Monday, November 28, 2005 9:44 AM by msemack
> But you still haven’t demonstrated that the
> problem is OVERLAPPING PARTITIONS.
Fine. I’ll repair this serious defect in my report of the facts.
If you attach a new SCSI drive (or newly initialized by zeroing the MBR) to a PCMCIA SCSI adapter, attach them to a computer running Windows 95, and run Windows 95 FDISK to create a DOS extended partition containing two or more DOS logical drives, then the last two logical drives in the extended partition chain might or might not be overlapping, but the last two logical drives in the extended partition chain are damaged in some serious manner. Since we can’t call it overlapping, and I’ll bite my tongue off instead of calling it by name, let’s say fickled. It is not immediately apparent that the last two logical drives are fickled, because some amount of writing can be done to them before the damage becomes apparent.
After running FDISK, ordinarily you have to reboot. Since this drive is connected by a PCMCIA card, it is possible to get Windows’s permission to disconnect and reconnect instead of rebooting, but it doesn’t matter. To feel safe, reboot.
Then format. Each fickled partition will appear to format fine.
Then copy a bunch of files to one of the fickled partitions. It will work. You can scandisk that partition and it is fine.
Then copy a bunch of files to the other of the fickled partitions. It will work. You can scandisk that partition and it is fine. But then look back at the partition that you previously copied files to. Sometimes it will still be OK, sometimes it will be ruined. That depends on how much writing you’ve done to the partition that you later copied files to.
Now that you see the fickling effects in one partition, reformat it. Copy a bunch of files to it again. Then look at the other partition again, and you’ll see the fickling effects there.
So, what does it mean when writing to one partition ruins the entire contents of an adjacent partition? Since we can’t call it overlapping, I called it fickling in this posting. Vendors who saw it understood it as overlapping, but what do they know.
Even if it’s not overlapping, the fickling is immensely destructive. How did it get fixed in Windows 98 when programmers didn’t even know about it. Why weren’t Windows 95 victims allowed to get the fix.
> How many people WERE AFFECTED BY THE ISSUE?
> Not how many people bought PCMCIA SCSI
> controllers. There is a difference.
The number of people who were affected by it was the number of people who used Windows 95 FDISK to partition hard drives via PCMCIA SCSI controllers (plus I think any SCSI controller that didn’t have geometry translation built into it). Yes that is different from the number of people who bought PCMCIA SCSI controllers, because some people bought PCMCIA SCSI controllers prior to Windows 95, and vendors provided reliable partitioning tools for Windows 3.1. Some used PCMCIA SCSI controllers with NT4 (which requires the nuisance of shutting down before connecting or disconnecting, but which worked safely). And some bought and used PCMCIA SCSI controllers with Windows 95 after some vendors provided partitioning tools for Windows 95. You are right, those numbers aren’t all equal. They’re still in the millions.
> And just WHAT gets your goat about this?
Because I see the amount of testing that gets to be done by customers instead of being done by Microsoft, AND Microsoft refuses warranty service, etc.
> it obviously didn’t hurt a lot of people.
Wrong. The fickling damage usually becomes apparent months after FDISK was used. I was comparatively lucky in getting hit with it in less than the pretended 90-day warranty period, but there was still no way to turn the pretended warranty into a real one. Even after getting hit, it took a few more weeks of experiments before proving that it always happened when Windows 95 FDISK was used (100% reproducible), did not happen when vendors’ partitioning tools for Windows 3.1 were used even though Windows 95 was used after that for formatting each partition and writing files, did not happen when Windows NT4 disk manager was used for partitioning even though Windows 95 was used after that for formatting each partition and writing files. Most victims were hit months later and had neither the skills nor the recollection of events to go back and reproduce it. Most victims probably just assumed they did something wrong, or that it’s just the way computers work (just like BSODs being just the way computers work).
If this problem really affected as many people as you suggest, we would find more than one person reporting the issue.
If it affected "millions" of people, then we would see more than one person (you), ranting about it on Usenet and Mr. Chen’s blog. Even if only 1% of the "affected" customers reported the problem, that’s over 10,000 people.
And yet, you’re the only person reporting this issue. All I can conclude is that the issue affected far less people that you think it did.
I still suspect that something was unique to your systems which triggered the issue.
"Because I see the amount of testing that gets to be done by customers instead of being done by Microsoft, AND Microsoft refuses warranty service, etc."
Microsoft does very extensive testing. I don’t think anyone will deny that (except maybe you). They probbly have the largest software testing budget of any company in the world.
It is unfortunate that your SCSI issue apparently made it through test. Stuff like this happens with ALL software (not just Microsoft). Furthermore, it will continure to happen for MANY years to come.
We haven’t come up with a way to write (non-trivial) software that is devoid of bugs.
We haven’t come up with a way to test and verify that a program is truly bug-free.
Let me know when you come up with a solution to the problem.
Tuesday, November 29, 2005 1:19 PM by msemack
> I still suspect that something was unique to
> your systems which triggered the issue.
Uniquely reproduced 100% on several different laptops, several different vendors’ PCMCIA SCSI adapters, and several different external hard drives. Uniquely reproducible 100% in minutes when we know how to do it, using 4 different versions of Windows 95. Uninquely not occuring with identical hardware in Windows 98, reproducible 100%.
Talk about crap.
Your previous statements too:
> Then don’t buy an OEM version of the
> software.
> Blame the PC vendor for not offering an
> NT4.0 OEM version.
Again talk about crap.
It is public knowledge what happened when IBM offered choices to customers.
By the way even among the few vendors that offered NT4 as a choice, why did they suddenly stop offering NT4 when NT4 SP4 came out? Why didn’t they resume when NT4 SP5 came out?
A few vendors were still allowed to offer it on crippled hardware. Just like Microsoft allowed IBM to offer choices, by charging higher prices and minimizing cooperation.
But no vendor was allowed to obey the supposed EULA which asserts a 90-day warranty for Microsoft’s software. OEMs said that even in cases where Microsoft had developed fixes, OEMs were not allowed to provide the fixes to their existing customers.
> We haven’t come up with a way to write
> (non-trivial) software that is devoid of
> bugs.
Wow, a sentence with no crap, amazing.
> Let me know when you come up with a solution
> to the problem.
Some of us have come up with a way to perform reproductions of bugs when 100% reproductions have been presented. Some of us do know how to honour warranties. Some of us do know how to give refunds when we can’t fix a bug. Some of us do know how to deliver fixes when we obviously have done a fix.
And guess what. Although I was angry on discovering that this bug was in Windows instead of something unique to my situation, I didn’t hate Microsoft yet. When Microsoft demonstrated how fake its asserted warranties were, would not allow submission of bug reports, developed hotfixes for numerous less serious bugs but refused to let existing customers download the fixes, I gradually learned to echo the spite which your company shows to its customers. But you’re right, I haven’t come up with a solution that would work for Microsoft. My solution only works for companies that are less arrogant than yours.
If you insist that my comments are "crap", then please explain why you are the only person reporting this issue. Until then, I fail to see how my comments are crap.
Regarding your comments on software warranties: You have yet to name a SINGLE software company that offers a warranty on their software. As soon as you find one, let me know.
Until then, your argument makes no sense. The lack of software warranties is an industry-wide issue.
Wednesday, November 30, 2005 9:20 AM by msemack
> If you insist that my comments are "crap",
At least one exception was noted.
> please explain why you are the only person
> reporting this issue.
I did.
http://blogs.msdn.com/oldnewthing/archive/2005/11/17/493891.aspx#497729
>> I was comparatively lucky in getting hit
>> with it in less than the pretended 90-day
>> warranty period, but there was still no way
>> to turn the pretended warranty into a real
>> one. Even after getting hit, it took a few
>> more weeks of experiments before proving
>> that it always happened when Windows 95
>> FDISK was used (100% reproducible), did not
>> happen when vendors’ partitioning tools for
>> Windows 3.1 were used even though Windows
>> 95 was used after that for formatting each
>> partition and writing files, did not happen
>> when Windows NT4 disk manager was used for
>> partitioning even though Windows 95 was
>> used after that for formatting each
>> partition and writing files. MOST VICTIMS
>> WERE HIT MONTHS LATER AND HAD NEITHER THE
>> SKILLS NOR THE RECOLLECTION OF EVENTS TO
>> GO BACK AND REPRODUCE IT. MOST VICTIMS
>> PROBABLY JUST ASSUMED THEY DID SOMETHING
>> WRONG, OR THAT IT’S JUST THE WAY COMPUTERS
>> WORK (JUST LIKE BSODS BEING JUST THE WAY
>> COMPUTERS WORK).
> You have yet to name a SINGLE software
> company that offers a warranty on their
> software.
Right. Several times I’ve alluded to a famous software company that PRETENDS to offer a warranty on their software. But the asserted warranty is a fraud, and the company doesn’t really offer the warranty that they assert. That is one of the reasons that I learned to hold the same opinion of that company as that company holds of its customers.
If you’ve done as much reading as you’ve alluded to, then you’ve seen the pretended warranty.