| Date: | May 21, 2009 / year-entry #145 |
| Tags: | tipssupport |
| Orig Link: | https://blogs.msdn.microsoft.com/oldnewthing/20090521-00/?p=18203 |
| Comments: | 61 |
| Summary: | Quite some time ago, a customer had forgotten that they were using an evaluation edition of Windows, and they were awakened one morning with the following error message: The evaluation period for this installation of Windows has expired. This system will shut down in 1 hour. To restore access to this installation of Windows, please... |
Quite some time ago, a customer had forgotten that they were using an evaluation edition of Windows, and they were awakened one morning with the following error message:
The customer submitted an urgent request for assistance. "Please advise how we can get this machine working again. We need it to run a demo for a major client later this week." In the customer's panic, they forgot to read the actual error message, which tells them what they need to do:
In other words, get a non-evaluation edition of Windows and perform an upgrade install. (Kick it off as soon as you finish booting up so you don't get the rug yanked out from under you if you dawdle and take more than an hour to get through setup.) This will preserve all your existing data while upgrading it from an evaluation edition to the real thing. Pre-emptive snarky comment: "The existence of evaluation editions with expiration dates proves that Microsoft is evil." |
Comments (61)
Comments are closed. |
I can not count the number of times I get a call and the user says "its not working". I ask "did the software give you any kind of error message"? And their reply is usually "Yes, but I forgot what it said".
This happens in email. I was on a thread where person A said, "Our product now does <foo>. This will not impact you because of <bar>." The message was literally two lines.
Someone else on my team immediately Reply-All’ed with "Will this impact us?"
Sheesh!
Sweet, sweet memories.
I recall a relatively common procedure we’d have to walk customers through, years ago when I would sometimes have to walk customers through procedures. It would result in a dialogue box coming up which would say "Please wait while windows builds a driver information database", with a percentage bar that would go from 0% to 100%. No buttons. The bar moved relatively quickly- Well, quickly enough that there was no reason to think the machine had crashed.
The conversation would invariably go as follows.
Customer: "It says please wait while windows builds a driver information database. What should I do?"
Me: "We’ll wait while windows builds a driver information database."
Customer: "Oh, ok."
There’s a corollary to this:
"Please put USEFUL information into the the error message",
If it’s useful then, perhaps, more people would read it, maybe. Write in English, (or National language of choice anyway) not technobabble. If possible report the names of the file/driver/hardware that’s missing and/or broken. Also try and tell the user how to fix it. Oh and finally keep it short :).
We’ve all seen messages like the dialog box that says "ERROR" and nothing else, yeah – what error, where? Or, just to pick on ms, "One of the components required to run your application is missing". Which component? Or the BIOS POST error "Keyboard is broken, press F1 to continue" Duh! Examples are legion.
@steve smith:
"Or the BIOS POST error "Keyboard is broken, press F1 to continue" Duh! Examples are legion."
This is not the exact message. It is more like:
"No keyboard error. Press F1 to continue."
And it makes sense – at least, when it was introduced. Back in the 80s, it was perfectly possible that the keyboard reported an error, but it was actually working. Thus, you could just press F1 although the BIOS did not recognise the keyboard.
Another option would be to attach a keyboard and press F1 -although, hot-plugging the DIN plug or the PS/2 plug is not really recommended, but it works.
> “The existence of evaluation editions with expiration dates proves that Microsoft is evil.”
Anyone submitting that snarky comment would be demonstrating muddy thinking. The existence of evaluation editions with expiration dates actually shows that some beta testers are evil and must be guarded against.
-Wang-Lo.
That comment isn’t very snarky.
HA! "That guy" thinks it is just fine to make the computer stop working after a certain period. But thinks it is totally bonkers to have the computer work only two hours at a time after a certain period.
It isn’t so much that people don’t read error messages (well they don’t) but they generally just want the dialog box to go away.
[Tell that to this guy. -Raymond]
On the one hand, 3 months is an excessive amount of time. On the other hand, people don’t read a damn thing, and this may get through. Or they could call tech support with ‘my computer crashes every two hours’
"Captain, a message box error was mention. I believe there is a 98.9% chance that the <<Keyboard not found, press F1 to continue>> BIOS message from the 80s will be mentioned."
I guess we see it from a "computer person" point of view. If an error light flashes up on my car’s dashboard, I can read the handbook to find out what it is, but I’m still going to pay someone who knows what they’re doing to fix it.
Stretching the analogy a bit, if you get a CEL, before spending a bunch of money on a mechanic, why not replace the air filter, tighten your gascap, and clean your MAF? Depending on the codes (which you can pull for free), you can fix a lot of the problems yourself cheaply for a bit of effort.
Yes, "that guy" is basically complaining that he feels someone at Microsoft botched the terminology; that, instead, the RC "expires" in March, and "dies" in June.
It’s a matter of interpretation that he has decided to make a big deal about.
It does not matter how plainly worded or friendly the language is. My parents will not read it!
@Chris L:
That is why Windows 7 RC will in no uncertain terms convey that the game is over. Maybe Microsoft should couple the shutdown with a video of a fat lady singing.
I’m generally allergic to ANYTHING that says "YOUR MACHINE WILL REBOOT in X hours and you can’t prevent it". fsck, this is MY computer and I won’t allow a piece of silicon command me!
LOL!
I have seen programmers fall into this themselves; I think there is a tendency to believe we deal with such difficult problems that a simple error message cannot possibly include a solution.
A few weeks ago at work, a friend asked me to take a look at his compilation problem. (Compiler messages used to be obscure, but in recent years they have improved dramatically). Right in the compiler message was the solution, which I had never before encountered, but I acted all smart and said "well, did you try…" and read the error message right off the screen! "Um, no." And guess what? It worked! :)
"In the customer’s panic, they forgot to read the actual error message, which tells them what they need to do:"
I suspect that they fully understood what the error message was telling them, but were hoping that there was a registry trick that would extend the trial period so they wouldn’t have to run out, buy a licensed copy of Windows, and install it before the demo.
Is there a way to know when my Windows 7 beta will expire? I know it’s a beta but I had no idea it’s time limited until now! I’m running build 7000, by the way.
Wesha: Then presumably you won’t be running pre-release software which Microsoft released to you (for free!) for testing purposes, right?
I think the reboot every 2 hours thing is perfectly logical and legitimate. People will ignore the "this evaluation copy of Windows will expire in 3 months" message and then complain when all of a sudden it stops booting. At least rebooting on you every 2 hours is not something you can ignore. It also gives you a chance to go out and buy a real copy, then you’ve got plenty of time to copy your important stuff off, install the new version and off you go.
Keep in mind that by the time the reboot-every-2-hours thing has started happening, the final release of Windows 7 will have been out for a couple of months (assuming it’s on time :p) so there’s no reason for you to still be running the pre-release version at all…
Sorry, that’s not very snarky. The guy who is complaining is an idiot.
Please note that I’m a frequent Microsoft basher. But I only like to bash them for the things they actually do wrong. Which is not nearly so large a list as some people make out.
The time-limiting of an RC build is the most sensible possible thing. And making the time out severely annoying is a good thing too – the whole bloody point is to get people to stop using the RC build and buy a freaking copy.
I think the only real thing to complain about here is that the PR says ’till June’, but the OS really becomes unusable in March. Oh well.
Once a guy who worked for me came in one morning and said he couldn’t log into Windows any more – some error message dialog. I asked him what it said, and he said he didn’t know, he didn’t read it. So I told him to read the error message and see if it said anything useful (by this time I’d already guessed the reason, but I viewed it as a valuable instructional aid for said person).
He came back and said it didn’t say anything useful. He was in the next office, so I stood up and went to the door and watched him log in again (from where I could see that my guess was indeed correct). Again he said it didn’t work and got frustrated. Again I told him to read the error message. Same thing again.
Then I told him to read the error message aloud to me, which he did. And then I looked at him and said "So…is your Caps Lock key on then?" (like the error message suggested).
Then he realised it was and turned it off and logged in, in a bad mood. Oh well.
Still, at least now you know why XP and later has that bubble help when you turn caps lock on when trying to log in :-)
http://lmgtfy.com/?q=windows+7+beta+expire
With regards to the "Kick it off as soon as you finish booting…."
Wouldn’t it be sensible and convenient for all, if after the upgrade process was started, Windows would in this circumstance allow it to finish and not re-boot in the middle of it?
[On the one hand, 3 months is an excessive amount of time. On the other hand, people don’t read a damn thing, and this may get through. Or they could call tech support with ‘my computer crashes every two hours’]
Except that Windows 7 should have been available to the public for several months at that stage, and have its own 60 day trial version.
And I doubt that any companies will be installing the RC on users computers. They may be using it in the IT deparment for evaluation purposes, but it would be insane to install it on a non-techie computer.
[I suspect that they fully understood what the error message was telling them, but were hoping that there was a registry trick that would extend the trial period so they wouldn’t have to run out, buy a licensed copy of Windows, and install it before the demo.]
No, chances are that the user simply did not read the message. Too many people have got into the mindset that error messages simply stop them from doing what they want to do, so they need to get rid of them as quickly as possible.
[On the one hand, 3 months is an excessive amount of time. On the other hand, people don’t read a damn thing, and this may get through. Or they could call tech support with ‘my computer crashes every two hours’]
Except that Windows 7 should have been available to the public for several months at that stage, and have its own 60 day trial version.
And I doubt that any companies will be installing the RC on users computers. They may be using it in the IT deparment for evaluation purposes, but it would be insane to install it on a non-techie computer.
[I suspect that they fully understood what the error message was telling them, but were hoping that there was a registry trick that would extend the trial period so they wouldn’t have to run out, buy a licensed copy of Windows, and install it before the demo.]
No, chances are that the user simply did not read the message. Too many people have got into the mindset that error messages simply stop them from doing what they want to do, so they need to get rid of them as quickly as possible.
@Tim Jarvis: If the setup process could circumvent the one-hour restriction then so could other apps.
Hehe, so naive…
Yeah, users don’t read.
my favorite one is from Server 2008 on a functional server
“NTLDR is missing”
why was the BCD cached in memory so it got trashed upon reboot?
bad design flaw homies
rm /vmlinuxon a running linux system. -Raymond]“[Tell that to this guy. -Raymond]”
Totally. I mean, If I get something for free, I can at least expect it to do what *I* want, instead of being grateful for getting it at all, right?
It reminds me that years ago someone phoned Microsoft support for workaround of initial release of WGA and they got the registry hack… Perheps they’re willing to try their luck one more time… :P
If the same expiration scheme is kept for the RTM 60 days trial edition, will my PC start rebooting every 2 hours 30 days before I install it? (Now, that would be mental)
> If the setup process could circumvent the one-hour restriction then so could other apps.
So perheps Microsoft should also offer the installer to perform upgrade on evaluation installations when they boot from the disc. (It seems that in Vista you can only upgrade an installation when you run the setup within the running OS you want to upgrade.)
*that guy* shouldn’t be writing for a magazine called "IT Pro". He obviously knows nothing about software; I imagine he knows less about hardware.
I’m willing to guess that the reason Windows 7 RC starts rebooting every 2 hours in March is by their nature testing machines don’t get used every day. Plus, it gives certain companies time to recall machines running beta and RC versions before they get too far along…
http://blogs.msdn.com/oldnewthing/archive/2006/02/14/531798.aspx
Just a point of correction – most decent Linux systems will actually survive that (as would most embedded Linux systems). /vmlinuz is the kernel, but it’s hardly ever used for booting Linux – those kernels normally reside in /boot, and are compressed (the kernel will decompress itself then boot). /vmlinuz is an uncompressed kernel image not used for anything much, really…
No, Raymond, if you want fun, try some key files. /sbin/init is a good one – pros know how to fix it, but most will sit blankly at the screen with either a shell prompt, or "Kernel panic: cannot find init".
I think ntldr or bootmgr will attempt to boot a system missing the boot.ini/bcd files, by assuming a somewhat standard configuration. A program update a few months ago comes to mind when they accidentally nuked boot.ini and the majority of people didn’t notice, but some were left with an unbootable system…
This hits close to home, having done tech support on and off for almost 10 years previously in life. Even now, as the unofficial "computer guy" in my department at work, I’ll have people come up and say things like "My computer at home has an error. Do you know what’s wrong?". No other information. Just "an error".
"So perheps Microsoft should also offer the installer to perform upgrade on evaluation installations when they boot from the disc."
"[That’s an awful lot of work for a fringe case. -Raymond]"
AIUI, the linked post is mostly saying "please don’t upgrade 7beta->7RC, because we’d rather you tested the Vista->7RC upgrade path"; not "please don’t do it because it’s too hard".
RC->final is surely a technically different task to beta->RC (or beta->beta), as there shouldn’t have been any major changes, just bugfixes (that’s what RC means, after all). I don’t see why it would be a insurmountable problem; I suppose the economics of it come down to how many people are running the RC vs total Vista users, and that’s probably something only Microsoft knows, though. But then there’s still the question of why under-running-system/booted-from-CD makes a difference to how hard the migration is.
Or even better, after supplying the RC with a valid RTM licence key, it updates the system automagically (via web or DVD).
I don’t think that the critic is against Microsoft requiring an update after a certain period of time, he is just against how it is done.
First, given that the evaluation period is a year, the user is likely to expect an issue or a requirement to upgrade at a year’s time. So, the user might allocate the last month to ensure that all systems work with the eval version .. i.e. the 11th month.
However, the system actually causes reboots at 9 months — which might not be planned for. Even if the system is just used for testing, there could be serious disruption to the testing schedule since 2 hour reboots would cause major havoc.
So, should they start posting warnings at 6 months that the system would be seriously disrupted in 9 months? If so, why not just use the same (non-disruptive) alert mechanism at 9 months?
His point is that Microsoft chose to do this in a kind of stupid way, and I have to admit I kind of agree.
My non-technical friends and family would have no idea what this actually means: "please upgrade this installation using a licensed distribution of this product"
As soon as you have so say "In other words," then perhaps, if you’d used those other words in the first place, it would be clearer. I know you don’t necessarily have control over these things – but just sayin’.
Even so, they likely *still* wouldn’t read it :-)
@”this guy” and “Larry Watanabe”
Look, there is absolute no excuse for “suddenly” being surprised that after 9 months you system starts rebooting. Also, since you acknowledge knowing that rebooting starts in 9 months, plan your testing schedule accordingly. Just because you want to plan for 12 months of testing, all the while knowing that you have 9, isn’t someone else’s fault.
Really you’re just complaining that you have 9 months. Why not just say that?
Steve Smith: Examples of bad error messages are everywhere. Not to pick on Microsoft, but SQL Server (yep, even SQL server 2008) has an error message that basically says “An insert couldn’t be done because one of your data fields is too long to fit in some column. Even though I know the column name, I’m not going to tell you what it is. I’ll let you figure it out yourself.”
Actually, the message is “String or binary data would be truncated”. SQL Server knows which column the string or binary data would be truncated IN, but it doesn’t put that information in the message. AArgh.
And I don’t see why the “string or binary data” part is there. Although if the column name was given, that part is not horrible; still, I would like to see “string data” OR “binary data” in the message, not “string or binary data would be truncated”.
Reminds me of Joel Spolsky’s principles of UI design:
1. Users don’t have the manual, and if they did, they wouldn’t read it.
2. In fact, users can’t read anything, and if they could, they wouldn’t want to.
If you want bad error messages, try MATLAB. They are rather consistently at the wrong place and with the wrong reason.
And then there is LaTeX, where a simple typo can result in a cascade of 100 errors, of which only the first (or sometimes the last) one makes sense.
@tim
"Then I told him to read the error message aloud to me, which he did. And then I looked at him and said "So…is your Caps Lock key on then?" (like the error message suggested)."
Here I was hoping you were going to say the the message was to see Human Resources…
The major misunderstanding among coders/programmers/managers in software companies is that a programming code or program execution "error" is *not* something that should be presented to the user as "an Error."
Just as an example, the latest web browser by the most successful software company in the world apparently checks if the encrypted connection is successfully established. If it isn’t, and user continues, what he sees as information is, in red letters: "Certificate error."
In that particular case, what’s most probable is that somebody is phishing the user, so the user should get anything else but "Certificate error." In the perception of the user: he didn’t do anything with these "certificates." He just clicked there. Them computers they are stupid.
Maybe something like "security alert: wrong site" which is much more meaningful to anybody who didn’t program that particular feature should be there.
If you "just click" and you get " Certificate error" (what is that!) of course the best you can do is to conclude that it’s futile to even read the messages that "the computer" (or maybe "hard disk" — everything is "hard disk" for a lot of users) spews out so often.
It’s really coders/programmers/managers who are making errors in most of the cases.
" I’ll have people come up and say things like "My computer at home has an error. Do you know what’s wrong?"."
The error is wrong. Just tell them that.
So often I think that it’s pointless to even add useful error messages to any software, users just ignore them. Same goes for doc. I wan thinking that we should add an Easter egg in our doc that promises a prize to the finder and see how many years it takes to be turned in.
"So often I think that it’s pointless to even add useful error messages to any software, users just ignore them. Same goes for doc. I wan thinking that we should add an Easter egg in our doc that promises a prize to the finder and see how many years it takes to be turned in."
There was a company that did this with their EULA. It was several years until someone emailed them about it.
I stuck a large poster above my mother’s computer that says "If all else fails, read the instructions".
Unfortunately this didn’t quite work the way I intended, she thinks it means that she should read them out loud to me (in excruciating detail) and then ask me what she should do.
@Dave: She could also interpret it in a way that "all else" includes asking you for help, so you’re still above reading the instructions on the list!
Actually, I’m saying that Microsoft should attempt to support that kind of upgrade from evaluation version to release version. And I’m not including beta and RTM versions.
Since they ought to carry similar version of libraries, I think this case will be more or less the same as repair installation of the target OS. (What do repair installation do when the disc is the basic version but the system has SP1 or higher version installed already?)
Maybe it’s pointless for users, but it’s life-saver for the support people.
When companies hire techinical supports, they don’t expect them to sit there and look blank to screen when they have error… or do you expect them to call support number every two days or so?
"Pre-emptive snarky comment: "The existence of evaluation editions with expiration dates proves that Microsoft is evil." "
Worse:
http://boycottnovell.com/2009/05/10/gratis-vista-7-is-like-free-tobacco/
OK, it’s true; Users are wombats. Even the ones who can read, can’t be bothered to read error messages, it’s far easier to ask someone else to fix the problem for them. Some of this attitude is our own fault, if we present users with messages that make no sense, there is actually no point in reading them, to paraphrase, Garbage out, garbage in!
However, there are some people who do read messages and the point I was trying to make is that it’s up to the people who develop the software, many of whom are reading this blog, to make the messages useful. We take so much care to make the UI intuitive and attractive (well, sometimes we do), we hire usability consultants and graphic designers and employ focus groups. We make sure that the product works, we refactor the code, we develop extreme testing scenarios and QA departments, then if the user tries to download a 5GB DVD image to FAT32 drive, we say “Drive full”. The drive isn’t full, it’s an invalid file. It doesn’t help the user to know that there may be a reason for this nonsense. They see stuff like this and, the next time an error occurs they say “it’s a waste of time reading these messages because they are nonsense”
To take an example, Raymond’s comment above about SQL server. He says “they could have created separate error codes for “string” and “binary”, I guess”. It’s actually fairly irrelevant whether it’s string or binary (or little green apples); The information that makes this error message most useful, is the name of the field! SQL server knows this information, it should be a responsibility of the developers to provide a mechanism for returning it. Sometimes it’s just too much effort, provide a feedback area for the fieldname and when I do, those b******s aren’t gonna bother reading it; Sod it, let ’em fry. I expected better from Raymond.
I’m not trying to pick on SQL server, it was just an example that cropped up, there are squillions of other products that are, at least, as flawed. It’s a mistake I have made too.
Well, SQL server has this information, why shouldn’t I expect it be reported in an error message.
Put it another way, you run a build or a make on widget.exe. 6 gazillion lines of code and 134 zillion source files. It mashes away making whirring clicking sounds and after 45 hours, after finding a missing semi-colon, it says "ERROR_SYNTAX_VARIABLE_NOT_FOUND". I can just see the look on your face. If YOU want decent error messages, so do the wombats.
I’m not asking for a core dump or print of the trace table or a novel, but a simple short sentence pitched to the user at right level; "Data truncation in table X at Row Y in Field Z" or "The update you requested to the customer table for customer "ACME" failed because the data supplied for customer name was too long" shouldn’t be beyond the wit of man (or woman). All it takes is a little thought.
Some programmers, the lazy unprofessional ones, will undoubtedly hate it, and no, they won’t get on board. They hate source and change control, or bug tracking or testing or even actually checking return values in their code; You know what, I don’t care, we would all be better off without these idiots.
To quote, from memory, one of your lot; "We don’t do these things because they are easy, we do these things because they are difficult".
"What do repair installation do when the disc is the basic version but the system has SP1 or higher version installed already?" -Cheong
In my experience, it usually results in a half-alive system that appears to work fine, most of the time. There always seems to be a few mismatched libraries, making the system better off having the data backed up and the OS rebuilt. (And reapplying any service packs and hotfixes, that will reapply, has never helped in my cases.)
"[How about if they said "Expires in 9 months (with a 3 month grace period)"? -Raymond]"
Much better, although if you’re advertising it, people may think the grace period is just like a normal one.
Better to say ‘with a 3 month grace period with limitations’.
Or you can not advertise it at all. Really, if you’re said "It’ll work for 9 months", and in fact I can (doing things harder) still use it for another month, I won’t complain.
So after the trial period is gone, you’d place a blocking MessageBox on load "Your trial period is over. Install a buyed OS. [Shutdown] [Let me enter to backup my things, please, please]"
You /should/ have a backup ready for the expiration time, but i could easily get caught on it, too. That messagebox forces you to *read* it.
Then present another one ‘You have two hours to pack your things and ge out of here’, and present a countdown timer (but not topmost!).
The only problem I see with this approach is that there will always be someone trying to backup things to a DVD in a process which takes 2 hours and 10 minutes [and calling customer help to complain about that]. But there’s not much that can be done with all of these, since for every duration someone will need a few minutes more.
Even if SQL-server could tell you the column, it probably would omit it anyway because of security concerns. It might show up on a wegpage and give an attacker an easy way to learn column names
The better "IErrorInfo" would be
char*
;-)
I think that guy at IT PRO is misunderstanding the situation, as did I for about 45 sec until I reevaluated the premise. It is, in fact, bat sh-t crazy to force a PC shutdown every 2 hours, beginning 3 months BEFORE expiration. However, that’s after NINE MONTHS of use (assuming a trial period of one year).
So really, you get a 9 month evaluation (~270 days, which is more than double the previous standard evaluation term that I remember) with a 3 month grace period. Not a bad deal for zero dollars.