Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Windows 95 almost had floppy insertion detection but the training cost was prohibitive (msdn.com)
37 points by nebula on April 2, 2009 | hide | past | favorite | 14 comments


I still don't get why they couldn't just spin the disk up, check if it exists using that message and then check the bit at the same time. Do that once during installation and there's no user intervention.


From the article:

"You can't just try to figure out what type of drive the user has by comparing the clever technique against the boring "turn on the floppy drive light and make grinding noises" technique, at least not without displaying a warning to the user that you're about to do this—users tend to freak out when the floppy drive light turns on for no apparent reason."

What they should have done is wait for the disk to be used for some other reason, and then test it, and record the results. Just in case the drive changed, then after every reboot verify that it still makes sense.


I saw that, and it was a cop-out excuse in my opinion. During installation, usually unattended too, your computer does all sorts of unexpected things. It reboots, the screen switches off and on, it freezes up for a while. Most old systems had the floppy grind away on boot as well. I fail to see the issue (and it doesn't sound like they even tested it with any users).


But the article provided a solution, too:

"You can't just [use the] "turn on the floppy drive light and make grinding noises" technique, at least not without displaying a warning to the user that you're about to do this—users tend to freak out when the floppy drive light turns on for no apparent reason." (emphasis added)


That would work until your floppy drive dies and you replace it. Then there's a 50% chance that your floppy drive always shows the wrong status (since there's no plug-and-play with floppy drives, and therefore no way to know when to redo the bit check).

That alone would be enough of a reason to cut the feature; imagine how much tech support would go into supporting this one issue (at a time before the internet was ubiquitous).


How about rather then at installation time, but the first time someone queries for the floppy drive? Every time you boot it will always spin the drive the first time, but not after that. That seems much better then always spinning up the drive.


Not to mention that they were Microsoft in the days of Windows95. If anyone could get floppy manufacture's to standardize they could.


I would guess that one problem with this approach is that in 1994/95, the setting for the existence of a floppy drive in the BIOS did not always work properly (and sometimes there was no setting for "no floppy"). Therefore, if windows tried to check for the floppy on boot when you had no floppy installed, it would hang the system for a long period of time.


If someone replaced the drive, then if you erroneously detected a floppy's in the drive, you try to access it and.. nope, no disk there. So then you flip the system. Fixed.

As far as I recall, the error codes for no disk versus an erroneous read were different, so you could just figure it out in case of a false positive.


The rationale given in the article was that a drive starting up with no obvious cue by the user would lead to confusion or concern on the user's part. Back then, before the widespread availability of the 'Net, users would probably think a "boot block" virus was attacking their system or some other scary scenario was happening.

Of course, I was a Mac-only user back then, so we had auto-detect (and auto-eject!) drives and learned to take them for granted. Part of me thinks the historical suckiness of DOS/Windows is less the fault of Microsoft and more the fault of the IT community of that era, who were tone deaf regarding end user satisfaction and empowerment.

(Even now, using Windows machines at work make me feel constrained compared to the relative freedom to do what I wanted with my classic era Macs. Sadly, since OS X, Macs have started to feel less "free" but I'm starting to digress too far afield now...)


Don't forget those excellent Macs where the (instant-off) power button looked like the floppy drive eject!

Oh what fun!


You know what's funny?

Most classic era Macs used to have a power off key on the keyboard, instead of requiring a button on the case proper. (Like FireWire, ADB was always "hot" and such machines used a rocker switch instead of a push button for "master power off.") But since Apple adopted USB, this functionality was taken away due to how USB is (was?) dependent on a actively powered CPU to arbitrate traffic.

Guess what that same button is used for now?


My thoughts exactly. But I'm sure there was more to it, or else they would have done that.

Still very interesting to hear the constraints faced by teams working on such important software.


I don't quite get why it matters. Guess, then when the drive is properly used for the first time, test, change over if required.

After that, quietly stop spinning up and grinding the floppy drive when not required.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: