Re: Effects of Magsafe...

zoara <me3@xxxxxxxxxxx> wrote:

On Mon, 10 Jul 2006 13:18:16 +0100, Woody wrote:

zoara <me3@xxxxxxxxxxx> wrote:

On Mon, 10 Jul 2006 09:52:47 +0100, Woody wrote:

True. I was thinking of memory keys and the like rather than USB hard
disks.

Given the size of memory keys these days I don't think there's a real
difference.

There is a difference in likely usage patterns. It is more likely a
drive will be plugged in all the time and a key wont.

Mmm, maybe. My USB key tends to be plugged in for longer periods than my
USB disk, excluding backups.

I would imagine my use is more normal though.

And backups really would benefit from a
"silently unmount" option, since the backup can resume where it's left off
when the drive comes back.

True

I'm talking from the perspective of a laptop, by the way, where yanking out
all the cables is a much more likely scenario than with a desktop.

Indeed. That is what I assumed.

However, there is nothing so annoying as a drive that won't eject
because something is holding it open. Ultimately you are still going to
remove it anyway.

Exactly my point. If a user is going to yank it anyway, why not get the
system to protect against data loss / corruption when the drive is yanked?

How can the system control that? If the drive is removed, what can the
system do about it? the drive is not there any more, so it cant. There
is nothing to stop half of that file being on the disk.

It can protect against data loss and corruption. Just because half the data
is on the USB drive and half on the internal drive doesn't mean that the
data should be *lost*. Just that it's halfway through a process.

True. Or at least it should be an option. At work I have an external
USB2 on a windows machine that contains datasets. A dataset is generally
around 1-4Gb and consists of 15000 xml files and 30000 graphics.
I don't want the system to ensure that every file is safely written
copying things to that as it would be a lot slower and the drive is
never unplugged.

As it stands, yanking a drive can cause dire problems. I'm proposing that
yanking a drive is coped with in a more fault-tolerant manner.

Hardly dire consequences though. I have only had one problem pulling a
drive out, and that just ended up with a load of files that it said it
existed and they didn't.

Why shouldn't the system just cater for you unplugging the cable without
kowtowing to its wishes? I thought computers were supposed to be our
slaves, not the other way around.

I work at home. I bring home my USB key. I merge the files from my key
to my active project. I work on them all day, I merge them back to the
key.

If I get home and the files are not there, the whole process is screwed.
It doesn't matter whether I don't want to bend to my computers wishes or
not, but if the files haven't finished copying, then they haven't
finished copying and even with the best programming in the world there
is nothing you can do to change that.

Ah, now there's something I hadn't thought of. I was thinking of the "pull
out the disk to take the laptop downstairs" scenario, where a half-copied
file would just mean going back upstairs and plugging the drive back in so
it could finish copying.

No, the files are considerably further than that.

Okay, how about this? Instead of the current "Oi, you muppet, you unplugged
the drive before I was ready" dialog (which appears regardless of whether
data is being written to the drive), why don't we change the dialog so it
only appears if a read/write operation is active, and get it to read "Oi,
you muppet, I was halfway through using that! If you want to finish copying
the data, plug it back in".

Hmm. that would work, although I believe windows does that. Although I
doubt it continues afterwards.

However, I don't think you are stupid if you put your hands in a fire; not
the first time anyway, if you havn't been told. It's lack of knowledge, not
stupidity.

yes - that was covered by : 'or not know better in which case you should
have people looking after you'

Okay. I misread that as "you're such a stupid muppet that you need people
to look after you". But you're right - if you lack certain knowledge, it's
good to have people around who have that knowledge who can look after you.

I don't think that precludes having safety mechanisms in place for those
people who lack knowledge and don't have mentors helping them every step of
the way, though.

Indeed, as long as the functionality isn't reduced by too much to deal
with them (or is switch offable though).

What I'm saying is that the system should allow for drives to be unplugged
without any notice, and should mitigate the risk of data loss in this
circumstance as much as possible (ie it should be "no risk at all"). And
the system shouldn't scold the user for 'doing the wrong thing'.

ok, so back to the way windows works then.

As I said, "That's halfway to what I'm talking about, but still doesn't
protect you if you're partway through writing a file when you pull it
out.". I actually think that the Windows method is less useful as it
doesn't encourage 'safe' removal. The Mac way, whilst still not perfect, at
least alerts the user to the fact that they did something wrong.

No, the mac way is completely irritating. The windows way at least bugs
you less (I haven't tried with vista - probably the same with more
dialog boxes).

Perfection would be to allow yanking without falling over or throwing a
hissy fit. Coping with expected human behaviour, in other words.

No argument there.

--
Woody

www.alienrat.com
.

Relevant Pages

  • Re: Copying a Hard Disk Containing Windows With Fedora
    ... His basic mistrake was using partitions rather than copying the entire ... fdisk showed my Windows HD had a main partition and a "hidden" ... windows from a PARTITION to a new partition of another drive, ... > Subject: Re: Copying a Hard Disk Containing Windows With Fedora ...
    (Fedora)
  • Re: Cant copy a file on hard disk
    ... I tried copying from the backup. ... Nothing seems to be wrong with the hard disk, ... Windows estimated the copying time at around 4 minutes. ... So it looks like I'll be getting another hard disk. ...
    (alt.sys.pc-clone.dell)
  • Re: Ouch!
    ... >>> wouldn't affect usb key applications as windows is handling the file ... That doesn't mean that they're a FAT disk, ... An IDE HDD is not a Windows IDE HDD, it's an IDE HDD that can be used on ...
    (Fedora)
  • Re: RAM usage during I/O
    ... >> is a Windows problem not Nero's because the same happens if I'm copying ... >> defragment disk because I already did that and it didn't work for a long ...
    (microsoft.public.windowsxp.help_and_support)
  • win2k will not start - missing ntoskrnl.exe after Partition Magic copy..
    ... Windows 2000 could not start missing ntoskrnl.exe ... After copying I replaced old disk with the new one, ... It's same size, first on the disk. ...
    (microsoft.public.win2000.general)