[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-usability
Subject:    Re: [KDE Usability] Users cannot find where to "safely remove" USB
From:       Peter <gostelow () global ! co ! za>
Date:       2010-04-12 17:53:34
Message-ID: 201004121753.37270.gostelow () global ! co ! za
[Download RAW message or body]

On Sunday 11 April 2010 21:26, Peter Grasch wrote:
> On Sunday 11 April 2010 20:02:28 Peter wrote:
> > On Sunday 11 April 2010 16:35, Dotan Cohen wrote:
> > > > That is of no interest for the user. Nowadays USB is highly used for
> > > > removable storage, operating systems need to adapt to that. All the
> > > > time some system is not used for what it was intended; but it is used
> > > > because it serves a need.
> > >
> > > I agree. My comments about "how to implement that" were not
> > > disagreement, rather, they were serious questions. Like I said, the
> > > machine should adapt to the user. Now come up with the technical
> > > implementation.
> >
> > While I understand your interest, I question whether this list is
> > appropriate for such a discussion. I raised the point that the problem
> > was a system one so members would realize KDE (with other desktop
> > developers) are attempting to resolve a problem they should not, imho,
> > have to address.
>
> To really fix this at a system level you would need to implement
> transactional file systems - file systems that can handle interruptions at
> any time without taking damage. Those exist but are not commonly used.
> Microsoft tried to push an extension version of FAT (TFAT) and even they
> failed to get users to use it
> (http://en.wikipedia.org/wiki/Transaction-Safe_FAT_File_System).

I agree TFS improve data integrity, but they do extensive writes, which 
reduces the memory's lifespan considerably, so sticks will wear out faster. 
GNU type systems require file systems to be added to and removed from a 
single file structure. Removing a stick, while it is still mounted, confuse 
these systems, even though the stick itself is okay.

>
> Personally I think the main issue is awareness. In my experience most
> people are ok to wait for something (within reason) if they know how long
> it's going to take.

There is no disagreement that users must remove sticks after applications are 
done with them. The question is what should happen when users don't wait. 
Clearly, users want their sticks to be automatically repaired. Any example 
therefore must include removing and reinserting the stick before a successful 
transfer. Furthermore, it should show what happens when a user removes one 
stick and inserts another.

>
> I seriously think that the performance drop of synchroneous disk access is
> an acceptable price to pay for that extra data safety. All the
> infrastructure is already there. With sync mounts, the progress indicator
> of transfers displays the _real_ transfer - complete with ETA.

While I appreciate the thought given in your examples, they don't explain how 
the system will recover when users remove sticks during transfers. The 
expectation is that the transfer should continue normally.

Removing the stick does not only affect the file being transferred, but can 
cripple the stick's whole file system, if it happens to be updating the 
directory at that instant. Most applications are incapable of accessing the 
underlying file structure and will need some utility to do the repairs for 
them.

Regards,

Peter
_______________________________________________
kde-usability mailing list
kde-usability@kde.org
https://mail.kde.org/mailman/listinfo/kde-usability
[prev in list] [next in list] [prev in thread] [next in thread]