Re: wifi usb adaptor

The message <k2uj235edjt4s4vrn81qa6qhhbdoiss0k8@xxxxxxxxxxxxxxxxxxxxxxx>
from Jaimie Vandenbergh <jaimie@xxxxxxxxxxxxxxxxxxxxx> contains these words:

On Sat, 21 Apr 2007 12:27:46 +0100, Johnny B Good
<jcs.computers***@xxxxxxxxxxxxxxxx> wrote:

The message <lunj23pm1jdkke0p76431m8tjb1rcc7445@xxxxxxxxxxxxxxxxxxxxxxx>
from Jaimie Vandenbergh <jaimie@xxxxxxxxxxxxxxxxxxxxx> contains these
words:

I've noticed, just recently, that there can be some variation in the
'tightness of fit' between different makes (or just samples of the same
make) of keyboard so it might be a case of your favourite keyboard
having a plug that's at the loose end of the range.

True, but not in my case (and my favourite keyboard is USB!). The
old-keyboards stack is occasionally raided, but it's normally coiled
cables and/or routing thereof causing pop-out.

I don't think I've any such coiled corded keyboards left in my 'spares'
collection (unless there's one lurking in the bottom of the heap :-).

I've often unplugged and replugged PS/2 keyboards and mice 'live' after
letting the system box boot with those devices plugged in in order to
make sure the interfaces were enabled during the POST.

Do you mean your hardware believes in them if you hotplug after POST?
I've never come across that.

Generally so (but they sometimes don't get recognised on plugging back
in) it depends on the OS I think (and which way the wind is blowing).

It strikes me that, with only a single 5v supply rail (plus ground
return), there's only exactly the same risk of damage as is present in
the USB interface when 'hot plugged', i.e pretty well none.

I've seen three motherboards popped from hotplugging keyboards - one
IBM, one Compaq, one Shuttle.

Possibly a lack of polyfuse protection, as per the infamous PC Chips
MoBos (I rescued one such board that had suffered 'user clumsiness'
damage to the USB ports which had blown a 5 volt circuit trace feeding
the usb _and_ the PS/2 ports by bridging with a 3 amp polyfuse after
replacing the damaged ethernet/usb rear panel socket assembly).

USB plugs are designed for live plugging, and have appropriate voltage
management.

They might be, but that doesn't prevent cack handed users from
destroying the socket (that PC Chips example above wasn't the first time
I'd seen such damage) and, in the absence of a polyfuse, blowing a
circuit trace rendering not only all the other usb ports powerless but
quite often the PS/2 ports powerless as well.

This turns out not to be the case - thus my phrasing above:
http://www.fs-driver.org/faq.html#acc_ext3

Well, I've just _re-read_ that page, and, just as I remembered, the
EXT3 volumes are simply being treated as Ext2 volumes by IFS.

D'oh. I was reading them thus: if they are being treated as EXT2, then
the journals would never fill, and the warning about IFS not fscking
them wouldn't be appropriate.

I wasn't thinking of sharing the FS between a real ext3-supporting
system and the Windows box.

Aye, could well be. Good old VIA. Maybe a £5 PCI USB card would help?

Yes, I've considered that, but the only USB2 PCI cards I have in stock,
are also VIA chipset so no point in trying them

(rummage) Yeah, me too. Does anyone else even make the things?

Actually, I just rescued my last combo card from being sold (I'd
mistakenly reserved it for a job that only required a firewire card). It
happens to be sporting an ALi chip which may or may not be an
improvement over the VIA chip (I'm not impressed with ALi chipped MoBos
of recent vintage). I'll pop the 3+1 firewire card out and replace it
'real soon now'. If it eliminates the problem, I'll probably not know
for at least a month unless I can try a 100GB transfer test (the FS
would go awol around the 50 to 70GB mark during any large backup jobs
with the current hardware).

3+1 firewire card to save 'crowding' the existing adapter card
population (there is a spare slot, but I do so hate crowds, don't you?
:-).

I actually had all slots filled in this box a while back - AGP plus
all 6 PCI. I'd never managed that before.

I'm back down to 1 PCI used now, much better.

In the circumstances, even the humble Ext2 is a vast improvement over
NTFS in the event of FS 'damage'. FSCK, at least, addresses the problem
in a much more efficient way than the 'bruteforce data recovery' methods
used by the commercial NTFS repair utilities which require hours rather
than minutes to grind away at the problem and require you to have
another drive spare to copy the recovered files to.

The design of NTFS should avoid most situations where you need to do
such grinding, though - the filesystem journaling (not data
journaling) is way better than the superblock-duplication of e2fs.

The major problem with NTFS is, as far as I can tell, the lack of any
decent FS _repair_ utilities within the OS. All the FS repair utilities
that can handle NTFS seem to be based on the assumption of data recovery
rather than actual FS repair

Along with mice and Xbox Live, NTFS is one of the very few decent
products from MS. Made rather useless outside Windows by being closed,
of course.

Agreed, but it's the lack of security in the OS that turns the
'security features' of NTFS into a deadly weapon aimed against the
typical home user. NTFS's resilience is probably the reason why a
knoppix boot on a fried winXp box (where chkdsk from the repair console
has simply, and immediately, curled up and died complaining it found
irrepairable problems) has been able to access users' data and allow
copying to another machine via the samba service.

If you know of an Ext3 FS driver add-on for NT5.x based systems, please
let me know. :-)

I'm holding out for ZFS/win32!

Looks like ZFS/OSX is going to happen in Leopard, though - that'll do
me.

A bit of overkill. I'd settle for Ext3 support (and a win32 version of
FSCK :-).

--
Regards, John.

Please remove the "ohggcyht" before replying.
The address has been munged to reject Spam-bots.

.