In message <Pine.LNX.4.61.0512102050160.1950@m24s24.vlinux.de>, Hubert Feyrer writes:
> On Sat, 10 Dec 2005, Igor Sobrado wrote:
> >> ===>	# cat <boot-big.fs >/dev/wd0b
> > [...]
> >
> > Just curious... why a cat(1) instead of a "dd if=boot-big.fs of=/dev/..."?
> > dd(1) seems more robust to raw copy filesystems to/from drives.
> 
> Try using tab completion with dd. :)

No problem at all!  dd if=/dev/rand<tab> -> dd if=/dev/random ...  :-)
(pdksh 5.2.14... I am more accustomed to the Esc-Esc sequence yet
but as you asked about *tab* completion...)

> No idea what you mwan about robustness...

I was thinking on cat(1) as a catenation (what a synonym for concatenation!)
command for files.  I suppose that it works with binaries and that block
size is not an issue but I certainly trust on dd(1) for copying device
related information.  I remember some issues with cat(1) in other Unix
flavour some years (perhaps a decade) ago when copying raw binary data.

On an unrelated issue: I really enjoyed the column of Steven M. Bellovin
(published with Matt Blaze and Susan Landau) on the november issue of
Communications of the ACM.  A fine analysis of the consequences of a
bad cryptography policy in weakening Internet security.  Nice to see
that people as him is working closely with TNF.  I liked his column
on spamming, phishing, authentication and privacy on CACM on december
2004 too.

Cheers,
Igor.