[PLUG] XML, text, and the development of unix

[Jeff Abrahamson <jeff@purple.com>]

I recently attended a talk by Dennis Richie, one of the original
authors of unix.  I thought later that I ought to have attempted to
ask him the following question:

He (or one of his colleagues) once said, in espousing unix's tools
approach to commands, encouraged the use of text, calling it the
universal interface.  Indeed, one of the reasons that the selection of
tools available to us in unix is so rich is that so many tools take
text as both input and output.  It allows me to pipe this email
through (grep foo | awk '{print $2}' | sort -u | wc -l) or some other
such thing probably unanticipated by the author of any one tool.

Thirty years passed, and we are beginning to see the rise of XML.  A
lovely thing, XML, but I see it pushing aside many venerable
text-based file formats.  Where once we might have been happy to let
some tabs or spaces or colons provide our documents with structure,
now we want XML.

But, as rich and valuable as XML is, it makes ad hoc tool use more
difficult.  For example, if we were designing unix today, we might
spec the system files to use XML.  But, for example, now I might say

    grep /bin/bash /etc/passwd | wc -l

to find out how many users have set bash as their login shell, I'm not
sure how I could do that in a single line if /etc/passwd were
/etc/passwd.xml.

What are we losing, I would have asked Richie, as more and more files
become XML?

I know some of you, especially sys admins, have extensive experience
both with XML and with the need to query in an ad hoc way using these
text-based tools.  What are your experiences, pro and con, with this
issue?

-- 
 Jeff

 Jeff Abrahamson  <http://www.purple.com/jeff/>
 GPG fingerprint: 1A1A BA95 D082 A558 A276  63C6 16BF 8C4C 0D1D AE4B

Attachment: pgpNAW6gMX19e.pgp
Description: PGP signature