> Date: Wed, 9 Apr 2003 10:09:15 +0100
> From: Matthew Dovey <[log in to unmask]>
> I've copied this to the main SRW developers list in case they wish
> to add (or contradict) what I say.
Thanks for doing this, Matthew. I will now take you up on your offer
Alexei, I am involved in the SRW effort because at the moment it seems
inevitable that the XML bandwagon will eventually roll all over
traditional IR practices. So it's better to be _on_ that bandwagon
than underneath it :-)
Nevertheless, for actually getting work done, Z39.50 remains a much
better choice than SRW today, primarily because It's There: the
implementations are out there, tested and battle-hardened, deployed in
varying situations and doing the jobs they're meant to do. In
contrast, SRW -- while it may well evolve into a stable, reliable,
protocol -- has not yet to my knowledge been deployed outside of
academia. It simply lacks maturity (and hence solidity.) And a brief
scan of this mailing-list's archives will show that SRW
interoperability is still afflicted by the level of difficulties that
Z39.50 implementations resolved the best part of a decade ago.
So if Z39.50 programming is so hard, what's a poor boy to do? As
Matthew says --
> People working with Z39.50 spend quite a long time getting to grips
> with the transport and encoding before even touching the actual
> standard itself.
> ii) In some ways, SRW is a marketing exercise to attempt to dispell
> the myths that Z39.50 is necessarily complex.
If you don't want to mess with low-level protocol stuff -- and why
should you? -- and if you're looking to implement IR _clients_, then
what you need can almost certainly be found in the ZOOM project
(zoom.z3950.org), which furnishes easy-to-use, fully-functional Z39.50
application libraries in a selection of languages including C, C++,
Java, Perl, Tcl, Visual Basic, Python, PHP and Scheme.
Writing Z39.59 clients is _much_ easier now than it was in the days
that SRW was first conceived.
If you're interested in the historical background, I wrote a few
short, informal papers that contain one man's (no doubt biased)
perspective. They are are on-line at
and, if you can put up with a certain amount of frothing at the mouth,
/o ) \/ Mike Taylor <[log in to unmask]> http://www.miketaylor.org.uk
)_v__/\ "Never ruin an apology with an excuse" -- Kimberly Johnson.
Listen to my wife's new CD of kids' music, _Child's Play_, at