> Date: Tue, 14 Dec 2004 22:03:35 +0000
> From: Dr Robert Sanderson <[log in to unmask]>
>
> There would be three potential POST services, all living at the same
> end point:
>
> SRU (no XML at all, just posted parameters, including x-foo)
> SRX (Just the XML, no SOAP wrapper)
> SRW (SOAP wrapper + XML)
Although you're right to spot the possibility of SRX, I think we can
(and should) continue to ignore it. I don't see that it offers
anything that the other two don't, and confuses the discussion.
> Currently SRW has pride of place, as it was expected that SRW would
> be at least as frequently implemented as SRU, which turns out not to
> be the case.
No indeed.
> But really, if you need Very Long strings to be sent, then just use
> SRW not SRU? Not easy to do from a web page form, granted, so there
> is some appeal to SRU.
We have learned -- the hard way -- that SRU is _much_ easier to
implement than SRW. On the client side, it is a literally trivial
matter of an HTML form to generate the URL; the server is slightly
less trivial, but really, a 20-line Perl script can implement an SRU
server.
By contrast, it seems that _no-one_ (with the possible exception of
Matthew) like working with SOAP. WSDL is nightmarish, and all the
SOAP toolkits out there just plain don't work much of the time. When
one does, another will misinterpret (or just refuse) the WSDL. My
impression -- and please, others, feel free to disagree! -- is that
SRU is very much winning the implementation race; I think SRW may
quite soon be seen as an irrelevant offshoot.
If that's a correct interpretation -- that the world is joining in
SRU's party, leaving SRW to sulk in the kitchen, mumbling into its
paper cup of 2003 Bulgarian cabernet sauvignon and complaining that
all the good food's already been eaten and there's nothing left except
carrot sticks and "crispbread", which isn't really bread at all --
then we need to make sure that all the people flocking to implement
SRU don't all turn around and flock right back the way they came in
six months when they have to submit a compex request and find their
URLs arbitrarily truncated.
What we're proposing here is a really, really simple fix which doesn't
seem to have any downside (theoretically at least): we let SRU client
implementors change
<form method="GET">
to
<form method="POST">
when their URLs get too long.
You know it make sense.
_/|_ _______________________________________________________________
/o ) \/ Mike Taylor <[log in to unmask]> http://www.miketaylor.org.uk
)_v__/\ "I took a speed reading course and read 'War and Peace'
in twenty minutes. It involves Russia" -- Woody Allen.
--
Listen to free demos of soundtrack music for film, TV and radio
http://www.pipedreaming.org.uk/soundtrack/
|