• To: "Michael Rys" <mrys@m...>
• Subject: Re: xsi:nil is a wart (was Interesting pair of comments)
• From: Paul Downey <paul.downey@w...>
• Date: Fri, 15 Jul 2005 08:33:17 +0100
• Cc: <xml-dev@l...>,"Michael Champion" <michaelc.champion@g...>
• In-reply-to: <BAE415CEAA831548800F68C44E905AF5029F3D8F@R...>
• References: <BAE415CEAA831548800F68C44E905AF5029F3D8F@R...>

On 15 Jul 2005, at 06:13, Michael Rys wrote:

> Xsi:nil is not central to data apps. It is an ugly wart...

Actually, I'd agree. minOccurs is usually good enough for
my purposes, except I've found myself littering schemas with
nillable="true" against each minOccurs="0" simply because some
tools insist on sending xsi:nil regardless.

I had to teach third line support from one notable vendor how
to read the spec before they conceded that this behaviour was
indeed a bug.

--
http://blog.whatfettle.com

• References:
  • RE: Interesting pair of comments (was Re: Schema Experience Workshop minutes online)
    • From: "Michael Rys" <mrys@m...>

• Prev by Date: Mixed content in data-binding (Was: Re: Interesting pair of comments (was: ...))
• Next by Date: Re: Mixed content in data-binding (Was: Re: Interestingpair of comments
• Previous by thread: RE: Interesting pair of comments (was Re: Schema Experience Workshop minutes online)
• Next by thread: can xml schema represent such a instance xml document here?
• Index(es):
  • Date
  • Thread