Frankly, I find it strange to have a <title> tag that is no such thing.
In other words, the <title> cannot be found in the <title>, it must be
found in <nonSort> (if such exists) and <title>. How intuitive is that?
But if that is the collective wisdom of the group, I can live with it.
Meanwhile, to use a local example, this is a sample from one of our
MODS records for our online books (see

<nonSort>A </nonSort>
<title>shield in space? [electronic resource] :</title>
technology, politics, and the strategic defense initiative : how the
Reagan Administration set out to make nuclear weapons "impotent and
obsolete" and succumbed to the fallacy of the last move

It is utterly unusable for any public display (e.g., what is
"[electronic resource]" doing in the <title> tag?!). Thank god we have
another metadata record from an internal UC Press database, from which
we can extract information we can actually put in front of a user. I
would like to see MODS actually become something usable beyond
perpetuating the many mistakes of MARC/AACR2, which I believe is the
rightful role of MARCXML.

On Jan 23, 2004, at 8:13 AM, Brian Tingle wrote:

>>>>>> [log in to unmask] 2004-01-23 10:20:12 >>>
>>> You should be able to define them as xs:string, in which case white
>>> space is retained. Well, in theory, anyway.
>> On Fri, 2004-01-23 at 07:30, Andrew E Switala wrote:
>>> Whitespace normalization is done at the level of the XML parser, so
>>> by
>>> the time schema validation happens the attributes have already been
>>> normalized.
> On Fri, 2004-01-23 at 07:49, Karen Coyle wrote:
>> Well, it sounds like "the parser" isn't schema compliant.
> Well, you have XML parsers, and some of them optionally implement
> schema
> validation and some of them don't.  But an XML parser that is just
> trying to get DOM from a well formed XML document has to make the same
> DOM weather it is bothering to do schema validation or not.
>> In which case
>> the same would be true for the strings within the <nonSort> element.
>> So
>> this whole nonSort thing really doesn't work.
> That does not follow at all.  He made an assertion about *attribute*
> space normalization rules in XML parsers, not white space in
> *elements*.  My understanding is that whitespace rules in elements are
> not the same as in attributes.
>> In which case, we should
>> go with Roy's solution, even though that's not how MARC does it.
> Okay, now this is really a stretch.  Please leave it the way it is
> now:)
> -- Brian Tingle
>    Content Management Designer
>    California Digital Library