Hi,
Am converting HTML to XML using java, XSLT. The HTML contains element having attribute value with white spaces, for example
<span name="There is double space here">some text</span>
The name value contains two white spaces. When I convert this to XSLT, the extra white space is stripped. I tried to use the but this is not working.
Actually I want to preserve whitespace only for attribute values. How can I preserve extra whitespaces for attribute value for all elements.
Thanks,
Kapil

You haven't said how you are converting the HTML so it's hard to say where it's going wrong. Handling of whitespace in XML attributes depends on the DTD-defined attribute type - for attributes declared as CDATA (which is the default) newlines and tabs are converted to spaces, but multiple spaces aren't collapsed. For some other attribute types, for example ID and NMTOKENS, multiple spaces are collapsed to a single space. Perhaps this is happening because your attribute <span name="..."> is defined as an NMTOKENS attribute. It's not clear to me that an English sentence is an appropriate value of the "name" attribute.

xsl:preserve-space does something completely different. Please don't try to guess what a construct does from its name - look it up in the spec. Programming by trial and error will get you nowhere.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Thanks for your reply Michael.
I am converting HTML to XML by passing html string to java XALAN transformer. There is no DTD associated.
Actually I had used the name attribute just as an example. Am getting data in the form of XML from a third party application and I am converting it to HTML for user editing and then transforming back the HTML to XML. There are some attributes which I have to preserve during this round trip transformation. Some of these may have double white spaces as value. I have to preserve value of these attributes.
I tried this transformation using XML spy and it correctly preserves the white spaces but not with java based transformation.
Thanks,
Kapil

Then I think Xalan is getting it wrong. Raise a bug report - or try Saxon ;)

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Could there be any alternate way to overcome this problem.
Thanks,
Kapil

I've given you two suggestions, how many more do you want?

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Ok, I will try to use Saxon for java.
Thanks for the help Michael.

Michael, I read the specification
http://www.w3.org/TR/REC-xml/#AVNormalize and according to this the attribute value should not contain multiple spaces.
Does Saxon still convert multiple spaces as it is?
Thanks,
Kapil