Problems editing XML documents

[charleski]

Originally Posted by pietvo

Firefox saves the file as is, including the xml declaration. And rightly so, because it saves files verbatim.

Actually, it strips the declaration and leaves the rest of the html untouched.

Originally Posted by Strether

It opened just fine in Word 2003, too, but with the special symbols displayed; and if they can be hidden, it would be nice to know how to do that.

I use Word 2007, so not sure if this applies to 2003 as well, but if you right-click a tag and select 'Show XML structure' there's a checkbox labelled 'Show XML tags in the document'.

Alternatively, you can apply an XSLT to the XML in Word. I've attached a very simple one which will just convert the declaration to html. Browse to it in the XML Data Views pane (again, this in Word 2007; 2003 may be different) and apply it to see the document without tags (though Word seems incapable of retrieving style info from an external css in this mode).

In general, it's easiest just to strip the xml declaration if you want to edit these files in Word.

[pietvo]

Originally Posted by charleski

Actually, it strips the declaration and leaves the rest of the html untouched.

Not for me. (Firefox 3.6.2 on Mac OS X). And I would find it nasty if it did.

[charleski]

Originally Posted by pietvo

Not for me. (Firefox 3.6.2 on Mac OS X). And I would find it nasty if it did.

Ah, checked again and found out what's happening - saving as 'Web Page, complete' strips the xml declaration, but saving as 'Web Page, HTML only' doesn't.

[pietvo]

I see. In that case it makes many more (syntactic) changes to let it look more like HTML instead of XHTML.

[charleski]

Originally Posted by pietvo

I see. In that case it makes many more (syntactic) changes to let it look more like HTML instead of XHTML.

Firefox will do a few tidying operations such as adding an http-equiv entry if the input doesn't have one and converting some empty tags like <p/> to <p></p>. But it doesn't modify fragment identifiers, which could trip up some user agents. Firefox doesn't interpret <br></br> correctly and therefore messes up the output, but that's a fairly unlikely construct (though valid in xml).