Re: versioning use case

read the article, very nice.  I only had one comment and hopefully I am 
not blantanly showing my ignorance of XQuery here due to never using it 
yet.  I believe XQuery is just missing a feature.  Instead of

   for $x in //xml-deviant 
    return $x/signature

It should be 
   for $x in //(DareObasanjo)xml-deviant
     return $x/signature 
(If the xml-deviant is not of DareObasanjo, you would get a runtime exception just like .NET)

I still think extension could be a very powerful mechanism if we 
required it to be that powerful mechanism that behaves as closely as 
possible to OO.  I sort of agree with the comment the first person 
posted on your article.

By the way, my base type unfortunately doesn't have a wildcard, and 
specs change very slowly.  They unfortunately forgot one in the error 
codes that are sent back to a user.  Until it is fixed, I am kind of up 
a creek as I have to meet the standard.

Totally agree with you on restriction.  In fact, I think restriction is 
basically one form of deprecation because if some clients are using the 
non-restricted element that is in the standard and you extend by 
restriction, their client will no longer work with you, and you are now 
in violation of the standard.  Noone can ever really use restriction to 
add on to a schema and expect to be compatible.

thanks,
dean

Dare Obasanjo wrote:

>comments inline 
>
>________________________________
>
>From: Dean Hiller [mailto:dhiller@avaya.com]
>Sent: Sat 11/22/2003 10:12 AM
>To: Dare Obasanjo
>Cc: www-tag@w3.org
>Subject: Re: versioning use case
>
>
>
>This is exactly what I mean though.  A company has extended the schema
>using "extension", yet I don't want their additional features.  I just
>want to validate the base schema is ok.  I can't.  It is all or nothing.
> I have to validate base schema plus the companies additional
>elements(their extension), or I can't validate it at all.
>
>[Dare Obasanjo] If the base complex type has a wildcard with processContents="lax" then you can get what you want. 
>
> The only thing is I think it is much cleaner to do extension.
> It is like OO. 
>
>[Dare Obasanjo] W3C XML Schema isn't like OO, it seems like it is but it isn't. I've gone into this in more detail at http://www.xml.com/pub/a/2003/10/29/derivation.html 
>
>  
>