RE: Elements vs. Attributes - was Re: [ubl-ndrsc] Minutes for 6 M arch2002 UBL NDR meeting

["Stuhec, Gunther" <gunther.stuhec@sap.com>]

Hello Eduardo,

I guess, "amount" is a Basic Core Component (BCC) for example. We have to
define all the BCCs according the ebXML Core Component Specification V1.85
And this specification says on page 11 that a Basic Core Component
represents a singular business concept with a unique business semantic
definition. A Basic Core Component will be constructed by using a Core
Component Type (CCT). And a CCT consists of one and only one Content
Component and one or more Supplementary Component which giving an essential
extra definition to the content component. 

The Content Component derives from a primitive type and will be used to
express the content of a CCT only. The Content Component itself have no
addtional meaning, like price. Therefore, your definition will be not agree
with the ebXML Core Component Specification. If you would like to use a
child-element for expressing the content it self, we have to define it as
following:

<amount currency="Y">
	<content>X</content>
</amount>

or

<amount>
	<content currency="Y">X</content>
</amount>

The ebXML CCS says that the Content Component will be derived from a
primitive type. And a primitive type is according the XML Schema Definition
a simple type definition. Therefore, it is not necessary to express the
content itself within an addtional child element. The following example will
be enough:

<amount currency="Y">X</amount>

And we have to define that all another possibilties for defining a Core
Component Type are prohibited.


Regards,
	
	Gunther

-----Original Message-----
From: Eduardo Gutentag [mailto:eduardo.gutentag@sun.com]
Sent: Mittwoch, 6. März 2002 18:45
To: Matthew Gertner
Cc: ubl-ndrsc@lists.oasis-open.org
Subject: Elements vs. Attributes - was Re: [ubl-ndrsc] Minutes for 6
March 2002 UBL NDR meeting


There's one thing that has just occurred to me, related to the discussion we
had this morning.

While we were talking, the mental model I had was that of an amount X,
expressed
as a currency Y; thus, I agreed to:
<amount currency="Y">X</amount>

This was clarified later as:
<amount currency="Y">
	<price>X</price>
</amount>

(for those who weren't in the discussion, please remember these are mental
models, or just examples, and do not intend to convey anything about element
tag names, etc.)

And looking now at it, I see that what was originally metainformation about
the value X has now migrated to be treated as metainformation of <amount>,
whereas it really should be metainformation of <price>; this may become
particularly important if <amount> were to become a more complex element
that contains more than price.

So, I'm not sure whether I'm nitpicking or not, I'd like to propose
that the actual example be
<amount>
	<price currency="y">X</price>
</amount>

Or am I totally wrong, and currency is actually metainfo on amount?

Matthew Gertner wrote:
> 
> The meeting was unchaired but we went forward anyway.
> 
> 1. Roll call
>     Bill Burcham                YES
>     Mavis Cournane      YES
>     Mark Crawford
>     Fabrice Desré
>     John Dumay
>     Matt Gertner                YES
>     Arofan Gregory      YES
>     Phil Griffin
>     Eduardo Gutentag    YES
>     Eve Maler
>     Dale McKay
>     Joe Moeller
>     Sue Probert
>     Ron Schuldt
>     Lisa Seaburg
>     Gunther Stuhec      YES
>     Paul Thorpe
> 
>     Quorum not reached as of x:15.  We proceeded informally.
> 
> 1'. Next week's call
>      No real chair. Needs to be discussed on the list.
> 
> 2. Acceptance of minutes of previous meetings
> 
>     Deferred.
> 
> 3. Adoption of agenda
> 
>     Short discussion of Bill's paper
>     Elements vs. attributes
> 
>     Adopted informally. Other items deferred.
> 
> 4. Action item review
> 
>     Deferred.
> 
> 5. Tag names
> 
> Short discussion of Bill's paper. Arguments made both for and against
> binding tag names in various ways to type names. Action to everyone to
read
> latest version of Bill's paper (will be posted to list) as input for final
> decision.
> 
> 6. Elements vs. attributes
> 
> Gunther has a new position paper. He argues that attributes could be used
> more than previously discussed. An example in currency code: making it an
> attribute eliminates the need for additional child elements, is more
> readable and is more efficient. Also proposed for:
> 
> * ID/IDREF
> * xml:lang (Arofan points out that there are difficulties enumerating this
> in practical cases, Gunther says that another attribute could be used for
> this purpose)
> 
> Gunther says he still has to outline treatment of empty elements. Arofan
> says that a lot of elements will be empty in this case, since all data
will
> be in attributes (which Eduardo doesn't like), but Gunther says this is
not
> so: the value itself will be in the element content.
> 
> Arofan points out one possible problem, which is when there is structure
in
> the value (e.g. exchange rate info). This might be solved by using mixed
> content, which no one likes. Eduardo says that this can be solved using
new
> attributes or sibling elements.
> 
> In general, consensus that elements be used for data and attributes for
> metadata, as per Gunther's paper. One caveat from Arofan (no example
given)
> is if metadata needs to have structure, in which case it has to use
> elements.
> 
> 7. Adjourn
>     Adjourned y:00.
> 
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>

-- 
Eduardo Gutentag               |         e-mail: eduardo.gutentag@Sun.COM
XML Technology Center          |         Phone:  (510) 986-3651 x73651
Sun Microsystems Inc.          |

----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>