Attributes vs. text content (Was Re: RFC: Attributes and XML- RPC)

Mark Birbeck Mark.Birbeck at
Tue Sep 21 23:15:51 BST 1999

joubin wrote:
> If, as a matter of convention, one (only) uses attributes to 
> convey element metainfo (i.e. information _about_ the element)
> then, in fact, you would have a much more coherent document
> (cum protocol) with clear distinction between information and
> meta-information.
> Consider:
> <?xml version="1.0"?>
> <methodCall type="asynchronous">
>    <methodName><methodName>
>      // .. etc.
> </methodCall>
>   v.s.
> <?xml version="1.0"?>
> <methodCall>
>    </asynchronous>
>    <methodName><methodName>
>      // .. etc.
> </methodCall>
> The latter mixes info and metainfo.  Not a good strategy, 
> IMO, in long haul; specially if one is aiming for simplicity,
> and clarity.

But why 'methodName' is also not 'information _about_ the element' is
not obvious. I think you'd be hard pushed to justify any particular
representation over any other - other than those that feel intuitively
right in a given situation.

Having said that, one rule I have adopted is to use attributes in
situations where I want to know something about a node without having to
read its children. This was as a consequence of developing some code to
return nodes from a server but rather than sending back all child nodes,
I simply returned XPath references to the children (a bit like the
fragments spec.). This meant that the client need only request
subsequent nodes if it wanted to. In that situation, you get better
performance by having some attribute information in the node itself,
which might help determine whether you even need to read the children.



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list