Which XML structure makes more sense?

Question

Without going into too much detail we are looking to use XML as meta-data to describe constraints on properties (This is a cutdown example and XSD did not support our proposed complex model), there are two options being considered, which of the following XML strucutures makes better sense?

Option 1)

<?xml version="1.0" encoding="us-ascii"?>
<Properties xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"&gt;
   <Property type="string">
      <name>quanitity</name>
      <contraints>
         <contraint type="isRequired">
            <value>true</value>
         </contraint>
         <contraint type="regex">
            <value>^[0-9]$</value>
         </contraint>
         <contraint type="regex">
            <value>^[a-zA-Z]$</value>
         </contraint>
      </contraints>
   </Property>
</Properties>

Option 2)

<?xml version="1.0" encoding="us-ascii"?>
<Properties xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"&gt;
   <Property type="string">
      <name>quantity</name>
      <IsRequired>true</IsRequired>
      <Regex>^[0-9]$</Regex>
      <Regex>^[a-zA-Z]$</Regex>
   </Property>
</Properties>

Answer 1

For flexibility, I would go with #1. This will allow you to add many different types of constraints and custom rules.

Answer 2

Personally, I'd go with the first one, I suspect it'd be easier to write code to parse and do error checking in. Plus it makes more sense semantically to my brain.

That said, I think you'll find it's 'personal choice', either is fine.

Answer 3

I would probably go for option #2 as the XML job does a better job of describing the data (which is after all what XML is supposed to do), and is therefore easier to read and less verbose. Option #1 is a little close to having <name> and <value> tags for my liking.

Answer 4

It depends on how you want to parse/process the XML and if you want to be able to validate it.

From a validation point of view (if you are considering automatic validation via an XSD at compile time) Option 2 is the better choice, as the content of the tags will have a certain type (e.g., a regular expression), while the <value> tag could not be checked for correct content.

Also, I would probably go with Option 2 concerning the parsing and processing if you are using a SAX parser, as you could switch states based on the tags.