Yep. The rest of it is, well, xml, and it's not going to be as clean looking as json , but why do I need to see the word json at the start and stop of every element? It could at least be abbreviated to just 'j'?

On the plus side I bet the output gzip's up rather nicely in transit assuming you're HTTP GET'ting this data....

Replace `xmlns:json="http://...` with `xmlns:j="http://...`.

Actually, just remove the colon and the prefix entirely. It becomes surprisingly readable:

    <object xsi:schemaLocation="http://www.datapower.com/schemas/json jsonx.xsd"
        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
        xmlns="http://www.ibm.com/xmlns/prod/2009/jsonx">
        <string name="name">John Smith</string>
        <object name="address">
            <string name="streetAddress">21 2nd Street</string>
            <string name="city">New York</string>
            <string name="state">NY</string>
            <number name="postalCode">10021</number>
        </object>
        <array name="phoneNumbers">
            <string>212 555-1111</string>
            <string>212 555-2222</string>
        </array>
        <null name="additionalInfo" />
        <boolean name="remote">false</boolean>
        <number name="height">62.4</number>
        <string name="ficoScore"> > 640</string>
    </object>

> On the plus side I bet the output gzip's up rather nicely in transit assuming you're HTTP GET'ting this data....

The XML is 901 bytes, gzip yields 401 (44.5%)

Of course the original JSON is 303 bytes and JSON is already pretty compressible, to 215 bytes (71%) in this case.

> but why do I need to see the word json at the start and stop of every element?

Because of what I just mentioned. If they had set the root namespace directly you wouldn't have to declare the namespace prefix on every element.

I don't know why you were downvoted. Setting the root namespace provides exactly the effect of dropping "json:" from everything.