The position attribute has two other possible values: replace and attributes. Using replace causes the selected element to be replaced with the content of the instruction element. Consequently, if you don't have any content, the selected element is simply removed. The preceding list or form view would cause the email field to be removed:

<field name="email" position="replace" />

Removing fields can cause other inheriting views to break and several similarly ugly side effects, so avoid that if possible. If you really need to remove fields, do so in a view that comes late in order of evaluation (refer to the next).

Using attributes has a very different semantics from all of the preceding. The processor then expects the element to contain the attribute elements with a name attribute. Those elements will then be used to set attributes on the selected element. If you want to heed the earlier warning, you'd better set the invisible attribute to 1 for the email field:

<field name="email" position="attributes">
    <attribute name="invisible">1</attribute>
</field>

An attribute node can have the add and remove attributes, which then in turn should contain the value to be removed or added to the space separated list the attribute value is expected to be. This is very useful for the class attribute, where you'd add a class (instead of overwriting the whole attribute) by saying the following:

<field name="email" position="attributes">
    <attribute name="class" add="oe_inline" />
</field>