itemids must be valid URLs, however the spec example microdata vocabularies give no guidance to valid global identifiers. The example UIDs listed in the referred RFCs also dont appear to be one of the IANA URI schemes, eg:

The type can include the type parameter "TYPE" to specify the format
   of the identifier. The TYPE parameter value should be an IANA
   registered identifier format. The value can also be a non-standard
   format.
Type example: UID:19950401-080045-40000F192713-0052
http://tools.ietf.org/html/rfc2426#section-3.6.7

Property Parameters: Non-standard property parameters can be
   specified on this property.
Description: The UID itself MUST be a globally unique identifier. The
   generator of the identifier MUST guarantee that the identifier is
   unique. There are several algorithms that can be used to accomplish
   this. The identifier is RECOMMENDED to be the identical syntax to the
   [RFC 822] addr-spec.
Example: The following is an example of this property:
     UID:19960401T080045Z-4000F192713-0052@host1.com
http://tools.ietf.org/html/rfc2445#section-4.8.4.7

As itemid is entirely up to the vocabulary itd help ppl to understand itemid usage if these specs were clear about what global identifiers are valid, ideally with an example using a global identifier. This would also help vocabulary writers (schema.org) to clearly define global identifiers  http://schema.org/Book only defines ISBN as an itemprop currently.

Ref:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=13452
http://microformats.org/wiki/vcard-errata#3.6.7_UID_Type_Definition
http://www.iana.org/assignments/uri-schemes.html

So there's two ways we can go with this:

 - make itemid="" any arbitrary string, with vocabularies defining the precise requirements on uniqueness and syntax

 - make the vCard and vEvent vocabs not use itemid="".

I think I'm leaning towards the second. The problem with the first is that it might encourage people to use itemid="" for things like ISBN number, which would lead them to difficulties if they decide their items aren't exactly one publication, so for example a hard cover edition and paperback edition would have the same item.

EDITOR'S RESPONSE: This is an Editor's Response to your comment. If you are satisfied with this response, please change the state of this bug to CLOSED. If you have additional information and would like the editor to reconsider, please reopen this bug. If you would like to escalate the issue to the full HTML Working Group, please add the TrackerRequest keyword to this bug, and suggest title and text for the tracker issue; or you may create a tracker issue yourself, if you are able to do so. For more details, see this document:
   http://dev.w3.org/html5/decision-policy/decision-policy.html

Status: Accepted
Change Description: see diff given below
Rationale: Concurred with reporter's comments.

Checked in as WHATWG revision r6746.
Check-in comment: Change the vCard and vEvent vocabs to not use itemid='' since their UIDs aren't URLs.
http://html5.org/tools/web-apps-tracker?from=6745&to=6746

So now there are no known vocabularies using itemid... Wouldn't it be just as good to drop the attribute entirely and let vocabularies that need something like this define a property themselves? Schema.org has the url property which is likely to often overlap with itemid, should they ever get around to defining what it means (http://www.w3.org/2011/webschema/track/issues/6).