xbrl us gaap contextRef standard?

its in xbrl 2.1 specification
http://www.xbrl.org/Specification/XBRL-RECOMMENDATION-2003-12-31+Corrected-Errata-2005-04-25.htm#_4.6.1

      <context id="BalanceAsOf_20120531">        <entity>          <identifier scheme="http://www.sec.gov/CIK">0000866729</identifier>        </entity>        <period>          <instant>2012-05-31</instant>        </period>      </context>      <context id="BalanceAsOf_20120531_AccumulatedOtherComprehensiveIncomeMember">        <entity>          <identifier scheme="http://www.sec.gov/CIK">0000866729</identifier>          <segment>            <xbrldi:explicitMember dimension="us-gaap:StatementEquityComponentsAxis">us-gaap:AccumulatedOtherComprehensiveIncomeMember</xbrldi:explicitMember>          </segment>        </entity>        <period>          <instant>2012-05-31</instant>        </period>      </context>      <unit id="USD">        <measure>iso4217:USD</measure>      </unit>    <us-gaap:AssetsCurrent contextRef="BalanceAsOf_20120531" unitRef="USD" decimals="-5">929900000.0</us-gaap:AssetsCurrent>    <us-gaap:Assets contextRef="BalanceAsOf_20120531" unitRef="USD" decimals="-5">1670300000.0</us-gaap:Assets>    <us-gaap:StockholdersEquity contextRef="BalanceAsOf_20120531_AccumulatedOtherComprehensiveIncomeMember" unitRef="USD" decimals="-5">-74200000.0</us-gaap:StockholdersEquity>

for example; consider these two context; ids needs to be unique with respect to their content(With respect to period entity; and entity with its sub elements; identifier and segment if present)
we will use its id to map financial facts. Here in example AssetsCurrent refers to unit--> whose id is USD and context information--> whose id is BalanceAsOf_20120531; and detail of this this context id can be found in context element.here everything should be unique other wise it will create duplicate facts/contexts.You can see here that; both context has same period but so why can't their ids? because one context has addition detail segment information.
(i am not going in detail here; if you want to know; how it is detecting duplicate/error consider example no 25 and 26 from the link.)
in other way you can also view facts/values with three axis as CIK period and xmlname;and this uniqueness makes them comparing with other company's facts with these axis.

Don't rely on the value of the contextRef attribute. The EDGAR FIler Manual gives advice to filers about creating mnemonic values, but all the software used to create filings uses their own ways of creating these values.

All the contextRef is good for is locating the context for a particular fact. It is the parts of the context that organize the data.