1. Introduction
This specification supplements RFC 3280prop [PKIX-1], which profiles
X.509 [X.509] certificates and certificate revocation lists (CRLs)
for use in the Internet.
The basic function of a certificate is to bind a public key to the
identity of an entity (the subject). From a strictly technical
viewpoint, this goal could be achieved by signing the identity of the
subject together with its public key. However, the art of Public-Key
Infrastructure (PKI) has developed certificates far beyond this
functionality in order to meet the needs of modern global networks
and heterogeneous IT structures.
Certificate users must be able to determine certificate policies,
appropriate key usage, assurance level, and name form constraints.
Before a relying party can make an informed decision whether a
particular certificate is trustworthy and relevant for its intended
usage, a certificate may be examined from several different
perspectives.
Systematic processing is necessary to determine whether a particular
certificate meets the predefined prerequisites for an intended usage.
Much of the information contained in certificates is appropriate and
effective for machine processing; however, this information is not
suitable for a corresponding human trust and recognition process.
Humans prefer to structure information into categories and symbols.
Most humans associate complex structures of reality with easily
recognizable logotypes and marks. Humans tend to trust things that
they recognize from previous experiences. Humans may examine
information to confirm their initial reaction. Very few consumers
actually read all terms and conditions they agree to in accepting a
service, rather they commonly act on trust derived from previous
experience and recognition.
A big part of this process is branding. Service providers and
product vendors invest a lot of money and resources into creating a
strong relation between positive user experiences and easily
recognizable trademarks, servicemarks, and logotypes.
Branding is also pervasive in identification instruments, including
identification cards, passports, driver's licenses, credit cards,
gasoline cards, and loyalty cards. Identification instruments are
intended to identify the holder as a particular person or as a member
of the community. The community may represent the subscribers of a
service or any other group. Identification instruments, in physical
form, commonly use logotypes and symbols, solely to enhance human
recognition and trust in the identification instrument itself. They
may also include a registered trademark to allow legal recourse for
unauthorized duplication.
Since certificates play an equivalent role in electronic exchanges,
we examine the inclusion of logotypes in certificates. We consider
certificate-based identification and certificate selection.
1.1. Certificate-based Identification
The need for human recognition depends on the manner in which
certificates are used and whether certificates need to be visible to
human users. If certificates are to be used in open environments and
in applications that bring the user in conscious contact with the
result of a certificate-based identification process, then human
recognition is highly relevant, and may be a necessity.
Examples of such applications include:
- Web server identification where a user identifies the owner of
the web site.
- Peer e-mail exchange in B2B, B2C, and private communications.
- Exchange of medical records, and system for medical
prescriptions.
- Unstructured e-business applications (i.e., non-EDI
applications).
- Wireless client authenticating to a service provider.
Most applications provide the human user with an opportunity to view
the results of a successful certificate-based identification process.
When the user takes the steps necessary to view these results, the
user is presented with a view of a certificate. This solution has
two major problems. First, the function to view a certificate is
often rather hard to find for a non-technical user. Second, the
presentation of the certificate is too technical and is not user
friendly. It contains no graphic symbols or logotypes to enhance
human recognition.
Many investigations have shown that users of today's applications do
not take the steps necessary to view certificates. This could be due
to poor user interfaces. Further, many applications are structured
to hide certificates from users. The application designers do not
want to expose certificates to users at all.
One situation where software applications must expose human users to
certificates is when the user must select a single certificate from a
portfolio of certificates. In some cases, the software application
can use information within the certificates to filter the list for
suitability; however, the user must be queried if more than one
certificate is suitable. The human user must select one of them.
This situation is comparable to a person selecting a suitable plastic
card from his wallet. In this situation, substantial assistance is
provided by card color, location, and branding.
In order to provide similar support for certificate selection, the
users need tools to easily recognize and distinguish certificates.
Introduction of logotypes into certificates provides the necessary
graphic.
1.3. Combination of Verification Techniques
The use of logotypes will, in many cases, affect the users decision
to trust and use a certificate. It is therefore important that there
be a distinct and clear architectural and functional distinction
between the processes and objectives of the automated certificate
verification and human recognition.
Since logotypes are only aimed for human interpretation and contain
data that is inappropriate for computer based verification schemes,
the logotype extension MUST NOT be an active component in automated
certification path validation.
Automated certification path verification determines whether the
end-entity certificate can be verified according to defined policy.
The algorithm for this verification is specified in RFC 3280prop
[PKIX-1].
The automated processing provides assurance that the certificate is
valid. It does not indicate whether the subject is entitled to any
particular information, or whether the subject ought to be trusted to
perform a particular service. These are access control decisions.
Automatic processing will make some access control decisions, but
others, depending on the application context, involve the human user.
In some situations, where automated procedures have failed to
establish the suitability of the certificate to the task, the human
user is the final arbitrator of the post certificate verification
access control decisions. In the end, the human will decide whether
or not to accept an executable email attachment, to release personal
information, or follow the instructions displayed by a web browser.
This decision will often be based on recognition and previous
experience.
The distinction between systematic processing and human processing is
rather straightforward. They can be complementary. While the
systematic process is focused on certification path construction and
verification, the human acceptance process is focused on recognition
and related previous experience.
There are some situations where systematic processing and human
processing interfere with each other. These issues are discussed in
the Security Considerations section.
1.4. Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in BCP 14, RFC 2119
[STDWORDS].
