This is a policy that a standards body (such as ISO, ITU, W3C, ANSI, IETF, and JEDEC) may choose to have with respect to patents on any technology that becomes incorporated into a standard they come up with. 'RAND' stands for 'Reasonable And Non-Discriminatory', and refers to the means by which patents under this policy must be licensed before they may be included in a standard. It means that patent holders whose patents become incorporated into a standard must do three things:

  1. Disclose any and all patents to the standards body, if they are part of the consortium making the standard,
  2. Give reasonable royalty fees on use of the patent when it is used as part of an implementation of the standard. What is meant by 'reasonable' is a policy set by the standards body, and
  3. License the patent to anyone and everyone who wishes to use it to implement the standard. The patent holder must not discriminate against anyone wishing to implement the standard who pays the license fees.

Some standards bodies add further details to this but that is the general spirit. This is thought by some standards bodies, especially those concerned with hardware standards such as JEDEC and ITU, to be a reasonable compromise between creating an open standard and preserving the integrity of the patent system. With this, patent holders whose patents have become incorporated into a standard are not allowed to charge extortionate rates for use of the patent to implement the standard, and they are also not allowed to refuse to license the patent to anyone, including their competitors, for purposes of implementing the standard.

Other standards bodies on the other hand, reject this policy and insist that either their standards be totally free of patents, or that if any patents exist they must be be licensed to everyone royalty-free (this is the so-called RF licensing policy). This is a policy that preserves the integrity of the open standard, but subsequent patent evasion may result in a standard of lower quality. The ISO has this policy, as does the W3C, though they presently ponder the pros and cons of RAND licensing.

This term first began entering public consciousness when the W3C, the organization that makes the standards that lie behind the World-Wide Web, began mulling the use of RAND licensing for technologies used by official web standards in the middle of 2001. Massive public outcry has caused the W3C to tread more lightly in this area, and as of this writing the W3C's patent policy working group has not yet decided on whether to use RAND licensing.

Update October 7, 2002: It appears that the W3C has, after a year of debate and discussion on the issue, finally released an official statement that it will not be using RAND licensing for any future standards it promulgates. Amen!