- Error 33
The term "error 33" is
jargon for the failure due to predicating one research project on the success of another, or alternatively for allowing one's own research into the critical path of another project. [http://onlinedictionary.datasegment.com/word/error%2033 "Datasegment Online Dictionary] ]Eric S Raymond 's [http://catb.org/esr/jargon/html/E/error-33.html Online Jargon File] ]Origin
"Error 33" was originally coined by Dr. George Pake, the first director of
Xerox Parc .Usage
As alumni of
Xerox Parc have spread to other companies throughoutSilicon Valley , this phrase has been broadened in scope to general research and engineering projects as well as research efforts.The advice to leaders of any project with a substantial development risk is to make sure that as many dependencies as possible use established technology, engineering, or research results.
Xerox alumnus
Alan Kay has elaborated this into two "sets of theories" that contradict each other.Kay, Allan (2004). [http://www.vpri.org/pdf/draper_RN-2004-001.pdf "The Power Of The Context"] (Draper Prize Lecture, Feb 24, 2004 "VPRI Research Note RN-2004-001").] In one direction, avoidance of error 33 suggests that teams avoid building their own tools - in his example, software languages and operating systems - because it can sink an incredible amount of time that doesn't move the primary project objective forward. In the opposite direction, error 33 is avoided by creating exactly the right tools for the job at hand, and avoiding reliance upon external vendors, teams, or projects for success. Alan cites the development ofEthernet as an example of avoiding error 33: Metcalfe, Boggs, Lampson, and Thacker avoided designing a protocol that depended upon an error-free network, and instead depended upon retry and recovery in a way that eventually would send messages perfectly, even under extreme conditions.In the same paper, Alan cites the Bravo text editor (precursor to
Microsoft Word ) as a project that deliveredWYSIWYG printing, but avoiding dependence on a reliable platform by incorporating the ability to replay right up to the point of a crash.Another Xerox alumnus,
Neil Gunther , suggests thatWeb 2.0 sites that depend heavily for their commercial success uponAmazon.com 's (at the time of writing in 2008) barely-out-of-researchElastic Compute Cloud (EC2) technology is just such an error 33.Gunther, Neil (2008). [http://perfdynamics.blogspot.com/2008/02/web-20-meets-error-33.html "Web 2.0 Meets Error 33"] .] A February 2008 failure of Amazon's EC2 systems did impact many websites.References
Wikimedia Foundation. 2010.
