Philip Emeagwali myths
This page briefly addresses some specific claims appearing on Emeagwali's self-promotional website and elsewhere. For a more in-depth discussion of Emeagwali's "outright lies, half-truths, and numerous unsubstantiated claims," see the article Self-Promotion and Self-Authentication: "Father of the Internet" by Chioma Ezeilo.
MYTH #1:
Emeagwali made the "world's fastest computation" in 1989
Each year since 1987, the judges for the annual Gordon Bell competition have given out multiple prizes for supercomputing applications, usually including one forperformance (fastest speed), and one for price/performance (best speed/cost ratio; specifically, "price-performance ratio as measured in megaflop/s per dollar on a genuine application").
Despite what you may have read on the Internet, the 1989 prize for the fastest performance was awarded not to Emeagwali, but to another entrant:
In the performance category, we awarded the prize to a team from Mobil Research and Development and Thinking Machines Corp. [...] Their solution of a seismic data-processing problem ran at almost 6 Gflops on a CM-2 Connection Machine.
"Special Report: 1989 Gordon Bell Prize," IEEE Software, May 1990, p. 101
The winning speed of 6 Gflops was almost double that of Emeagwali's entry (3.1 Gflops, a speed that Emeagwali has repeatedly misrepresented as the "world's fastest"). The Mobil/TMC team achieved the best price-performance ratio too, but since no entry was allowed more than one prize, the price/performance award passed to Emeagwali (who also used a CM-2 Connection Machine) even though his score in that category was about 20% worse than the leading score:
We awarded the price/performance prize to Philip Emeagwali [...] His model ran at a price/performance of slightly less than 400 Mflops per $1 million.While the Mobil/TMC team achieved almost 500 Mflops per $1 million, we decided to award only one prize per entry.
"Special Report: 1989 Gordon Bell Prize," IEEE Software, May 1990, p. 101
See the yearly results for the performance and price/performance categories. In no year did Emeagwali achieve what he claims he did.
MYTH #2:
Emeagwali won computing's equivalent of the Nobel Prize
In spite of efforts by Emeagwali and his admirers to hype up his $1000 Gordon Bell award as the "Nobel Prize" of computing, the rest of the computing world continues to associate that lofty label with the Turing Award, which Emeagwali has never won. The Gordon Bell prize, however, is just one of many other annual computing awards, respectable but not at all comparable to either the Nobel Prize or the Turing Award in prestige or prize money. Nor are the selection criteria analogous: Emeagwali was awarded for his performance in an annual competition, not in recognition of lasting contributions to his field.
MYTH #3:
Emeagwali is a "Father of the Internet"
1. Emeagwali's paternity claims to the Internet are judged and found baseless. Emeagwali has not shown any evidence that he was involved with ARPA or any other research organization or company connected with the genesis of the Internet, nor did he express his ideas in technical journals or any other channel through which he could have influenced the development of the Internet during its formative stages.
2. Is there a father of the Internet? A discussion of the main contenders: JCR Licklider, Bob Taylor, Paul Baran, Donald Davies, and Lawrence Roberts.
MYTH #4:
The "Connection Machine" was invented by Emeagwali
A few Web sources expand the fallacy that Emeagwali created the fastest computer program, insisting that he was also responsible for the massively parallel supercomputer on which the program ran. However, the 65,000-processor "Connection Machine" that allowed such speedy computations was actually the brainchild of Danny Hillis and was built by Thinking Machines Corporation, the company Hillis co-founded. The fact is extensively documented on the web and is so well established that I won't devote more attention to it.
MYTH #5:
Emeagwali has lots of patents or patent applications
The number of Emeagwali's patents or patent applications is sometimes claimed to be as high as 30. However, a search of the USPTO and esp@cenet patent databases (as of September 2003) reveals the true number to be zero. The U.S. Patent and Trademark Office, which has a fully searchable online database for all U.S. patents since 1976 as well as all current applications, simply has nothing on file for Emeagwali.
MYTH #6:
"Dr." Emeagwali?
Emeagwali is frequently referred to on his own and other websites as "Dr. Philip Emeagwali" or simply "Dr. Emeagwali." Though he did in fact enroll in a PhD program at the University of Michigan, he did not get the degree after failing his qualification examinations twice and having his thesis rejected by a committee of faculty members. Emeagwali subsequently sued the university, alleging civil rights violations and racial discrimination. His case was dismissed without trial. When Emeagwali appealed to a higher court, a three-judge panel rejected his discrimination claims unanimously.
Suite 817C
Atlanta, GA, USA.
30303
Cell: (+1) 404-573-9697
Blog: www.farooqkperogi.blogspot.com
"The nice thing about pessimism is that you are constantly being either proven right or pleasantly surprised." G. F. Will
--
You received this message because you are subscribed to the "USA-Africa Dialogue Series" moderated by Toyin Falola, University of Texas at Austin.
For current archives, visit http://groups.google.com/group/USAAfricaDialogue
For previous archives, visit http://www.utexas.edu/conferences/africa/ads/index.html
To post to this group, send an email to USAAfricaDialogue@googlegroups.com
To unsubscribe from this group, send an email to USAAfricaDialogue-
unsubscribe@googlegroups.com
No comments:
Post a Comment