zlacker

DeWitt clauses have spread outside of databases too, unfortunately. John Regehr got a nastygram from Coverity/Synopsys when he tried to post static analyzer benchmarks (not even his own; the benchmarks were done by researchers at the Toyota InfoTechnology Center):

https://blog.regehr.org/archives/1217

This has effectively dissuaded me from trying to perform benchmarks of Coverity and other commercial tools using LAVA [1], so I can attest to its chilling effect.

[1] https://seclab.ccs.neu.edu/static/publications/sp2016lava.pd...

>>moyix+(OP)
That sounds like something SciHub would be very useful for --- "the benchmarks they don't want you to know."

>>userbi+6i
Sure, someone could post a paper like this anonymously. But anonymous papers don't benefit from all the incentives of the standard academic publishing system.

As a fun aside, here's an example of one of the few fully anonymous papers I know of:

http://census2012.sourceforge.net/paper.html

>>moyix+(OP)
The EDA (Electronic Design Automation) industry is full of such clauses. Every software in that industry is selling based on its quality of results or faster processing time. Every tool claims 10x improvements but there is no way to verify it. To be fair to them, understandably designs are different and may yield different quality of results.