Subject: Re: [xsl] Tree Comparing Algorithm From: "Michael Kay mike@xxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> Date: Tue, 4 Feb 2020 09:03:24 -0000 |
> Even if the probability of two objects to have the same hash code is low, we should take this into account -- for example, we may choose to calculate a pair of hash codes for each object, using two independent hashing algorithms. > This reminds me of an incident during my time with ICL when we were designing a hardware accelerator containing search logic, and it did precisely this. The software people (like me) were horrified, but the hardware designers were entirely comfortable with doing statistical calculations on the probability of failures, and so long as the expected number of failures was significantly less than one in the design lifetime of the device, they were happy to go with it. Michael Kay Saxonica
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: [xsl] Tree Comparing Algorithm, Dimitre Novatchev dn | Thread | [xsl] xsl:mode and priorities quest, Christophe Marchand |
Re: [xsl] Tree Comparing Algorithm, Dimitre Novatchev dn | Date | Re: [xsl] Tree Comparing Algorithm, Martin Honnen martin |
Month |