Re: [xml-dev] Designing an experiment to gather evidence on approaches t

[Richard Salz <rsalz@us.ibm.com>]

> Anecdote either way.  Scientific (as opposed to legal) evidence requires 
a
> planned test set ...

Don't tell Ptolemy, Copernicus, et al.

> In the nature of things, such gold-standard evidence testing is not
> always possible.

Indeed, I'd say such a Platonic ideal is a fool's errand in the real world 
let alone one with so many variables as "IT projects."  All we have in 
this situation is some data points of experience to help guide our 
architectural decisions. 

        /r$

--
STSM, WebSphere Appliance Architect
https://www.ibm.com/developerworks/mydeveloperworks/blogs/soma/