There is a useful heuristic in critical thinking, that " if you can't cite a source for your claim, then your claim is probably bogus". Thanks to Wikipedia etiquette, this even has a succinct form: "Citation needed". (See http://xkcd.com/285/ for a great example.)
But it scares me that some people apparently operate under the reverse heuristic: "if you can cite a source for your claim, the claim is probably true". Lest you think I'm making this up, one fellow here has even made up an acronym for it: http://blogs.popart.com/2009/02/inbsiycnts-1/
This is madness.
Take the particular paper the above person is referring to, "An Economic Release Decision Model" by Grady, 1999. This paper is available nowhere online. It was presented at one of the first editions of a little-known SQE conference on "Applications of Software Measurement".
This paper is often cited as one of the recent sources for the "rising cost of defects" claim; people even quote specific "cost multipliers" depending on phase that come from a chart in this paper. Unfortunately, its extremely limited availability makes it hard to check what data it may refer to.
After a lot of digging around, one finds that the author was with Hewlett Packard, and that the study in question is probably the same one that was referred to in one chapter of his book on "Practical Software Metrics", titled "Dissecting Software Failures". If this is indeed the case, then our search for verifiable information has paid off, since the following is online: http://findarticles.com/p/articles/mi_m0HPJ/is_n2_v40/ai_7180006/
Turn to page 2, where we find the following quote:
> "The data for this example is taken from a detailed study of defect causes done at HP. In the study, defect data was gathered after testing began. [...] This study didn't accurately record the engineering times to fix the defects, so we will use average times summarized from several other studies to weight the defect origins.
I want you to read this several times until it sinks in, especially the bit in bold.
But it scares me that some people apparently operate under the reverse heuristic: "if you can cite a source for your claim, the claim is probably true". Lest you think I'm making this up, one fellow here has even made up an acronym for it: http://blogs.popart.com/2009/02/inbsiycnts-1/
This is madness.
Take the particular paper the above person is referring to, "An Economic Release Decision Model" by Grady, 1999. This paper is available nowhere online. It was presented at one of the first editions of a little-known SQE conference on "Applications of Software Measurement".
This paper is often cited as one of the recent sources for the "rising cost of defects" claim; people even quote specific "cost multipliers" depending on phase that come from a chart in this paper. Unfortunately, its extremely limited availability makes it hard to check what data it may refer to.
After a lot of digging around, one finds that the author was with Hewlett Packard, and that the study in question is probably the same one that was referred to in one chapter of his book on "Practical Software Metrics", titled "Dissecting Software Failures". If this is indeed the case, then our search for verifiable information has paid off, since the following is online: http://findarticles.com/p/articles/mi_m0HPJ/is_n2_v40/ai_7180006/
Turn to page 2, where we find the following quote:
> "The data for this example is taken from a detailed study of defect causes done at HP. In the study, defect data was gathered after testing began. [...] This study didn't accurately record the engineering times to fix the defects, so we will use average times summarized from several other studies to weight the defect origins.
I want you to read this several times until it sinks in, especially the bit in bold.
