ISO 29119: The Redundancy in the Documentation Document
When it comes to ISO 29119, the new international standard for software testing, there is a great deal of secrecy around it. The current plan is to create a total of five documents (three are already published) at a suggested price of two hundred dollars each. For a thousand dollars, what do you get?
Most people, including those signing the petition against it, don’t actually know what the documents say, as the price puts the documents beyond the realm of the typical tester. I have the information published so far, but the copyright disclaimer means I can’t share it.
At least, I can’t share all of it. Reading the text carefully, I have come to the conclusion that small snippets, in order to comment and review, are fair use.
29119 part 3, the “documentation document,” starts with a series of definitions. Here are a few examples (retaining the original capitalization from the text):
4.17
Test Data Readiness Report
document describing the status of each test data requirement
4.21
test environment readiness report
document that describes the fulfillment of each test environment requirement
4.22
Test Environment Requirements
description of the necessary properties of the test environment
4.31
test status report
report that provides information about the status of the testing that is being performed in a specific reporting period
Section 5 is basically a collection of documents. Each has the same things in common: an overview, a unique identifier, the issuing organization, the approval authority, change history, and then the thing that document is actually supposed to be. And yes, the document descriptions continue the proud tradition of defining themselves. For example, once you actually get to the thing the test process document is—after introduction, scope, references, the glossary, and objectives—you hit part 5.2.4.2, “test process,” which “identifies the test process that the organization will follow.”
Reading this, I was amazed at the redundancy of the document. Instead of listing an identifier, title, overview, scope, history, and change log for each document, the standard could say that every document needs these things and that they typically come for free in modern version control systems. Then each document could be described at the single bullet point level, and the entire document would become ten pages long instead of 127.
Oh. Perhaps that is why they didn’t do it.
