Talk:Software regression

	This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing
Low	This article has been rated as Low-importance on the project's importance scale.
	This article is supported by WikiProject Software.

WP:NOR violation?

This article seems to me to be largely original research. That the term "regression" has "evolved" past meaning, simply, reintroduced bugs is an unsourced claim that is in contradiction with the article on regression testing, which, itself, discusses regressions, at length, as reintroduced bugs. TerraFrost 15:32, 18 January 2009 (UTC)

I disagree. The regression testing article's first section describes the activity as one of uncovering "new errors...in existing functionality" as well as "faults [which] have re-emerged". That is, not just "reintroduced bugs" alone as would be required for a contraction. That said, I think the claimed shift in emphasis is being overstated, at least in the BSD world. Consider the (modern) NetBSD description that "Regression tests are used to exercise a particular bit of the system to check that it works as expected, and to make sure that old bugs are not reintroduced"[1]. NetBSD added systematic regression testing in 1993[2] and originally described the regression tests as generic ("A regression test is run by a makefile in a test directory"[3]). The "regression system originally came from NetBSD"[4]. I also found a reference to compiler regression tests from 1979 as "tests that compare the performance of a new compiler against the performance of an older one"[5] and "checks that this compiler produces the same output as the earlier one"[6]. I found no descriptions that explicitly refer to re-emerging bugs, probably because at that time (a) bugs and defects were probably not systematically tracked and (b) were expensive to automate. Instead, coarse metrics such as output size or time taken could be recorded. DLeonard (talk) 01:39, 23 June 2011 (UTC)[reply]

This seems resolved to me since no one has had anything to add in 8+ years and the argument seems sound. Kenyoni (talk) 01:52, 22 December 2019 (UTC)[reply]

Not really. Just because no one has edited this corner of the article can we assume that the article does not still contain original research. Walter Görlitz (talk) 21:23, 31 December 2019 (UTC)[reply]

Opening sentences too verbose?

The description on the disambiguation page for "regression" puts it clearly and succinctly: "the appearance of a bug which was absent in a previous revision." --Keith111 (talk) 05:24, 12 July 2009 (UTC)[reply]

On second thought, it's not a regression if the new bug is with a new feature (that didn't work correctly to begin with), so maybe the wordiness is warranted. --Keith111 (talk) 23:33, 13 July 2009 (UTC)[reply]

No, failure on switch to DST is not a regression.

A failure on switch to DST is only a regression if the software previously handled the switch.

A regression is purely an inadvertent loss of functionality. Some feature of the program which used to work suddenly breaks.

216.31.219.19 (talk) 23:23, 17 March 2014 (UTC)[reply]

Suggestion to add terminological explanation

... The word "regression" (in the term "software regression") is short for "bug regression", a bug that was reintroduced by a subsequent change... --Krauss (talk) 00:17, 5 October 2014 (UTC)[reply]

First, you inserted before the opening sentence and MOS:LEAD, specifically WP:LEADSENTENCE states "If possible, the page title should be the subject of the first sentence" and so your insertion was inappropriate.

Second, it was an attempt at a definition and WP:NOTDICDEF states that we should avoid such.

Third, we call your addition original research unless it can be supported with a verifiableor reliable source. I have reviewed the sections in Testing Computer Software: Second Edition (Kaner, Falk and Ngyuyen) and it states that regression testing is testing what has been tested and say nothing about a regression bug or that software regression is at all associated with the idea. (pages 295 and 357). Boris Beizer's Black-Box Testing (pp. 235) refers to it as a "rerun of a major subset of the original test suite" and references Burke's 1988 work. No mention of it leading to a bug regression. I have read other books and none mention "bug regression" but I have heard of "regression bugs" and simply "regressions".

So let's see what StickyMinds has to say. 12 hits for "bug regression" 27 of "regression bug". Google has 75,000 for "regression bug" and 50,100 for "bug regression". Walter Görlitz (talk) 03:59, 5 October 2014 (UTC)[reply]

Should I edit this?

Greetings! Let me start off by saying that I, many years ago, was in a library which was hard to get into, and I decided to use my time there to research primary and secondary sources for a topic on wikipedia, and to provide sources, or discredit some unsourced assertions on a page which seemed like had no activity for a long time. I believe I clicked on the "give me a random page" and then proceeded to spend an entire afternoon researching and providing wikipedia with the information I had found. Unfortunately, there was a "revert all edits" button, and someone reverted all changes I made to the page- even grammatical and spelling changes, which I separated from the others in case people wanted more sources for certain content. I told myself that I would never ever edit a wikipedia page again or waste any time on this, however I saw a sign posted in another library, told another person about my bad experience, and he convinced me to post again.

I can fix this post, and provide you with sources from MDN, LKML, and a few other places. However, how do I do it so that one disgruntled individual does not revert my changes? I'm not willing to waste my time again. I would like to contribute to scientific, engineering, and mathematical wikipedia posts since I have benefited from them for many years, but I cannot justify a huge waste of time to get reverted with a single click. I am fine submitting my posts and research here to have it evaluated by others who can select content, but a "one click reverts all" is not for me.

Anyway, this is a post which would not take me long to 'fix'. I'm not going to waste time reading 'howto' deal with the community//posting content and the like, because I did it years ago and it was a waste of my time. A quick explanation is sufficient. I can help with a number of other software engineering pages.

I would absolutely prefer to have my content and changes evaluated by another individual who can decide if they should be made or not. I'm fine posting in the 'Talk' section since editing the main page did not work out for me in the past, but I'm not willing to get hit with the 'revert all' again. Would it be acceptable to post changes here, and if they are 'approved', would someone else change the main page and monitor it? No worries about having an 'open book' on my end- I am quite used to having my work and research evaluated before it is published, but the 'evaluation' part before 'deletion' is quite important to me.

Hotchkiss87 (talk) 02:47, 22 October 2016 (UTC)[reply]

As long as your edits supply references rather than remove refs, I will not revert. I definitely won't revert spelling changes, as long as it's not changing between International (British) and Amercian spellings. Walter Görlitz (talk) 22:19, 23 October 2016 (UTC)[reply]

Maintenance template

I have removed the maintenance template for the following reasons:

Original research seams to have been resolved 8 years ago.
I added several citations and believe the material to be adequately cited.

Article still needs improvement in other ways.Kenyoni (talk) 02:01, 22 December 2019 (UTC)[reply]

I added a new tag as I don't believe the study is particularly notable and I removed two blogs. More templates should probably be applied. Walter Görlitz (talk) 21:24, 31 December 2019 (UTC)[reply]

Did some cleanup/extensions

I had a bit of time to spare, so I decided to do some cleanup with this article, as the topic falls well within my research area. Here are the major changes I applied.

In the introductory paragraphs, I clarified that software regressions can occur as a result of both changes to the software under test and the environment in which it is running.
I added a citation to the list of software regression types
I added two new subsections – one on Prevention and Detection, and another on Localization. I tried to write these as generally as I could, to avoid violating WP:OR. In cases where I mention specific techniques proposed in either research or industry, I included the proper citation to the relevant literature (from reputable conferences/journals), and kept the details to a minimum.

I decided to keep the "unreliable sources" tag in case anyone would like to review and/or anyone still thinks there are issues with the sources that haven’t been addressed. Otherwise, I think we can remove it. Please let me know if there are any comments/suggestions.

S@lo (talk) 06:45, 8 November 2021 (UTC)[reply]

Consider merging Software aging and Software rot

This article seems similar to Software aging and Software rot. Please consider merging the three into one. Timhowardriley (talk) 18:59, 10 March 2022 (UTC)[reply]