- The following is an archived discussion of the DYK nomination of the article below. Please do not modify this page. Subsequent comments should be made on the appropriate discussion page (such as this nomination's talk page, the article's talk page or Wikipedia talk:Did you know), unless there is consensus to re-open the discussion at this page. No further edits should be made to this page.
The result was: promoted by Z1720 (talk) 14:41, 15 July 2022 (UTC)
| DYK toolbox |
|---|
Pytest
Pytest logo
... that technology projects from across the internet, including those of Mozilla and Dropbox, are switching to Pytest from other frameworks for software testing?
Quote: In fact, projects all over the Internet have switched from unittest or nose to pytest, including Mozilla and Dropbox.Okken, Brian (September 2017). Python Testing with Pytest (1st ed.). The Pragmatic Bookshelf. ISBN 9781680502404. Retrieved 19 March 2022.
Created by Thomas Meng (talk). Self-nominated at 01:29, 19 March 2022 (UTC).
An interesting topic and clearly notable. However, the article is correctly tagged as being in need of rewriting, to be less like an instruction manual, and more like a NPOV article. And in that should make hopefully make the article understandable to a normal reader- I understand the article and its details, but only because I work in the field. This will need to be fixed before this DYK can proceed. Joseph2302 (talk) 17:41, 22 March 2022 (UTC)
- @Joseph2302: Thank you for your feedback. In the past few days, I took up an effor to fix those issues you mentioned. Now I think the article is in better shape. Please let me know how far it is now from DYK's standard. Thank you. Thomas Meng (talk) 01:54, 31 March 2022 (UTC)
- Apologies, this slipped my mind. Reviewing properly now:
- The article still has multiple paragraphs without citations. The minimum amount of sourcing I'd expect is one source per paragraph- if the sources already in the article support the text where I've added citation needed tags, then that should be quick to fix
- The text is better, but it still very technical (which does seem to be the case for a lots of computing articles I've noticed). I understand that it's a technical topic, but there's almost nothing in the article that an average reader would understand. Some articles like Node.js for example has a "History" section, which would be beneficial to a less technical reader. There's still so much code in this article that it's too technical and confusing, and still feels to me like it's a manual on how to use it. Joseph2302 (talk) 16:56, 6 April 2022 (UTC)
- @Joseph2302: Here's what I've done to address the problems you pointed out:
- Modified every section so that each section's first paragraph(s) would only include pytest concepts, and implementation details are saved for the end. Additionally, wording/explanations are improved where possible.
- Added a History section for less technical users to read. The lead section should also be understandble for them.
- Added ~20 wikilinks for programming related concepts.
- The citations problem is also fixed.
- Unecessary code templates (e.g. for file, project names) that hinder readability are removed.
- Thank you for your time and I look forward to hearing back from you. Thomas Meng (talk) 02:19, 27 April 2022 (UTC)
- @Joseph2302: Here's what I've done to address the problems you pointed out:
- @Joseph2302: Thank you for your feedback. In the past few days, I took up an effor to fix those issues you mentioned. Now I think the article is in better shape. Please let me know how far it is now from DYK's standard. Thank you. Thomas Meng (talk) 01:54, 31 March 2022 (UTC)
- Hmm, I can try to shepherd this through. The prose is okay- not great, it'll need a bit more work, but the more immediate concern is sourcing. Some sources are good, some are iffy due to their status as primary sourcing, and some just shouldn't be used:
- Dane Hillard's "Effective Python Testing With Pytest"
- tim's "Assertion rewriting in Pytest part 1: Why it’s needed"
- Microsoft's "Unit test basics"
- Klein's "Testing with pytest"
- Perfecto's "Pytest marks"
- These all appear to be secondary, yet non-professional sources ranging from personal blogs to coding lessons to company blogs. I don't think any of those meet DYK's reliable sourcing standards, and material relying on it needs a more well-developed source like a book, magazine, newspaper article, scholarly journal, or otherwise. theleekycauldron (talk • contribs) (she/they) 01:51, 27 May 2022 (UTC)
- Thank you theleekycauldron for taking the time to review. I'll start editing this article in the next few days and probably finish improving the sources by the end of this weekend. Regards, Thomas Meng (talk) 03:57, 2 June 2022 (UTC)
@Theleekycauldron:, thank you for your patience. I have now fixed the five problems you pointed out by
- Removing all four references to Real Python — since three of which had already been corroborated by Okken's book, I only had to delete one short paragraph of actual content.
- Removing Tim's blog and replacing content and sourcing with Oliveira's book.
- Removing this crowd-contributed source and replacing it with a new book — Unit Testing Principles, Practices, and Patterns.
- Removing source and replacing with Okken's book with specific page references.
- Same as No.4
Regards, Thomas Meng (talk) 13:32, 4 June 2022 (UTC)
- Looks good, Thomas Meng! Next, we're gonna need to clear up the prose, it's quite jargon-y; and a new hook will be needed, as I don't think ALT0 is viable. theleekycauldron (talk • contribs) (she/they) 19:35, 8 June 2022 (UTC)
- Thanks theleekycauldron. I've actually tried to fix the jargon-y issue for several rounds now, and I ended up with the current version. It seems that most programming-related articles do rely on jargons (e.g. Node.js and Python (programming language), which is GA), but with wiki-links to them. So that's what I've been trying to do — adding wiki-links and improving explanations for programming concepts where there aren't wiki-links. Maybe could you be more specific on where exactly you'd like the prose improved? I'll be in a better position to fix it then.
- The current hook is paraphrased from Okken's book, which I thought was quite indicative of pytest's popularity. Could you please elaborate on why it's not viable so that I know how I should fix it? Thank you. Thomas Meng (talk) 03:02, 10 June 2022 (UTC)
- To me, it sounds like an ad. Promotional and without any real connection to the topic itself. —David Eppstein (talk) 00:14, 13 June 2022 (UTC)
- Agreed: it reads like an ad. I've struck it; a new hook will need to be found. BlueMoonset (talk) 15:19, 13 June 2022 (UTC)
- To me, it sounds like an ad. Promotional and without any real connection to the topic itself. —David Eppstein (talk) 00:14, 13 June 2022 (UTC)
- Thank you for the feedback BlueMoonset, prof David Eppstein. Here are two alternatives that give some interesting facts:
- ALT2:
... that pytest, an open-source software testing project in Python, supports various types of software testing, including unit tests, integration tests, end-to-end tests, and functional tests?
- ALT2:
While pytest is useful for unit testing, integration testing, system or end-to-end testing, and functional testing, the strategy for testing the Tasks project focuses primarily on subcutaneous functional testingOkken, Brian (September 2017). Python Testing with Pytest (1st ed.). The Pragmatic Bookshelf. ISBN 9781680502404. Retrieved 19 March 2022.
- ALT3:
... that pytest is an all-volunteer contributed, open-source software testing tool in Python and has been classified as a key ecosystem project on the PyPI with over 9 million weekly downloads?"pytest". pytest.org. Retrieved 15 June 2022."Python Package Health Analysis". snyk.io. Retrieved 15 June 2022.
- ALT3:
Regards, Thomas Meng (talk) 19:18, 15 June 2022 (UTC)
- I was gonna say the same about the original hook- and unfortunately, I think it applies to both of these, as well. It does feel a little promotional. theleekycauldron (talk • contribs) (she/they) 19:22, 15 June 2022 (UTC)
- [edit conflict] You're still writing a hook that reads like you are trying to persuade Python programmers to use pytest. That's what it means to be an ad. Try taking the point of view of an encyclopedia reader who is not a programmer. What about pytest could you write that would intrigue a reader and get them to read an encyclopedia article about pytest, even if they have no intention of becoming a programmer? —David Eppstein (talk) 19:23, 15 June 2022 (UTC)
- ALT4:
... that pytest is a Python framework for writing software tests that help make sure programs are bug-free?How about this one?
- ALT4:
David Eppstein, theleekycauldron It's a bit difficult to write a hook that can hook non-programmer's interest into a programming tool — not going into any of pytest's features (e.g. parametrized testing or assert re-writing) due to understandability, while also not including any easy-to-understand facts like download trend or popular usage. I see that prof David Eppstein has had experience writing DYK hooks for technical articles. Perhaps you could help compose this hook?. Thank you. Thomas Meng (talk) 20:46, 15 June 2022 (UTC)
- I don't know, my usual strategy is to only make DYK nominations when I think there is something in an article or hook that would stand out to a general audience. When I have a chance to nominate an article, but it is more purely of technical interest, I skip it. Not everything needs to go to DYK. I tried reading through the article a couple of times but nothing stood out to me. —David Eppstein (talk) 06:43, 16 June 2022 (UTC)
ALT2 or ALT4 are hooked and quoted. Long enough, (posted) new enough, DYKCheck all green, no copyvio. All of the issues raised above have been addressed long ago. Neither the hook nor the article strike me as promotional in their current form. This was posted two months ago and now the hangup is "not everything needs to go to DYK"?! Enough already, this is good to go. Maury Markowitz (talk) 18:29, 17 June 2022 (UTC)
- It's more than that. The hooks are not interesting to a broad audience. Maybe theleekycauldron or a different prep builder will promote one of those hooks, but I'm against it. SL93 (talk) 22:35, 18 June 2022 (UTC)
- @SL93: I think ALT4 is perfectly interesting to the average reader. Maury Markowitz (talk) 14:02, 21 June 2022 (UTC)
Back under discussion again after editors at Wikipedia talk:Did you know questioned whether "bug-free" could ever be an appropriate description for software. —David Eppstein (talk) 19:25, 29 June 2022 (UTC)
- Maybe change "bug-free" to "limit bugs" or some synonym of that? It would suggest that bugs are still possible, just that efforts are being made to eliminate them. Narutolovehinata5 (talk · contributions) 00:20, 30 June 2022 (UTC)
- Oddly, our article itself is almost bug-free: its only use of the word "bug" in the whole article involves pypy but not pytest. In fact, it says little or nothing about why you might want to test your software. That is going to make it difficult to write any hook involving bugs. I don't see how the previous hook could have been approved without anything about "bug-free" software appearing in the article. —David Eppstein (talk) 19:31, 30 June 2022 (UTC)
- David Eppstein This is the third instance of this reviewer approving a hook that isn't in the article. SL93 (talk) 16:21, 2 July 2022 (UTC)
- Oddly, our article itself is almost bug-free: its only use of the word "bug" in the whole article involves pypy but not pytest. In fact, it says little or nothing about why you might want to test your software. That is going to make it difficult to write any hook involving bugs. I don't see how the previous hook could have been approved without anything about "bug-free" software appearing in the article. —David Eppstein (talk) 19:31, 30 June 2022 (UTC)
- ALT5
... that pytest rewrites assertions in Python programs, which check data consistency and trigger exceptions if problems are found, into software tests that can detail the causes of failures?
- Maybe change "bug-free" to "limit bugs" or some synonym of that? It would suggest that bugs are still possible, just that efforts are being made to eliminate them. Narutolovehinata5 (talk · contributions) 00:20, 30 June 2022 (UTC)
- Based on pytest § Assert rewriting, something that differentiates pytest from other testing frameworks such as unittest. I tried to write this at a level readable to a general audience; not sure if I succeeded. —David Eppstein (talk) 22:09, 1 July 2022 (UTC)
- Unfortunately no, that hook is way too technical for a general audience. Narutolovehinata5 (talk · contributions) 15:32, 2 July 2022 (UTC)
- ALT6
... that pytest detects errors in Python computer programs using tests that are structured in an arrange, act and assert sequence known as AAA?Article: "Fixtures practically constitute the arrange phase in the anatomy of a test (AAA, short for arrange, act, assert)." Source 1: "The tests are written in the classical style and use the typical three-phase sequence: arrange, act, and assert (AAA for short...) " Source 2: "I prefer pytest because of some advanced features (fixtures, plug-ins, etc.),... One of the most common patterns you'll find is the 3A or AAA test pattern. AAA stands for Arrange-Act-Assert."
- I'm suggesting this as a possible hook because it is in words a general reader can be curious about. :-) This isn't a hook that will catch an expert's interest, but expert hooks are likely to be overly technical or seen as promotional or both. Mary Mark Ockerbloom (talk) 20:39, 3 July 2022 (UTC)
- Unfortunately I don't think this hook solves the "too technical" concerns. Someone knowledgeable in tech would get it, but the average reader would just see it as word salad. Narutolovehinata5 (talk · contributions) 07:37, 4 July 2022 (UTC)
- Sometimes people click on word salad just to see what in the world it could be talking about -- and then they get a glimpse into a topic area they previously knew nothing about. In fact, I'd make it even a bit more obscure by dropping "computer program":
- ALT7
... that pytest detects Python errors via tests structured in an arrange, act and assert sequence known as AAA?
- ALT7
- EEng 15:03, 6 July 2022 (UTC)
- If it makes a difference, I think Alt 6 will do. --evrik (talk) 16:48, 6 July 2022 (UTC)
- To forestall going through some of the same bad-promotion-undo cycle we went through already with this: Where in the article do you see a sourced claim that pytest detects errors in Python computer programs? The rest of the hook is ok, but where is that part from? It may seem obvious to people who know about software tests that that's what tests do, but it doesn't seem to be in the article. To be pedantic, it's not the errors in the programs that the tests detect, but the corresponding error in the behavior of the programs. Figuring out where the error is in the program can be significant additional effort from that point. And even if the "detects errors" claim can be found, is an unqualified assertion that it detects errors accurate? All of them? —David Eppstein (talk) 17:03, 6 July 2022 (UTC)
- @David Eppstein: here: Pytest#Pytest_fixtures, albeit, I was inferring from the text and reading it into the hook. @EEng and Thomas Meng: can anyone cite that sentence. --evrik (talk) 18:47, 6 July 2022 (UTC)
- To forestall going through some of the same bad-promotion-undo cycle we went through already with this: Where in the article do you see a sourced claim that pytest detects errors in Python computer programs? The rest of the hook is ok, but where is that part from? It may seem obvious to people who know about software tests that that's what tests do, but it doesn't seem to be in the article. To be pedantic, it's not the errors in the programs that the tests detect, but the corresponding error in the behavior of the programs. Figuring out where the error is in the program can be significant additional effort from that point. And even if the "detects errors" claim can be found, is an unqualified assertion that it detects errors accurate? All of them? —David Eppstein (talk) 17:03, 6 July 2022 (UTC)
- If it makes a difference, I think Alt 6 will do. --evrik (talk) 16:48, 6 July 2022 (UTC)
- Sometimes people click on word salad just to see what in the world it could be talking about -- and then they get a glimpse into a topic area they previously knew nothing about. In fact, I'd make it even a bit more obscure by dropping "computer program":
- Unfortunately I don't think this hook solves the "too technical" concerns. Someone knowledgeable in tech would get it, but the average reader would just see it as word salad. Narutolovehinata5 (talk · contributions) 07:37, 4 July 2022 (UTC)
- ALT6
It actually might be difficult to find a source explicitly stating that the purpose of a test facility is to find bugs -- sort of like finding a source explaining that the purpose of a gun is to fire bullets. But I don't think we can go wrong with this:
- ALT8 ... that pytest uses tests structured in an arrange, act and assert sequence known as AAA?
I wish I could help more but I really can't. EEng 20:41, 6 July 2022 (UTC)
- Six is better, but this could do. --evrik (talk) 20:46, 6 July 2022 (UTC)
- I'd disagree here. ALT7 and its variants are still quite technical, and I'm saying this as a techie myself. I don't think the typical reader would understand this hook, and instead of being inspired to click and understand, they'd probably only be confused further and walk away. Narutolovehinata5 (talk · contributions) 01:16, 7 July 2022 (UTC)
- Look, this isn't anyone's Nobel acceptance speech, just a fucking DYK hook. As it stands, ALT8 is the only hook that's actually correct. You're overthinking this. EEng 02:40, 7 July 2022 (UTC)
- DYK hooks are supposed to be
interesting to a broad audience
(emphasis mine). It's even in the reviewing instructions. This does not count. And no, before you say "it's vague on purpose because it will make people click the link to find out more", that only works for some hooks. As for this one, I don't think it's one of those times. Non-techies are not going to know what an assert sequence is. Even if we give the hook the benefit of the doubt and say it's the only correct option, it isn't even hooky. It doesn't say anything interesting about the subject. Narutolovehinata5 (talk · contributions) 05:28, 7 July 2022 (UTC) - I also have to note that the article currently has a citation needed tag. Narutolovehinata5 (talk · contributions) 09:21, 7 July 2022 (UTC)
- DYK hooks are supposed to be
- Look, this isn't anyone's Nobel acceptance speech, just a fucking DYK hook. As it stands, ALT8 is the only hook that's actually correct. You're overthinking this. EEng 02:40, 7 July 2022 (UTC)
- I'd disagree here. ALT7 and its variants are still quite technical, and I'm saying this as a techie myself. I don't think the typical reader would understand this hook, and instead of being inspired to click and understand, they'd probably only be confused further and walk away. Narutolovehinata5 (talk · contributions) 01:16, 7 July 2022 (UTC)
- Six is better, but this could do. --evrik (talk) 20:46, 6 July 2022 (UTC)
- ALT9 ... that pytest verifies that programs will perform as expected by using tests structured to arrange, act and assert? Article text is updated and cited, citation needed tag has been removed. Mary Mark Ockerbloom (talk) 18:42, 8 July 2022 (UTC)
- "Did you know that a program testing library made Wikipedia editors spend over 2,800 words arguing over how to present it on the site's main page?" Alas, "verifies" is too strong a word, although it could be tweaked. I think ALT8 is passable and it's unlikely we'll be able to milk anything better out of the article. I did a brief reading into pytest's history and found nothing particularly interesting. Ovinus (talk) 19:14, 8 July 2022 (UTC)
- It's not merely too strong a word; it's the wrong word. In this context, it's a technical word that has a specific meaning, different from what was intended. Program verification is a completely different approach than software testing for improving reliability. —David Eppstein (talk) 19:36, 8 July 2022 (UTC)
- To misquote Vizzini, working on this is like making the mistake of fighting a land war in Asia. I say we use a version of 6 or 8 and be done with it. --evrik (talk) 21:30, 8 July 2022 (UTC)
- If an article nominated for DYK doesn't have a usable hook, we don't need to force ourselves and squeeze out a hook. A nomination can be rejected for lack of suitable material, there's no rule that requires technically eligible articles to be approved as long as they have a hook, even if said hook is ultimately uninteresting. You can't make pure lemonade out of orange no matter how hard you try. Narutolovehinata5 (talk · contributions) 23:30, 8 July 2022 (UTC)
- To misquote Vizzini, working on this is like making the mistake of fighting a land war in Asia. I say we use a version of 6 or 8 and be done with it. --evrik (talk) 21:30, 8 July 2022 (UTC)
- It's not merely too strong a word; it's the wrong word. In this context, it's a technical word that has a specific meaning, different from what was intended. Program verification is a completely different approach than software testing for improving reliability. —David Eppstein (talk) 19:36, 8 July 2022 (UTC)
- "Did you know that a program testing library made Wikipedia editors spend over 2,800 words arguing over how to present it on the site's main page?" Alas, "verifies" is too strong a word, although it could be tweaked. I think ALT8 is passable and it's unlikely we'll be able to milk anything better out of the article. I did a brief reading into pytest's history and found nothing particularly interesting. Ovinus (talk) 19:14, 8 July 2022 (UTC)
- David suggested the following hook at WT:DYK below. Personally I'm more partial to it even though it's admittedly not the best hook. At least, I think it's better than ALT6/8 and ALT9 above.
- On the other hand, if an article really has no suitable hook facts, failing is always an option. Narutolovehinata5 (talk · contributions) 23:30, 8 July 2022 (UTC)
- Correction: Gatoclass suggested it. —David Eppstein (talk) 01:53, 9 July 2022 (UTC)
- Alt 10 is okay. FWIW, how are failure to agree on a hook doesn’t mean this nomination should fail..--evrik (talk) 03:28, 9 July 2022 (UTC)
- If no hooks can be proposed about a subject, it can be a reason to fail a nomination. Not all articles are meant for DYK, especially if they do not have any information that would be appealing to broad audiences. Narutolovehinata5 (talk · contributions) 04:31, 9 July 2022 (UTC)
- Given that nobody else appears to have spelled this out: the nominator is QPQ-exempt, with just one credit to their name. Schwede66 01:48, 12 July 2022 (UTC)
Can we approve Alt 10 and let this move forward? --evrik (talk) 23:20, 13 July 2022 (UTC)
Approving ALT10. I am late to the party but don't see anything wrong with the article, QPQ excempt, and the hook seems at least moderately interesting to me (and I have nigh zero knowledge of any programming). I could name a dozen only moderately interesting hooks that got approved, so this should not be failed imo. --LordPeterII (talk) 09:01, 14 July 2022 (UTC)