Talk:Cargo plane crashes in Kyrgyzstan, killing over 30

Foreign lang sources
Per WN:CS, foreign language sources "...severely restrict the pool of potential reviewing editors. Non-English sources should be restricted to those absolutely essential..." I doubt we have any active reviewers who can read Turkish or Kyrgyz sources, sadly, and I also very much doubt it's impossible to find good English-language coverage of what is a major disaster. I'd come here wanting to review or expand, but find I cannot do the former and the latter may well go stale on the review queue based on the foreign-language sources.

In an ideal future we would have reviewers capable of understanding all manner of languages. We are, unfortunately, a long way from that. BRS (Talk)   (Contribs) 16:25, 16 January 2017 (UTC)
 * I just added a ton more English-language sources. Hth! :) —Firespeaker (talk) 21:03, 16 January 2017 (UTC)
 * Instrad of adding "a ton more English", just add those which can help verify the story without requiring the foreign language to be used. Adding so many sources would consume lot for time for a small article like this. Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  21:13, 16 January 2017 (UTC)
 * And, LMGTFY, really? It is not the job of reviewers to find sources for someones articles. If they do, they might not be able to review the article since they would no longer be an uninvolved reviewer. Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  21:15, 16 January 2017 (UTC)
 * Thanks for the Welcome!—Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * BBC and Al Jazeera's source can verify the majority of the story. What I could not verify was
 * BBC source says «Deputy PM Muhammetkaly Abulgaziev said on state TV: "According to preliminary information, the plane crashed due to a pilot error."» —Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * I removed this material that I had initially found in a local source, as I couldn't verify it in an English-language source. —Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * BBC source says "There was fog at the time but authorities said other planes had landed successfully before the crash" —Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * Almost every English-language source says that. —Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * I linked to the other news story on wikinews; see the sources cited there (is that how this is done?). —Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * I would say, remove all those sources which were added just to tell that there are English news sources to write about it, and list only those sources which can help in the verification of the whole story. Adding a bulk of sources just does not work. Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  21:28, 16 January 2017 (UTC)
 * They all verify different aspects of it at this point. I was told there were too few English-language sources and now there are too many.  So what's the magic number? —Firespeaker (talk) 00:04, 17 January 2017 (UTC)
 * There's no formula; there are various principles, sometimes there's tension between them. I'm honestly surprised to see such subtle sourcing issues coming up on a first article here.  --Pi zero (talk) 00:38, 17 January 2017 (UTC)
 * Could you share some of the principles? —Firespeaker (talk) 05:19, 17 January 2017 (UTC)
 * "Manas international airport is located outside the capital city of Bishkek, and has been the location of other fatal crashes. These include a Boeing 737 en route to Iran, which crashed shortly after take-off in 2008, killing 68 of the over 90 people on board." Since you are using "and", and I can not find some part of it, I would say that I did not find the entire thing because Judy saying a part would make it meaningless. If BBC has the information, why not delist all the unnecessary sources? Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  03:52, 17 January 2017 (UTC)
 * I'm confused why you guys are surprised. And I'm confused why you needed to remove most of the sources, including one of the most complete ones.  And I'm confused about what I still need to do to make the article publishable. —Firespeaker (talk) 05:18, 17 January 2017 (UTC)
 * It is not a good practice to list the sources which were not used to write the story. Since BBC can verify most of the information, as you have pointed out, I chose BBC to keep in the list. As the history indicates, you just added the English sources, left a LMGTFY link on the talk page, but did not modify the content, it is likely that you did not use the source to prepare the story. Since Wikinews is not Wikipedia, adding one more source saying the same stuff does not help. At Wikipedia, the ref's tells us that this piece of information can be found at these links. On Wikinews, each source is used for the verification of the story, and what I felt after reading the English sources, yesterday, is most of the sources say the same thing. BBC has longer story, and Al Jazeera has a pretty good article. (NYTimes is paywalled, and I need to clear cache and delete cookies to read the article) Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  05:50, 17 January 2017 (UTC)
 * That addresses the 2nd thing I was confused about, but not the others. —Firespeaker (talk) 06:30, 17 January 2017 (UTC)
 * I don't think there is anything which can prevent it for publishing (if the copyright, NPOV, Style and verifiability is not a problem). I don't know in what context do you think we are surprised.
 * "Manas international airport is located outside the capital city of Bishkek, and has been the location of other fatal crashes. These include a Boeing 737 en route to Iran, which crashed shortly after take-off in 2008, killing 68 of the over 90 people on board." Since you are using "and", and I can not find some part of it, I would say that I did not find the entire thing because Judy saying a part would make it meaningless. If BBC has the information, why not delist all the unnecessary sources? Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  03:52, 17 January 2017 (UTC)
 * I'm confused why you guys are surprised. And I'm confused why you needed to remove most of the sources, including one of the most complete ones.  And I'm confused about what I still need to do to make the article publishable. —Firespeaker (talk) 05:18, 17 January 2017 (UTC)
 * It is not a good practice to list the sources which were not used to write the story. Since BBC can verify most of the information, as you have pointed out, I chose BBC to keep in the list. As the history indicates, you just added the English sources, left a LMGTFY link on the talk page, but did not modify the content, it is likely that you did not use the source to prepare the story. Since Wikinews is not Wikipedia, adding one more source saying the same stuff does not help. At Wikipedia, the ref's tells us that this piece of information can be found at these links. On Wikinews, each source is used for the verification of the story, and what I felt after reading the English sources, yesterday, is most of the sources say the same thing. BBC has longer story, and Al Jazeera has a pretty good article. (NYTimes is paywalled, and I need to clear cache and delete cookies to read the article) Agastya Chandrakant   ⚽️ 🏆 🎾 🎬 🎤 📰  05:50, 17 January 2017 (UTC)
 * That addresses the 2nd thing I was confused about, but not the others. —Firespeaker (talk) 06:30, 17 January 2017 (UTC)
 * I don't think there is anything which can prevent it for publishing (if the copyright, NPOV, Style and verifiability is not a problem). I don't know in what context do you think we are surprised.

The first commenter here, BRS who started the thread, is an authorized reviewer, and quoted from a policy page. I am an authorized reviewer. AGastya is making a sincere effort to be helpful; no, he's not an authorized reviewer. I think AGastya may be discovering something I've been aware of for some time: if some particular point of policy/practice isn't already elegantly articulated in our documentation, we probably don't yet know how to explain it succinctly. What happens with these issues is, each time they come up in a review, the reviewer thinks them through and, if they're central to the review outcome (especially if that outcome is "not ready"), the reviewer then attempts to explain in comments on the review template. Over time, probably many years, our practices get honed and we slowly learn what explanations work and what don't. And then eventually we hope to find time, between reviews, to create an essay explaining things. Such an essay, once created, is just the tip of an iceberg reaching well back into the history of the project, notwithstanding when and by whom the essay was created. We may indeed be coming close to ready to spawn an essay on use of non-English sources.

A few points on what to avoid:
 * Avoid using non-English sources in the first place unless there's a good reason. Information not available in English might be a good reason, but don't then rely on the non-English source for more than necessary, and it's probably good to provide a note on the talk page explaining what specifically it's used for.  Understand what the problem is:  non-English sources are almost always in a language the reviewer isn't fluent in (and may be completely clueless about), and automatic translators are astoundingly bad; they sometimes get basic meaning wrong, and are sure to lose the nuances/overtones that are an essential part of what a reviewer needs to get at.  The more limited, and the more objective, the information being taken from the non-English source, the better; the reporter had better have some modest degree of fluency in the source language (rather than relying on automatic translation themselves), and the reporter and reviewer both have to exercise good judgement in recognizing when there's likely to be a problem.  Sometimes we do use non-English sources to good effect; a simple but rewarding example is where we provide an English-translation direct quote and then also provide the non-English original so readers can assess the translation for themselves (we've a template for that, translated quote), and sometimes we can draw on non-English sources for quotes that simply aren't available elsewhere in the English press.
 * (Btw, re the horrors of automatic translation, I somewhat recall a case where automatic translation from a non-Indo-European source language produced an assertion that a high official of a country (Prime Minister, maybe?) had beheaded someone, when as best I could tell on intensive study the source was saying that one or the other of them &mdash; I'm not sure I was ever able to tell which with certainty &mdash; had a head cold. There are things a reviewer can do to get extra milage out of automatic translation of a given passage, involving dictionaries, and experiments with running the automatic translator on many different fragments of the passage, and reasoning; it's not that hard to spend huge amounts of time studying a single sentence that way, with mixed results.)

Keep in mind, the amount of volunteer labor needed to review articles is a bottleneck for us (yes there's a long-term plan to alleviate the problem; it's been underway for several years so far); that is, for instance, why this article hasn't been reviewed yet. --Pi zero (talk) 14:01, 17 January 2017 (UTC)
 * If a source wasn't actually drawn on to write the article, only add it if needed to help with corroboration (two-source rule or, in this case, facts hard to extract with confidence from foreign sources). And, of course, document on the talk page that that's why you're adding it (or document it in embedded html comments in the article markup, and note on the talk page that you're doing that).  Keep in mind, whenever adding any source, that there should be some benefit to the article from it because the volunteer labor required to review the article goes up with each source added.
 * If a source was drawn on to write the article, don't remove it, or note on the talk page what was removed (it matters what was used, even if we do draw on other sources for corroboration).