「Wikipedia:Bot/使用申請/InternetArchiveBot」の版間の差分

削除された内容 追加された内容
271行目:
::::要点を翻訳します。(Translation) In short, the questions from Open-box-san are (1) Can you add {{tl|リンク切れ}} only if the page gives 404 error? If the page gives other errors (e.g. 403 error for this case), it is possible that the site blocks non-Japanese users. In such cases, it is better to add {{tl|webarchive}} but not adding {{tl|リンク切れ}}. (2) Also, for the case currently in question, just wonder why the archive date used is a very old date one - the one chosen by IABot is from 2006, but there is a [https://web.archive.org/web/20150320180356/http://www.mutsuwan-ferry.jp:80/ newer version from 2015]. (3) When the link is contained within a ref tag, both {{tl|リンク切れ}} and {{tl|webarchive}} are used correctly; but when it is just a plain link (as in External Link section), only {{tl|リンク切れ}} works fine. Namely, when there is an archived version available, the URL is simply changed to the archived one, without using webarchive. Questions from Takisaw-san: (1) Even very slight changes are ignored by IABot and simply switched to archived page - for example, in [https://ja.wikipedia.org/w/index.php?title=%E3%82%A2%E3%82%A4%E3%83%B3%E3%82%B9%E5%AE%97%E8%B0%B7&diff=prev&oldid=67750663] one only needs to remove "index.html" to show the page correctly; in [https://ja.wikipedia.org/w/index.php?title=%E3%81%8B%E3%82%82%E3%81%84%E5%B2%B3%E5%9B%BD%E9%9A%9B%E3%82%B9%E3%82%AD%E3%83%BC%E5%A0%B4&diff=prev&oldid=67750214] one only needs to remove "index.php". If these were done by human, such errors would not have happened in the first place; but since it looks difficult for a bot to cope with such cases, it sounds questionable to me whether such changes should/could be handled by a bot. (Open-box-san has one more example but I am skipping it - almost impossible to be detected by a bot.)
::::何かニュアンスが伝わっていないところがございましたらお申し付けください。--[[利用者:ネイ|ネイ]]([[利用者‐会話:ネイ|会話]]) 2018年3月18日 (日) 15:17 (UTC)
:::::{{ping|ネイ}} For translation: {{ping|Open-box}} So with my years of experience from this, dead URLs come in varying ways, including 403s. Then there are working pages that deliver valid content AND return a 404 for some reason. IABot works to try and avoid these confusing cases, but from my experience, geo-restricted domains is rare. They may be higher in Japan, I cannot confirm this, however, these domains can be whitelisted entirely so IABot no longer checks them and they will be maintained as live links. As for the archive date, IABot will typically pick a date closest to the time the site was initially accessed on Wikipedia, or to the time it was initially added to Wikipedia. In the case of a link outside of a ref tag, it may have gone to grab one from early on. You can of course instruct IABot to use a different URL by going to https://tools.wmflabs.org/iabot?page=manageurlsingle&wiki=jawiki and looking up the original URL. There you can change the archive URL IABot should use for it. As for number 3, when IABot edits a reference, it will format freely to keep as much data as possible for the sources. However outside of references, IABot assumes it may be editing a link that is integrated into the article and in order to ensure the article's readability isn't being disrupted, it will either tag the link dead, or outright replace the original with the archive to ensure the final render doesn't disrupt the page text. {{ping|Takisaw}} Any change in a URL is not "slight". If the old does not point to the new URL in the header information IABot receives from the page, there's no way for it to tell where the content's new ___location is. In any event, one should be concerned about the readers. Most don't even know how a URL works. They just type it in and click links to get to where they want to go. It won't occur to them to remove index.php, and they will very well see a broken link. If a site recently changed their index file, then a local bot should go and change them to the new correct URLs. This is done routinely on the English Wikipedia when an editor realizes the original content is still accessible, but just under a different URL. In the meantime IABot would keep the original link alive by providing archives. I hope this helps.—[[:en:User:Cyberpower678|'''<span style="color:darkgrey;font-family:monospace">CYBERPOWER</span>''']] <span style="font-family:Rockwell">([[:en:User talk:Cyberpower678|<span style="color:red">会話</span>]])</span> 2018年3月19日 (月) 02:58 (UTC)