Just trying to clear a few things off my hard drive. Here was one that was particularly curious when I was investigating what was going on with Bing: the files submitted by Cloudflare’s IndexNow. The theory: it would send Bing the newest accessed pages to add to the index. The reality: these are not new. In fact, these are ancient, many aren’t even web pages (they’re PDFs and web fonts). And sure enough, some did make it into the 10–55 pages that Bing is capable of indexing for Lucire these days—it’s a very tiny index in reality, regardless of how many results it claims to have for a given search, as we discovered.
In other words, IndexNow, as I saw it implemented, is a total crock, and not worth the bother.
I wish these companies would test these things first, but we are talking Microsoft, where we’ve been doing the job as unpaid QA for decades.
It does get worse. Looking inside Bing Webmaster Tools, these (below) are the pages it says it has for Lucire’s root directory. I’ve alluded to how bad it was earlier, but upon going through these, the main index pages, which Bing always had till recently, are missing. The home page is also missing (although when I first started investigating in July, it was still there, which a friend can confirm; and the structure of it has not changed other than the removal of some links to 404s). All that’s left are pages from the early 2000s, plus entries for pages that have never existed. You can check these against the Wayback Machine, but we have never had pages in the main directory called nguoi-noi-tieng, arts-culture, podcast, form-single.html, archivi or cv-generator. Yet Bing believes these phantom pages exist. Well done, Microsoft, you can’t even get this right. This isn’t how spidering works.