https://scribe.rip/p/the-forgotten-pixel-art-masterpieces-of-the-playstation-1-era-8b453dfe00bf is the Scribe.rip mirror for https://onezero.medium.com/the-forgotten-pixel-art-masterpieces-of-the-playstation-1-era-8b453dfe00bf .
4 days ago, the scribe.rip version was a full archive of the essay (which I know because I saved a copy while testing out HTML optimization).
However, a reader alerts me that when I linked to the scribe.rip as a good mirror of it, I was wrong: because now the Scribe.rip version is only a very incomplete version, containing only the first screen up to a paywall nag on the live Medium version. (The Medium version does not seem to have changed in any way in those 4 days, and I would be surprised if a post from 5 years ago had been modified.)
So something has broken on Scribe: a page that was working correctly 4 days ago is broken now.
Hi ~gwern. It looks like you put a bunch of effort into validating this. Thanks for doing that; it makes it a lot easier for me to understand what's going on.
Scribe uses a unlisted Medium API for fetching articles. When I launched scribe, it would give back the full article, even if it was paywalled. Slowly, over a number of months, I've seen more and more reports of paywalled articles showing only showing part of their content.
At this point, I expect all paywalled articles to show as incomplete. I'm tracking the issue over in the issue: Some paywalled articles require a login cookie so I'm marking this as duplicate.
As for your article showing fine, and then later showing as truncated, I'm not sure why that is. It's possible that the paywall API kicks in only after some time? Or perhaps you caught the one free view of it? I don't know for sure because I don't work at Medium.
I'm not surprised if Scribe uses API tricks or if it can't get a full article at all, and an article is broken when I try the Scribe version (if you can't get it, you can't get it); but I don't understand why the Scribe version was working but then broke. (I did look at that other thread about requiring login cookies, but thought it was unrelated: because obviously this pixel-art article didn't require a login cookie or else Scribe wouldn't've'd a full copy to begin with.) I would expect either it to start broken, or remain fixed. Is Scribe not keeping a copy at all, so it just breaks the moment Medium breaks it?
Is Scribe not keeping a copy at all, so it just breaks the moment Medium breaks it?
That's correct. It does not store copies of articles. It calls the Medium API every time.