“tampering with the public record”

This was a pretty good call-out on Twitter for changing out Embedded Tweets work from Kevin Marks.

Twitter has always provided semantic HTML for Embedded Tweets, in the form of a <blockquote>, then a <script> to enhance it up into a fancy embed. It exhibited an interesting (and good) behavior: if the Tweet no longer existed, it would fall back to a standard HTML rendering of the <blockquote>. The change was that it wouldn’t do that anymore, it would always render the fancy embed, only show a “missing tweet” UI, in order to “better respect when people have chosen to delete their Tweets”.

I can understand that perspective, but I can better understand the perspective of a publisher. If I’m putting an Embedded Tweet into an article, it’s there for a reason. It’s there to provide context or perhaps be the content. Suddenly and without warning a bunch of articles now had contextual holes in them because tweets that used to just be showing some HTML of the original text were now showing an empty box.

It’s a trust problem. If we can’t trust that Embedded Tweets say what they say at the time of embedding, then we just won’t do it. We’ll use a <blockquote> without the script, or if we’re angling for the social proof, perhaps a screenshot of the tweet. I imagine some folks already do that, if they are strongly performance and/or privacy-focused and using Twitter’s third-party JavaScript isn’t on the table. But for all those sites out there that have used it, Twitter ought not mess with the content presented.

This whole thing will come up again if Tweets become editable. That, too, messes with the public record and absolutely ensures nobody will use Embedded Tweets. The second a major publication puts an Embedded Tweet that is editable in an article: boom, goatse.