What happened to Metalink — the web’s almost-standard that almost nobody used?

What Metalink was

Metalink was an XML-based way to describe multiple mirrors, checksums, PGP signatures, and chunked download strategies for a single file so clients could choose the best source and verify integrity. It even landed in the standards stack: it has been reported that two RFCs (5854 and 6249) were published. The pitch was tidy: make downloads resilient, fast and verifiable without reinventing the wheel. Sounds useful, right?

Why it didn’t stick

But it didn’t. It has been reported that curl removed Metalink support in 2021 after roughly nine years of life in the codebase, and no major browser ever implemented it. Why? For one thing, you need both server-side adoption and client support — a classic chicken-and-egg problem. Browsers and CDNs never signed on, so the feature never reached everyday users. Meanwhile, the web evolved: CDNs got smarter, HTTPS became ubiquitous, HTTP/2/3 and range requests improved parallelism, and BitTorrent remained the go-to for wide distribution. Add the overhead of creating and hosting Metalink metadata and the lack of a killer consumer app or big corporate champion, and the momentum faded.

The emotional truth and what’s left

There’s a little frustration here: a neat, standards-backed idea that could have eased mirror management quietly withered because the market chose different shortcuts. It’s a reminder that good technology doesn’t win by merit alone; it needs the right timing, ecosystem buy-in, or a dominant platform to carry it. Allegedly, a few niche download managers and package systems kept Metalink alive in corners, but those pockets weren’t enough.

Could it come back?

Could Metalink stage a comeback? Unlikely unless a major browser, CDN, or package manager decides multi-source manifests are worth their investment. More plausibly, the same goals are being met piecemeal — multi-CDN routing, signed manifests, content-addressed delivery and P2P overlays — rather than by a single XML format. Metalink didn’t die because it was bad; it lost the race to network and market trends. That’s almost worse.

Sources: Lobsters