The Butterfly Net and the Sieve: Collecting vs. Curating the Public Web

When we talk about saving the web, we usually picture a grand, almost frantic, act of collection. A digital Noah’s Ark, tasked with gathering two of everything before the flood of obsolescence and link rot washes it all away. This is the domain of institutions like the Internet Archive, wielding what I’ve come to think of as the "butterfly net" approach: broad, automated, and vast in its aspirations. It aims to capture the sheer, teeming ecosystem of the web in flight, collecting billions of pages with the understanding that comprehensiveness, even with imperfections, is the primary goal. It’s a monumental, essential effort to preserve the whole swamp, complete with its mosquitoes and orchids.

But there is a second, contrasting philosophy at work in the shadows of our digital preservation efforts. It is quieter, more methodical, and operates with a jeweler’s eye. This is the "sieve" approach, practiced by specialized archives, academic projects, and community-led initiatives. Instead of a net, they use a fine mesh, designed to let the sand of the everyday web fall through, catching only the specific, glittering grains of a particular event, theme, or community. Where the butterfly netter sees a whole meadow, the sieve-keeper focuses on a single, rare type of pollen.

The difference is starkest when you look at what each method captures. A butterfly-net archive of a news website from a specific day might give you the front page, the sports section, and the classifieds—the entire organism. A sieve archive of the same day, focused on, say, the local arts scene, might ignore the headlines entirely. Instead, it would capture a specific gallery’s event page, a now-deleted blog post by a featured artist, the comments on a related social media thread, and the raw data from a public arts funding database. It captures context and depth, not just breadth. It seeks to preserve not just the page, but the conversation happening around it.

Neither approach is inherently superior; they are deeply complementary, two halves of a whole preservation strategy. The butterfly net creates the baseline—the massive, searchable corpus that allows us to track cultural shifts and recover lost pages we didn’t know we needed. It answers the question, "What was on the web on this date?" The sieve provides the narrative—the curated collection that allows a future historian to understand a specific event in rich detail. It answers the question, "What was the lived experience of this community during this crisis?"

We need both. Without the net, the sifted collections risk becoming isolated artifacts, detached from the wider digital landscape that gave them meaning. Without the sieve, the colossal net-fuls of data can become overwhelming and impenetrable, a hoarder’s attic where the most significant stories are buried under mountains of the mundane. The true art of web archiving lies in understanding when to cast the net wide and when to deploy the sieve with precision, ensuring that our digital memory is not just vast, but also intelligible and meaningful for the questions we have yet to ask.

Notes & further reading

A few pages I came back to while writing this: