And join Jack live on @source.coop's Great Data Products podcast on 11/19 to hear how we're making cultural memory collections easier to access and harder to delete. greatdataproducts.com/housekeeping...

This series is part of our work investigating not just the technical, but the human and societal routes to long-term digital preservation. You can also check out @maxy.bsky.social's Century-Scale Storage on our website.

lil.law.harvard.edu/century-scal...

One answer from Frank Cifaldi who is the Founder and Director of the Video Game History Foundation:

"I would have the best copyright lawyers in the country figuring out how we can actually make this work."

lil.law.harvard.edu/generational...

One answer from Rebecca Frank who is an Assistant Professor at the University of Michigan School of Information:

"The glib answer is the money itself would be the solution."

lil.law.harvard.edu/generational...

One answer from Amelia Acker who is an Associate Professor in the School of Communication & Information at Rutgers:

"I probably wouldn't build a system. I'd build a bureaucracy."

lil.law.harvard.edu/generational...

One answer from Che-Wei Wang and Taylor Levy, aka CW&T, who are designers, fabricators, and artists:

"Wait, only a hundred years?"

lil.law.harvard.edu/generational...

What are we all missing? Anything you can't get by clicking from link to link like EOT, or downloading datasets directly from data.gov. If there's things you care about preserving that fit that description, that's where to focus.

Another limited but vital effort is @eotarchive.org. They collect a huge amount from .gov domains every four years, and make it discoverable through @archive.org

Our collection from data.gov is limited: if an entry points directly to the data, such as a csv, we have the data. If it points to an html landing page, we just have the landing page. This means many, many datasets are not included. What we have from data.gov adds up to 15 or 20TB.

Speaking of telling someone, here’s our update: we have copies of all metadata from data.gov, and all of the dataset URLs it points to (shallow crawl); all federal Github repositories with issues, comments, etc.; and articles from PubMed.

Third, tell someone. Archive.org is one good place to store public data for discovery, and we at LIL will consider storing and signing data in some cases as well. Just posting data somewhere search engines can find is good too.

FOIA requests are another great way to scale up — check out @muckrock.com to get started.

Next, scale up. If you’re a programmer (or can team up with one), write a python script to download a full collection — say, everything from the data portal of a given government website. Run it yourself, and share it so we libraries can use it too.

If you’re a data scientist, good news — your work isn't just downloading data and publishing about it, but also keeping safe copies!

To keep access to stuff you care about: first just make a copy. Use ArchiveWeb.page to click around and download all the parts of a website you’re interested in. We like the desktop version to avoid capturing login cookies or extensions, but the browser extension is good too.