The TAB report on the UMN affair

Posted May 7, 2021 15:05 UTC (Fri) by calumapplepie (guest, #143655)
Parent article: The TAB report on the UMN affair

I find the use of Google Drive somewhat annoying, though I do understand it (easy filesharing service). Files in google drive are vulnerable to linkrot, and are difficult to automatically archive: if you visit archive.org for these files, you get an endless loading screen. If google ever changes their URL scheme, or decides to delete a bunch of rarely accessed files, or gets , these letters will be gone. Then, Mr. Future Historian writing the history of the Fedebian Starship Operating System Kernel will be unable to read them.

Repalacing */edit and */view with */preview produces a reasonably archive-able version: not quite the .PDF source, but still pretty OK. Getting the original pdf files archived in an easy way would be harder. Fortunately for Sir Future Historian, the /preview pages for these specific documents have been preserved in the in the internet archive by some random person who then referred to that in a comment on an LWN article.

to post comments

The TAB report on the UMN affair

Posted May 7, 2021 15:43 UTC (Fri) by tytso (✭ supporter ✭, #9993) [Link] (2 responses)

I've seen links to web pages and pdf files saved in content management systems (such as what is used by the Linux Foundation) that have become inaccessible because when the CMS is changed, the links break. When you migrate to a new CMS, maintaining old links links is *hard*. You would think that would be easier to just dump pdf files into a static directly and let that be served up by Apache, but for large organizations, it's hard; the people who manage content generally don't have the technical ability (nor would it be safe!) to give them ssh access to the organization's web server. This is why large organizations use CMS systems, whether it's WordPress, or Drupal, or other systems.

All things considered, Drive links have a likelihood of being stable more than many other alternatives. If people really cared, I suppose the another good alternative would be archiving them in a git repo. When you push it out to some git server, such as github or gitlab, it's still not guaranteed to be stable (both of those are companies that could go away, or the user's account where the git repo was hosted could go away for any number of reasons), but at least it would be easier for multiple copies of the archive to be easily replicated. Does the Internet Archive support archiving git repos for posterity? Maybe it's something they should consider.

The TAB report on the UMN affair

Posted May 7, 2021 15:48 UTC (Fri) by mathstuf (subscriber, #69389) [Link]

> Does the Internet Archive support archiving git repos for posterity? Maybe it's something they should consider.

I don't know about IA specifically, but LWN has had articles on archival efforts before. Found just one, but maybe the others were just Brief mentions.

https://lwn.net/Articles/693471/ "Preserving the global software heritage"

The TAB report on the UMN affair

Posted May 7, 2021 16:26 UTC (Fri) by calumapplepie (guest, #143655) [Link]

Github archived all their (active) git repositories (as of Febuary 2020) underneath a mountain: see https://archiveprogram.github.com/ .