A link archive 1.0 Background line 14 1.1 The manually curated link collection line 34 1.2 The demise of link collections line 42 2.0 Using archive.org to stabilize a link collection line 58 3.0 Experiences line 70 I have wondered whether there is a way for a curated link archive, in the spirit of the early days of the WWW, to be useful at the time of writing. This is an experiment where I tried to use archive.org links to stabilize the link targets. 1.0 Background A functional information network provides three core services: 1: Means of sharing. 2: Means of discovering what is shared. 3: Means of providing feedback. The early days of the web provided (1), built (3) on top the fact sharing an e-mail address was a lot less foolhardy and therefore quite common back then and used home pages, link collections, and so on to try and solve (2). Currently, open systems provide both means of sharing and means peer to peer communication, but discoverability is firmly in the grasp of search giants and walled gardens. An index is by its very nature a monolithic beast; dividing an index into smaller parts both requires domain expertise when splitting it as well as the greater problem the user must now know which index to ask. In other words, web search is extremely friendly to monopolies. 1.1 The manually curated link collection From home pages and early attempts at identifying different subsets of the web grew the larger, curated link collections, most famously Yahoo. The growth of Yahoo in that period should be ample demonstration these collections were considered useful resources. 1.2 The demise of link collections There might be no single factor which killed off the popularity of manually maintained link collections. Some factors may be: - The explosive growth of the web itself making the collections outdated at an accelerating pace. - The lack of a way to make them scale without adding manpower. - The increasing quality of search engines. - Single page applications and other hindrances to deep linking, like walled gardens in general. - Whenever a major site changed content management system, large parts of a collection could, and often would, become dead links, as well as other sources of link rot. 2.0 Using archive.org to stabilize a link collection Instead of linking directly to the original resource, this experiment always uses an indirection through archive.org's Wayback Machine. Using links to archive.org is an attempt to avoid long term link rot. Also, if what is available through archive.org is insufficient, the actual resource may be available (and updated), or the data on archive.org may provide sufficient bread crumbs to locate the original or new resources regarding the same problem. The problem of scaling with the web really does not interest me at all. :) 3.0 Experiences I chose a topic for storing a few links, which by experience were hard to dig up and could be quite hard to find again given what seems to be a continous degradation of the quality of web search. After more than a year, what I have is a pretty nice, little bookmarks file, with short comments. The links work, the information is retrievable, but I have become utterly disillusioned in regards to this approach. First of all, archive.org has been sued. Trusting information integrity to the litigious nature of the US public life seems like a fundamentally bad idea. Second, archive.org is still yet another centralized service. The institution of the library is one of civilization's greatest success stories, while the demise of the library of Alexandria is remembered as a tragedy millennia after it burned. For bookmarks into a cache to be reliable and trusted, the cache must be distributed and replicated. The bookmarks must be tended regularly, like a garden, to avoid it becoming stale and divorced from possible live, maintained copies. The goal was discoverability of useful resource, not pure archival work. In other words, my conclusion is that we ideally would have multiple public digital libraries. I think that is a good idea in general, but perhaps something could be done on a smaller scale? One alternative might be topical archives maintained by smaller organizations? I don't really know where to go from here, I just wanted to share my rather mixed success and reality check. Edited 20220218T152539Z. Steinar Knutsen, 20231001T183425Z, 8E0DDAAF