internetarchive / wayback-machine-versions Goto Github PK

Based on prior interactions about this effort with @markjohngraham , I must admit that I don't understand the rationale for diverging the approach from the Robust Link specification (http://robustlinks.mementoweb.org/spec/), which it previously aligned with.

The Robust Link specification is a major result from a 3-year Mellon funded project that quantified link rot and content drift at unprecedented scale (see papers https://doi.org/10.1371/journal.pone.0115253 ; https://doi.org/10.1371/journal.pone.0167475) and explored approaches to ameliorate their negative impact (see the analysis that led to the Robust Links spec at http://mementoweb.org//missing-link/)

Concerns with the here proposed approach can be summarized as follows:

1. Semantics - The semantics of Robust Links are very clear: each link leads to an archived snapshot that was captured at/around the datetime of linking and hence provides fallback mechanisms to revisit the linked resource as it was at the moment the link was put in place. The semantics of links to multiple archived copies created at many different datetimes, on the other hand, as presented in this approach, are unclear. For example, what is the (difference in) meaning of each link? How many links to archived copies should one include? Which archived copies should be linked to, the last 3, the first 7, a random set of 11? How does this scale for popular and (hence) frequently archived pages such as cnn.com?

2. Functionality - When clicking one of the links in the menu to an archived copy, as currently presented in this approach, one is lead to the archived resource in the Wayback Machine. The IA banner shown there provides access to all other archived copies for the linked resource. Why would a user return to the link menu to click one of the other linked archived copies listed there when all available snapshots are presented right there in the IA banner? So simply from the usability point of view, it is not clear what added value the multiple links bring.

3. A (maybe minor) technical concern is about using a comma as the delimiter of URIs since a comma is an allowable character in a URI.

Note: This issue was discussed prior to posting with @hvdsomp , @phonedude, @hariharshankar who were instrumental in devising the Robust Links spec.