HN Explorer

CiteSeerX has broken all paper links

3 points by Mathnerd314 ↗ HN
For example http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.219.7269&rep=rep1&type=pdf "Modern B-Tree Techniques" from https://news.ycombinator.com/item?id=15409757 no longer works and says "This page could not be found". It seems from searching for the paper that they are using a new scheme based on hashes, so for example https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=0b19f413ffb5bc68b43f3bd05a97c282a7c6d6ab, but it is quite frustrating that the old DOIs do not work.

3 comments

[ 0.20 ms ] story [ 19.8 ms ] thread
jamessb
For many entries (like your example), what was listed as the DOI was not actually a valid DOI, but rather some other identifier that is presumably specific to citeseerx.

A valid DOI includes a slash: the part before the slash indicates the registrant, and the part after it is the suffix that indicates which item managed by the registrant is actually referred to [1].

[1]: https://www.doi.org/doi_handbook/2_Numbering.html#2.2