DOI registrant using digital fingerprints of JATS files

Content Registration
,
1

I am exploring a highly automatable process that can be used by a DOI registrant. I suspect this process is novel and enables a new kind of DOI registrant **. For lack of better term, I’ll refer to this process as that of a “fingerprinter”. I use this term in the sense of a digital fingerprint [1]. My hunch is Crossref is the right registration agency for a fingerprinter.

Before I get too far documenting this process, I wanted to ask whether anybody is aware of similar efforts or documents I should read?

The object identified by a fingerprinter is an automatic procedure that uses a changing digitally recorded history of declarations by authors about digital scholarly documents. At a given point in time, this procedure resolves a fingerprint-based DOI link to a JATS XML file. The procedure also identifies rich information about a history of declarations by authors about JATS XML files. As the digital history of declarations by authors changes, so to does the exact digital JATS XML file resolved.

I would like to be a DOI registrant to develop and test this new fingerprinter process hosted on popgen.es using some of the scholarly documents that I write (e.g. :castedo.com/doc). I hope some of my braver colleagues will test this out with me too. I think it might lead to important insights and open-source code that future DOI registrants will want to use.

In the particular implementation I would like to test, the digital scholarly document is a JATS XML file. The digital identification of authors is via ORCID. The digitally recorded history of declarations by authors is done with cryptographic hashes [2]. Cryptographic strength is preferable but not a hard requirement.

Does this ring any bells for anybody? Anybody thinking of something similar?

I plan to post more details about this and ask about testing this idea out with a new DOI prefix.

Thank you,
Castedo Ellerman

** The closest discussion I have seen to this topic is [3].

[1] :en.wikipedia.org/w/index.php?title=Fingerprint_(computing)&oldid=1084076041

[2] :en.wikipedia.org/w/index.php?title=Cryptographic_hash_function&oldid=1077451766

[3] :pidforum.org/t/creating-resolvable-pids-without-registering-them-in-a-registry-through-cids-and-ipfs/1165

1 Like
2

This seems related to using content addressing, e.g. work by Ben Trask where a cryptographic hash of the content (e.g., md5, sha1) is the identifier. Resolution of the identifier was left open to any service anyone wanted to set up, Trask himself set up hash-archive-org where you could discover where content with a given hash could be retrieved. This service seems to be offline at the moment.

1 Like
3

For the love of God can someone on this site please fix the “no links in post” problem! Instead of ham-fisted rules to prevent spam, how about trusting people?

1 Like
4

Hi @rdmpage ,

We updated our security settings in January 2022 because of the amount of spam we were having to review each day. Prior to that there were no restrictions on posts in this space.

Unfortunately, the lack of basic security settings was taxing for our staff - to review and determine if all messages, and specifically the links accompany those messages, were legitimate/inoffensive. We now have some basic settings in place to prevent spam. Those settings have prevented almost all of the spam we were seeing. We understand there are tradeoffs when decisions like this are made. For now, we’re planning to retain the basic settings, so our staff and visitors to the community forum, as a whole, are not subjected to spam.

Thank you,
Isaac

5

@ifarley thank you for the quick reply, I understand that spam levels can make fora like this a nightmare without sometimes aggressive rules in place. Does this forum use the askimet plugin for Discourse? I’d send you the link. but…

6

Thanks for the suggestion for the askimet plugin. We’ll take a look. You still can’t send links @rdmpage ? Your account has met the minimum threshold to do so.

7

Thanks again for this recommendation @rdmpage . According to the Akismet plugin documentation: ‘For sites that are on our hosting, the Akismet plugin is pre-installed and enabled for you.’ This community forum is hosted by Discourse, so we’ve been using the plugin since we launched community.crossref.org. Unfortunately, the plugin has not eliminated all spam, so we have some additional measures to protect against spam in the forum.

-Isaac