Citation semantics in HTML

Dear Sir or Madam,

Re: Citations in HTML

I'm working on an HTML editor frontend for a distributed authoring and versioning application to implement a read-write web.

Forgive me for addressing this email so generally. I was unable to find a chapter for the United Kingdom. I'm writing because I have an idea for a future version of HTML that may help uphold democratic values as well as improve academic and research publishing.

Here's my idea:

Supporting semantic citation data for the <cite> tag would have a number of advantages. It would:

1) improve reliability of information on the world wide web
2) improve the accessibility of reliable information for academic and research purposes
3) enable GPT-type AI implementations to properly support their responses with citations (this is a real problem for people using GPT)
4) encourage a more thorough approach to publishing verifiable information in a media culture so often concerned with impact over truth.

Given that referencing standards already exist and are relied on heavily in academia, (e.g. Harvard, Chicago), would not the inclusion of citation data be feasible in a future specification?

An example implementation might be as follows:

<p id="introduction"> In
  <cite
      authors="Berners-Lee, T. and Fischetti, M."
      date="1999"
      title="Weaving the Web : the past, present and future of the World Wide Web by its inventor"
      publisher="London: Orion Business."
    >
      Weaving the Web : the past, present and future of the World Wide Web by its inventor (Berners-Lee, T. and Fischetti, M., 1999)
  </cite>,
the world wide web was originally envisaged as a creative space for collaborative writing, editing and dissemination of information, marked up semantically and structurally to both format and categorize the parts of a document/text according to meaning and the context of the content.</p>

Or, perhaps:

<p id="abstract"> Our analysis of the dataset &quot;
  <cite
      authors="Tatman, R."
      date="2017"
      title="Every Pub in England"
      location="online"
      address="https://www.kaggle.com/datasets/rtatman/every-pub-in-england"
      accessed="July 03, 2024"
    >
      Every Pub in England
  </cite>
  discovered at least 2 ancient establishments that may date to the Dark Ages, but all important
  historical information about has been lost due to historians not making use of semantic web
  technologies to allow machines to process, store and disseminate information to humanity's benefit.
</p>

With the above citation attributes, it would be possible to do this:

<footer>

<p>See also:
<ol id="bibliography">
</ol>

<script>

  const bibliography = document.querySelector( 'ol#bibliography');
  const sources = document.querySelectorAll( 'cite');

  sources.forEach( el => {
    const li = document.createElement('li');
    li.innerHTML = `${el.authors} (${el.date}) <em>${el.title}</em>. ${el..publisher}.`;
    bibliography.appendChild(li);
  });

</script>

</footer>

Hence, the information in the citations is readily usable for further computational purposes and for user consumption.

A corollary to this approach is then possibility of a new tag <bibliography> that would be automatically populated with the citation data in the page content.

<bibliography for="introduction abstract" />

A single "for" attribute would allow a space-separated list of content elements by id that should be parsed for the population of the bibliography, producing:

- Berners-Lee, T. and Fischetti, M. (1999) Weaving the Web : the past, present and future of the World Wide Web by its inventor. London: Orion Business.
- Tatman, R. (2017) Every Pub in England. Kaggle.com. [Online] [Accessed on 3rd July 2024] https://www.kaggle.com/datasets/rtatman/every-pub-in-england.

There are a great many possibilities for making better use of citation metadata. It would be possible to have numbered, superscript-style linking such that inline references link to footnote bibliography entries much as in Wikipedia.

A great amount of the user's markup work can be automated. The parsing of URL schemes and namespaces in the user's citation data would make citation markup easier,. For example, URN schemes such as urn:isbn could be used for book citations, or DOI for papers. A Javascript Citations API would be able to recognise when the user is referencing a particular type of source, making use of namespaces, and formatting citations properly by way of a Citation Builder.

The Citations API would support citation indexing, summarisation and management of citations, as custom Javascript classes with properties such as title, date, last-accessed, authors, etc..

There's a world of opportunity here, and like the lost historic pubs of yesterday, how much knowledge is being lost by our not leveraging our very best technology? Moreover, how much is democracy being abused and undermined by fake news and bogus, unverifiable information because we can't build a web of trust with insufficient tracing of sources?

Kind Regards,

David Philip Greenwood