- From: W3C Community Development Team <team-community-process@w3.org>
- Date: Wed, 1 Oct 2025 17:02:23 +0000
- To: public-tdmrep@w3.org
The chair described the status of the IETF AI-Pref initiative, which is starting a 3-day meeting in Zurich. Links to the current public drafts had been circulated before; they are: - https://datatracker.ietf.org/doc/html/draft-ietf-aipref-vocab-03 - https://datatracker.ietf.org/doc/html/draft-ietf-aipref-attach-03 Although there is no assurance that the IETF group will succeed in gaining consensus on such a vocabulary, a favorable outcome would be beneficial for TDMRep, as our specification can easily evolve to include such signals. As a reminder, the `tdm-reservation` signal indicates that the copyright owner reserves his rights for any processing of "his" content via TDM techniques. It is not a NO-TDM signal, but a RIGHTS-RESERVED signal. The complementary `tdm-policy` URL directs users to an ODRL file, which outlines the procedures for obtaining mining rights and provides contact information for the copyright owner. The chair outlined a proposal for the evolution of TDMRep, incorporating a fine-grained vocabulary. It is based solely on the evolution of the TDM Policy document, hosted on a Web Server. The proposal will soon be made available on the TDMRep W3C CG Page (https://www.w3.org/community/tdmrep/). The significant advantage of this solution is that the large volume of publications already released with embedded `tdm-reservation` and `tdm-policy` properties do not need to be modified in any way. Additionally, no modification of TDMRep implementations is required in HTTP headers or the server-hosted tdmrep.json file. Only the ODRL file, which is referenced by `tdm-policy`, needs to be modified if the provider wants to provide fine-grained signals; this file is typically unique for a given provider. One reason for adopting fine-grained signals may be for Web Search management. Search actors firmly state that modern Search engines rely on AI in their processing pipelines. As AI pipelines include TDM techniques (tokenization is considered by most people to be a TDM technique), reserving all TDM Rights could imply that Search engines need to obtain an agreement from content providers before indexing their content. TDMRep CG members advocate that robots.txt is an effective way to signal web crawlers an opt-in for Search, and therefore, constitutes a complementary solution for the TDM opt-out provided by TDMRep. However, this complementarity is not specified in the TDMRep document and remains subject to interpretation. Therefore, an explicit signal, free from specific bot names, could be more secure for both copyright owners and search actors. In this case, display constraints like 'no more than 100 characters" — as proposed by Microsoft in the IETF initiative — would be a good addition to the Search opt-in. The present members of the TDMRep CG decided during the call to work on a proposal for a vocabulary that covers the inference side of the problem currently being discussed at IETF. It may be proposed as an RFC. The group had previously decided to examine the outcome of the IETF initiative before determining a path for completing and standardizing TDMRep. The IETF initiative is taking more time than initially expected. Still, we'll follow it until the end of 2025 before deciding whether to move forward with our own set of fine-grained signals or rely on an AI-Pref consensus. ---------- This post sent on TDM Reservation Protocol Community Group 'Notes, September 30th, 2025' https://www.w3.org/community/tdmrep/2025/10/01/notes-september-30th-2025/ Learn more about the TDM Reservation Protocol Community Group: https://www.w3.org/community/tdmrep
Received on Wednesday, 1 October 2025 17:02:23 UTC