12:58:21 <RRSAgent> RRSAgent has joined #pmwg
12:58:25 <RRSAgent> logging to https://www.w3.org/2026/06/04-pmwg-irc
12:58:25 <Zakim> RRSAgent, make logs Public
12:58:26 <Zakim> Meeting: Publishing Maintenance Working Group
12:58:29 <wendyreid> chair: wendyreid
12:58:33 <wendyreid> date: 2026-06-04
12:58:55 <GeorgeK> present+
12:59:10 <wendyreid> present+
12:59:39 <MasakazuKitahara> MasakazuKitahara has joined #pmwg
12:59:53 <shiestyle> shiestyle has joined #pmwg
12:59:56 <MasakazuKitahara> present+
12:59:57 <toshiakikoike> toshiakikoike has joined #pmwg
13:00:07 <toshiakikoike> present+
13:00:41 <shiestyle> present+
13:01:01 <AvneeshSingh> AvneeshSingh has joined #pmwg
13:01:02 <kimberg> kimberg has joined #pmwg
13:01:14 <sueneu> sueneu has joined #pmwg
13:01:21 <kimberg> present+
13:01:23 <AvneeshSingh> present+
13:01:29 <sueneu> present+
13:02:15 <gman> gman has joined #pmwg
13:03:02 <AvneeshSingh> AvneeshSingh has joined #pmwg
13:03:02 <toshiakikoike> toshiakikoike has joined #pmwg
13:03:02 <shiestyle> shiestyle has joined #pmwg
13:03:02 <wendyreid> wendyreid has joined #pmwg
13:03:02 <GeorgeK> GeorgeK has joined #pmwg
13:03:02 <hadleybeeman> hadleybeeman has joined #pmwg
13:03:02 <mgarrish> mgarrish has joined #pmwg
13:03:41 <mgarrish> mgarrish has joined #pmwg
13:03:41 <AvneeshSingh> AvneeshSingh has joined #pmwg
13:03:41 <toshiakikoike> toshiakikoike has joined #pmwg
13:03:41 <shiestyle> shiestyle has joined #pmwg
13:03:41 <wendyreid> wendyreid has joined #pmwg
13:03:41 <GeorgeK> GeorgeK has joined #pmwg
13:03:41 <hadleybeeman> hadleybeeman has joined #pmwg
13:03:55 <DaleRogers> DaleRogers has joined #pmwg
13:04:08 <DaleRogers> present+
13:04:32 <Hadrien> Hadrien has joined #pmwg
13:04:55 <Hadrien> present+
13:04:57 <ajellinek> ajellinek has joined #pmwg
13:04:57 <mgarrish> mgarrish has joined #pmwg
13:04:57 <AvneeshSingh> AvneeshSingh has joined #pmwg
13:04:57 <toshiakikoike> toshiakikoike has joined #pmwg
13:04:57 <shiestyle> shiestyle has joined #pmwg
13:04:57 <wendyreid> wendyreid has joined #pmwg
13:04:57 <GeorgeK> GeorgeK has joined #pmwg
13:04:57 <hadleybeeman> hadleybeeman has joined #pmwg
13:04:58 <CharlesL> CharlesL has joined #pmwg
13:05:12 <Amanda_at_eBOUND> present+
13:06:21 <CharlesL> CharlesL has joined #pmwg
13:06:21 <ajellinek> ajellinek has joined #pmwg
13:06:21 <mgarrish> mgarrish has joined #pmwg
13:06:21 <AvneeshSingh> AvneeshSingh has joined #pmwg
13:06:21 <toshiakikoike> toshiakikoike has joined #pmwg
13:06:21 <shiestyle> shiestyle has joined #pmwg
13:06:21 <wendyreid> wendyreid has joined #pmwg
13:06:21 <GeorgeK> GeorgeK has joined #pmwg
13:06:21 <hadleybeeman> hadleybeeman has joined #pmwg
13:06:22 <LaurentLM> LaurentLM has joined #pmwg
13:06:31 <LaurentLM> present+
13:07:24 <wendyreid> https://pad.w3.org/p/PMWG_Minutes_2026-06-04
13:09:30 <shiestyle> scribe+
13:09:30 <shiestyle> scribe+ sueneu
13:09:34 <shiestyle> q?
13:11:01 <gman> present+
13:11:12 <shiestyle> Topic: Rechartering
13:11:36 <shiestyle> https://w3c.github.io/publ-maintenance-wg-charter/
13:12:55 <shiestyle> https://github.com/w3c/publ-maintenance-wg-charter/issues/58
13:15:19 <AvneeshSingh> q+
13:15:25 <wendyreid> ack AvneeshSingh
13:15:31 <shiestyle> Matt: usually there is something in the charter that allows things from the CG to come through if it is sufficiently ready for spec. It might be implied, and allowed within the W3C process. We don't need to follow every group's process.
13:16:11 <shiestyle> s/Matt/mgarrish/
13:16:26 <wendyreid> https://w3c.github.io/publ-maintenance-wg-charter/#coordination
13:16:35 <shiestyle> wendyreid: We have a line about that in the 3.4 section, we could bring it forward or move it to a section for all versions
13:17:38 <shiestyle> AvneeshSingh:  Gautier asks that we should use incubation from the CG, Ivan was saying it was obvious. But accessibility TF needs to be considered. Although Gautier and Ivan are not here to clarify their positions.
13:19:18 <shiestyle> wendyreid: We have in section 5, we have text indicating we should work with the CG. Maybe this is where we should change the wording. Maybe we need something more open like we will accept items that reach an appropriate level of incubation
13:19:20 <mgarrish> q+
13:19:25 <wendyreid> ack mgarrish
13:21:02 <Hadrien> q+
13:21:03 <shiestyle> mgarrish: What is a specific incubation going to accomplish for 3.5, since we aren't planning a long revision. If there is something specific, we could put it in the charter. We could take on any proposal that is brought up by the group, even if it ends up in a note. I'm not clear what we should put in the charter specifically
13:21:19 <wendyreid> ack Hadrien
13:23:02 <LaurentLM> LaurentLM has joined #pmwg
13:23:02 <CharlesL> CharlesL has joined #pmwg
13:23:02 <ajellinek> ajellinek has joined #pmwg
13:23:02 <mgarrish> mgarrish has joined #pmwg
13:23:02 <AvneeshSingh> AvneeshSingh has joined #pmwg
13:23:02 <toshiakikoike> toshiakikoike has joined #pmwg
13:23:02 <shiestyle> shiestyle has joined #pmwg
13:23:02 <wendyreid> wendyreid has joined #pmwg
13:23:02 <GeorgeK> GeorgeK has joined #pmwg
13:23:02 <hadleybeeman> hadleybeeman has joined #pmwg
13:23:40 <CharlesL> q+
13:23:45 <wendyreid> ack CharlesL
13:23:50 <shiestyle> Hadrien: I will do my best to have something for the week of the 15th. Another thing of interest, we didn't keep up with different multimedia formats in the last charter. It wouldn't hurt to have  a sentence about keeping up with image and audio formats.  Google mentioned EPUB in a bnlog post about jpegXL.
13:24:41 <Hadrien> https://opensource.googleblog.com/2026/06/journey-to-jpeg-xl-how-open-source-experiments-shaped-the-future-of-image-coding.html
13:24:47 <shiestyle> CharlesL: Do we need to call out anything regarding AI? New metadata to support AI content or contribution?
13:24:53 <GeorgeK> present+Q+
13:25:19 <shiestyle> wendyreid: We have open issues about AI authoring. We will have to have that discussion. and talk about aligning with other standards.
13:26:40 <shiestyle> GeorgeK: right now we have capabilities with screen readers and ai describing an image. Is that relevant to us? And when AI can answer questions about the book you're reading. Like a summary that is AI generated, or language simiplfication, or talking to a bot about understanding the content
13:27:01 <Hadrien> q+
13:27:26 <wendyreid> ack Hadrien
13:29:15 <shiestyle> wendyreid: We have little control over this, as they are reading system or screen reader features.
13:29:15 <shiestyle> The only thing that comes to mind is TDM, which is technically under our control. Can a publisher block those kinds of features
13:29:43 <shiestyle> s/The only/... The only/
13:29:48 <wendyreid> s/technically under our control/not technically under our control/
13:31:10 <AvneeshSingh> q+
13:31:28 <wendyreid> ack AvneeshSingh
13:31:58 <shiestyle> Hadrien: I haven't tried the latest updates to AI image descriptions, it is initially in the absence of alt text. On some devices it can do this locally and some online. I don't think there is much we can control, since it is triggered by the platform. It is more sophisticated now because you can get information beyond the description. It's an interesting question, its not necessarily covered by TDM since we aren't talking about trai
13:31:58 <shiestyle> ning a model, but something that happens in real time. It is raising questions and concerns on the publisher side. It is worth considereing. I think we will increasingly have these features at the system level. Maybe we could have impact on reading system features with more metadata. It is dangerous for us to invent new things, we should look at what other groups are doing
13:32:26 <shiestyle> s/trai/traing/
13:32:39 <shiestyle> s/ning/.../
13:33:01 <shiestyle> AvneeshSingh: AI is more of a busineess group discussion, business models, copyright issues. Maybe we add a line that we are coordinating with the business group and cg. It would give us room to  observe and do what is required.
13:33:52 <shiestyle> ... for media overlays, in the coordination group we should also have synced media. We should coordinate with them, since everyone esp from Daisy may not be able to join the working group but can join the CG
13:35:00 <shiestyle> ... in the impact W3C on the world, we are having a discussion to bring charters to conider their impact statements about what kind of impact this charter is trying to have in the world.
13:35:14 <shiestyle> wendyreid: It might be nice to have this discussion, and ground people into why we propose some things over another.
13:35:30 <shiestyle> ... Any other thoughts? Advancements to current features? New features?
13:35:58 <Hadrien> q+
13:36:00 <GeorgeK> q+
13:36:07 <DaleRogers> q+
13:36:15 <wendyreid> ack Hadrien
13:36:16 <shiestyle> sueneu: What about TTS? It's not something we'll drive on our own, I'd like to leave open that we address it, it is getting more important to the community, I don't want to lock us out of doing anything.
13:38:05 <shiestyle> Hadrien: It depends on what we want to do, the audio community group is considering incubating this and including it in a future charter. At first this was speech to text, but there is interest now in text to speech. It would be interesting to stay in touch with this, we may be implementers of web speech API
13:38:44 <wendyreid> ack GeorgeK
13:38:47 <wendyreid> q+
13:38:52 <shiestyle> ... I am concerned about the ways we try to provide information to TTS information. The proposed ones are currently useless. We should be selective about the work we choose to follow.
13:39:11 <shiestyle> ... the web site of this is worth tracking.
13:40:35 <Hadrien> q+
13:40:58 <shiestyle> GeorgeK: one of our epub test lists is readaloud. We test if TTS is synchronized. We question if alt text should be included in readaloud, and mathml. Lots of questions that we will see in OSLO, should the speed slow down for math? Should there be pauses?
13:41:02 <wendyreid> ack DaleRogers
13:41:11 <shiestyle> ... I know Gautier is looking atht CG in readaloud and language shifts. Should we try to model what happens in a human narrated audio book? Lots of questions in this area about best practices. I'm not sure this is in scope for this working group.
13:41:51 <AvneeshSingh> q+
13:41:57 <wendyreid> ack wendyreid
13:41:59 <GeorgeK> +1 to what Dale said
13:42:17 <shiestyle> DaleRogers: In higher ed, we asked if we are preparing our students for technologies that exist today, because we don't know what is coming 5 years down the road? Can there be a broad, open statement that we will monitor updcoming technologies
13:43:41 <shiestyle> wendyreid: There is a certain amount of flexibility we can introduce, but we can't make the staement too broad. We have to be specific in the charter partly for intellectual property protections. We could be more specifc about which technologies we are watching
13:45:16 <shiestyle> ... I've reached out to the APA and said we are willing to work with other groups on TTS, but there is no clear single home for this. We have to have a conversation about the specs that are out there now, are over 10 years old, without widespread implementation.
13:45:34 <wendyreid> ack Hadrien
13:45:48 <shiestyle> ... A new approach is needed, but who will recommend it? There is a lot of overlap with publishing, news online, media. I don't know who is owning spoken-html.
13:46:16 <shiestyle> ... we could put something in the charter about TTS since it is on our minds.
13:47:58 <shiestyle> Hadrien: there is no one really in charge of TTS, but the audio group is looking to go to spec. What is important to point out is to use machine learning for voices, the more natural they are, the less control we have. Even if this information is in epub files, you can't realy use it.
13:48:50 <wendyreid> ack AvneeshSingh
13:49:08 <shiestyle> ... about readaloud and conformance, we don't say much about specific settings like margins, do we want to do so for readaloud? Maybe this is a best practices document or minimum requirements. If we are very broad I think they could go in the spec, but it would be weird to do readaloud in more detail.
13:50:33 <gman> q+
13:50:37 <wendyreid> ack gman
13:50:43 <shiestyle> AvneeshSingh: Our charter is already broader and has incubation features not ususally found in W#C documents. Incubations should happen in the CG and when they are mature enough, they move to the working group. So we don't need to put this in the charter. We need to move more of this work to the CG, where we can get wider participation.
13:52:47 <shiestyle> gman: The AI concerns are more about business models, but we could delve into this deeper if we want. It depends on what we consider appropriate in the continuum of paramnoia. Data is worth more than gold nowadays specifically for organizations who build AI.
13:53:25 <shiestyle> ... Should we trust them to receive data like image descriptions, or is it still on their servers? Will they use whatever they want to train their models? For me, do we trust them or not trust them. Wee could discuss this more broadly.
13:54:37 <shiestyle> wendyreid: This has even been a discussion within TDM, since it is a statement but it is not technically binding. Even though there may be legal consequences.
13:54:59 <shiestyle> ... we should as the business group to have more of a discussion on AI. I comes down to what publishers want to do about it. Do we need metadata about what can be used for training or and what can't
13:55:35 <shiestyle> gman: maybe there should be a key
13:57:46 <shiestyle> wendyreid: (minutes lost)
13:58:07 <Hadrien> +1 for skipping the call next week, too many people traveling
13:58:16 <shiestyle> wendyreid: We have a few more ideas to add to the charter, these are all good points. We are not having a meeting next week because of the Digital Publishing Summit.  We will reconvene in two weeks. Hopefully we will have updates about horizontal reviews.
13:58:24 <CharlesL> CharlesL has left #pmwg
14:00:27 <shiestyle> s/(minutes lost)/Copyright is also complex as there is an accessibility angle and many of the features mentioned earlier have accessibility implications. In many countries copyright can be waived for accessibility reasons, but then it raises the question of what is an accessibility feature and who uses it, which is thorny./
14:00:45 <wendyreid> rrsagent, please make minuts
14:00:45 <RRSAgent> I'm logging. I don't understand 'please make minuts', wendyreid.  Try /msg RRSAgent help
14:00:48 <wendyreid> rrsagent, please make minutes
14:00:49 <RRSAgent> I have made the request to generate https://www.w3.org/2026/06/04-pmwg-minutes.html wendyreid
14:01:19 <wendyreid> scribe+ shiestyle
14:01:27 <wendyreid> rrsagent, please make minutes
14:01:28 <RRSAgent> I have made the request to generate https://www.w3.org/2026/06/04-pmwg-minutes.html wendyreid
14:02:21 <wendyreid> rrsagent, bye
14:02:21 <RRSAgent> I see no action items