00:09:30 Josh Moore: Then Aybuke 00:09:38 Josh Moore: Then Brian 00:10:03 Josh Moore: Then Caterina 00:10:13 Josh Moore: Then Christian 00:10:40 Josh Moore: Then Damir 00:10:50 Josh Moore: Then Dan 00:11:14 Josh Moore: Then Dave 00:11:39 Josh Moore: Then Eric 00:11:53 Josh Moore: Then Fernando 00:12:12 Josh Moore: Then the HHMI crew 00:12:35 Josh Moore: Then Jesse 00:12:51 Josh Moore: Then Ken 00:13:10 Josh Moore: Then Kiya 00:13:36 Josh Moore: Then Koji 00:13:44 Josh Moore: Then Lucy 00:13:58 Josh Moore: Then Matt 00:14:09 Josh Moore: Then Melissa 00:14:21 Josh Moore: Then Nathan 00:14:31 Josh Moore: Then Nick 00:15:05 Josh Moore: Then Norman 00:15:30 Josh Moore: Then … “samee” 00:16:48 Josh Moore: Will, Wouter, then Ziwen (Wouter isn’t last this time ;) ) 00:18:40 Josh Moore: Links from my side (1) Zarr V3 https://github.com/orgs/zarr-developers/projects/2 (2) OME / Zarr-java https://zarr.dev/blog/ome-2022/ 00:21:20 Josh Moore: josh@openmicroscopy.org 00:26:10 Josh Moore: Apologies, I forgot to mention: Live notes for the session are available in https://hackmd.io/BqnK9Wm4QpGYAhYOoaFBQQ Where possible, help to structure the notes for later publication rather than commenting in Zoom’s chat. Thanks! 00:35:57 Aybuke K Yoldas (EMBL-EBI): Just to clarify: REMBI is guidelines for metadata for imaging datasets, developed by cross-community working groups. BioImage Archive has adopted and implemented it - as it offers metadata guideline for many different imaging modalities - which we are a repository of. 00:43:24 Norman Rzepka: OME-Zarr needs a logo! 00:44:21 Josh Moore: Reacted to "OME-Zarr needs a log..." with 👍🏽 00:44:35 Caterina Strambio De Castillia: Reacted to "OME-Zarr needs a log..." with 👍🏼 00:47:11 Damir Sudar: +1 Matt 00:47:15 Dave Mellert: I +1 the idea of a minimal metadata spec with the ability to add one's own scheme. 00:47:37 Aybuke K Yoldas (EMBL-EBI): Agreed there needs to be minimal metadata that we can use to make sense of the image. 00:47:38 Kenneth Ho: +1 Dave 00:47:46 Davis Bennett (he/him): I’m guessing that many of these existing metadata standards have attempted to find something common to all imaging modalities 00:47:53 Josh Moore: Replying to "Agreed there needs t..." but _whose_ minimal? 🙂 00:47:53 Dave Mellert: but minimal really needs to be minimal 00:48:12 Aybuke K Yoldas (EMBL-EBI): Replying to "Agreed there needs t..." Mine 😛 00:48:13 Dave Mellert: I think Juan said it well (I'm looking back at the morning session notes) 00:48:27 Kenneth Ho: Minimal means useful ;-) 00:48:29 Josh Moore: Replying to "I think Juan said it..." Yes, there was the suggestion of a t-shirt ;) 00:49:39 Dave Mellert: one other think to think about, at JAX we think a lot about Systems of Record and who should own the metadata about a thing. e.g., I wouldn't ever pack sample metadata into an image file if the sample is captured in another database 00:51:15 Davis Bennett (he/him): Do people really think we can take N existing metadata standards and distill them to 1, with the total acceptance of the authors of those N formats? 00:52:16 Nathan Hotaling: @Dave actually I had that question as well. Does all metadata need to be embedded. Especially in the clinical world all metadata shouldn’t always be embedded in the file. 00:53:47 Damir Sudar: Doesn't the linked data concept solve that issue? Some metadata can/should be embedded and some of the metadata can/should exist in other repos. 00:53:54 Nathan Hotaling: I would also advocate for moving away from xml to a more modern web format, i.e. json/yaml 00:54:04 Dave Mellert: JSON-LD has been discussed 00:54:12 Damir Sudar: Reacted to "I would also advocat..." with 👍 00:54:21 Caterina Strambio De Castillia: Just a word of warning that while the idea of minimal is certainly attractive it is actually not as easy as it might appear… Deciding what is minimal is not at all trivial and it depends very much on imaging modality etc. 00:54:23 Damir Sudar: Reacted to "JSON-LD has been dis..." with 👍 00:54:30 Kenneth Ho: @Davis, we have different manufacturers using all sorts of different terms to describe their systems in their proprietary formats. Those metadata become useless. Therefore, there is a need to come to some consensus of common terms used and definition. 00:54:50 Caterina Strambio De Castillia: Reacted to "@Davis, we have diff..." with 👍🏼 00:55:45 Dave Mellert: but why not point to a unique identifier for e.g., an instrument that is captured in an ontology external to the OME-NGFF scheme? 00:55:50 Nathan Hotaling: Reacted to "@Davis, we have diff..." with 👍🏼 00:55:58 Nick Schaub: So maybe metadata becomes something like they do in ML with onyx? Basically, let everyone define their own metadata, and have a common translation format. 00:56:04 Caterina Strambio De Castillia: Also we have to keep in mind that one of the roles of metadata is to ensure reproducibility and reuse and therefore see level of consensus is needed 00:56:22 Aybuke K Yoldas (EMBL-EBI): Reacted to "Also we have to keep..." with 👍 00:57:00 Nick Schaub: Oops, I mean ONNX 00:57:18 Josh Moore: Replying to "one other think to t..." but a link to? and if both the pointer and the pointee are RDF does that help? 00:57:23 Nathan Hotaling: @nick isn’t that why pick up of ONNX is pretty low? That flexibility makes it a nightmare to map anything you are doing to what some one else is doing 00:57:38 Dave Mellert: @Josh you know I am all in on the RDF bandwagon 00:57:51 Josh Moore: Replying to "Do people really thi..." Did I cover this sufficiently? tl;dr — I don’t think anyone is proposing this. 00:58:05 Josh Moore: Reacted to "Doesn't the linked d..." with 👍🏽 00:58:08 Norman Rzepka: Reacted to "Did I cover this suf..." with 👍 00:58:15 Davis Bennett (he/him): Replying to "Do people really thi..." “Create a minimal set of metadata” sounds a lot like this 00:58:17 Nick Schaub: Replying to "@nick isn’t that why..." Of course it is, but its better than having metadata silos right? 00:58:33 Nathan Hotaling: Replying to "@nick isn’t that why..." fair 00:58:39 Dan Toloudis: We want some info right there locally in the zarr directory and for other data it is totally fine for it to be a URL to something potentially remote. There is the archival aspect and the “I just need to open the image and make sense of it”. We also have computed images where there is no microscope but some processing provenance and maybe parent images etc. 00:58:59 Josh Moore: Replying to "I would also advocat..." ❤️ but I often find myself missing XML when I’m faced with raw json. 00:59:10 Aybuke K Yoldas (EMBL-EBI): Replying to "one other think to t..." This case need to make sure databases and ontologies are maintained. 00:59:11 Dan Toloudis: Reacted to "I would also advocat..." with 👍 01:00:03 Nathan Hotaling: Replying to "I would also advocat..." Hahah, I can honestly say XML readability is so low that I have never thought “I miss the days when XML was pervasive” 01:00:15 Josh Moore: Replying to "Do people really thi..." True. In my mind that would be one of the metadata models that someone could put in NGFF but, as Virginia said, not to be created by this group. 01:00:27 Nick Schaub: Replying to "I would also advocat..." I was thinking the exact same thing Nathan. If XML died, I would shed no tears. 01:00:55 Norman Rzepka: Reacted to "I would also advocat..." with 👍 01:00:56 Aybuke K Yoldas (EMBL-EBI): Reacted to "We want some info ri..." with 👍 01:01:03 Josh Moore: Replying to "I would also advocat..." For the format, I agree. The tooling and specs, however, .... 01:01:27 Kenneth Ho: Reacted to "I would also advocat..." with 👍 01:02:14 Kenneth Ho: Replying to "I would also advocat..." +1 Josh 01:02:51 Aybuke K Yoldas (EMBL-EBI): Agree. 01:02:54 Caterina Strambio De Castillia: Agreed! 01:04:05 Josh Moore: Replying to "So maybe metadata be..." do you have a link which explains more how that works? 01:05:10 Nick Schaub: Replying to "So maybe metadata be..." https://onnx.ai/ 01:05:19 Dave Mellert: I have to head out now...thanks for the discussion, everyone! 01:05:41 Nathan Hotaling: Replying to "I would also advocat..." Yeah, that’s true Josh. The tools are useful now 01:06:24 Josh Moore: Reacted to "@Josh you know I am ..." with 🙂 01:07:08 Nick Schaub: The requirement is that metadata must have associated schema that can be used for validation? So we don't say we support X, Y, Z metadata standards, we just say support any metadata format that has validation schema? 01:07:55 Davis Bennett (he/him): As long as we are using something like JSON, someone will 100% put something extra in it. 01:08:01 sasker: The PDB for protein structures format is also quite liberal, but then software can struggle to handle all possible variants 01:08:20 Josh Moore: Reacted to "The PDB for protein ..." with 👍🏽 01:08:31 Josh Moore: Replying to "So maybe metadata be..." I’ll dig in. 01:08:46 Kenneth Ho: I don’t mind people putting a tons of things extra. What we have problem is that we cannot use them! Unless you are the author. 01:09:40 Caterina Strambio De Castillia: +1 Ken and Aybuke and Nathan I disagree with going nuts or also “whatever works for you”… There has to be some level of consensus building so that we can communicate otherwise it will end up being useless 01:09:44 Davis Bennett (he/him): This is an argument that people should use a standard, which I totally support, but I don’t think the standard should be “blessed” by ome-ngff to be effective 01:10:05 Caterina Strambio De Castillia: +1 Davis I agree 01:10:37 Davis Bennett (he/him): To be clear, when I said “go nuts” I meant “as a community, use whatever metadata works for you”, not that individuals should literally go nuts in json 01:11:23 Caterina Strambio De Castillia: +1 Nathan 01:11:40 samee: I have some example files where "structured annotation" nodes are recursive xml 01:12:09 samee: so much fun parsing as a dev :-) 01:12:38 Josh Moore: Replying to "I would also advocat..." imaging the xslt now … 🙂 01:19:05 Caterina Strambio De Castillia: Reacted to "I would also advocat..." with 👍 01:20:37 Caterina Strambio De Castillia: I like NextGen Metadata Framework as a project name 01:21:57 Dave Mellert: yay may meeting ended early I am back 01:22:04 Josh Moore: Reacted to "yay may meeting ende..." with ❤️ 01:22:21 Aybuke K Yoldas (EMBL-EBI): Reacted to "I would also advocat..." with 👍 01:22:24 Caterina Strambio De Castillia: Reacted to "yay may meeting ende..." with ❤️ 01:22:27 Josh Moore: Replying to "Agreed there needs t..." Naturally. 01:22:45 Aybuke K Yoldas (EMBL-EBI): Reacted to "Naturally." with 🖖 01:23:00 Norman Rzepka: Reacted to "Naturally." with 🖖 01:23:47 Davis Bennett (he/him): Is there any metadata that currently could not be expressed in OME-zarr? 01:25:01 John Bogovic: add 🖖 01:25:32 Josh Moore: https://github.com/ome/ngff/issues/78 01:29:27 Kenneth Ho: Reacted to "yay may meeting ende..." with ❤️ 01:29:45 Kenneth Ho: Reacted to "Naturally." with 🖖 01:32:19 Dave Mellert: I agree with Davis on the matter of rendering settings 01:32:57 Norman Rzepka: there has been a collection proposal. that could be a place to store rendering settings 01:33:05 Josh Moore: Reacted to "there has been a col..." with 👍🏽 01:33:16 Josh Moore: Replying to "there has been a col..." The collection (and the label) proposals are also on my short list 🙂 01:33:38 Nathan Hotaling: Sorry all, we have to step out, great discussion. Definitely interested in being involved! 01:33:47 Josh Moore: Reacted to "Sorry all, we have t..." with 👍🏽 01:33:52 Norman Rzepka: Replying to "there has been a col..." Happy to collaborate on that 01:33:54 Nick Schaub: I also have to step out. Thanks all! 01:35:45 Josh Moore: Reacted to "Happy to collaborate..." with 👍🏽 01:36:02 Christian Tischer: Replying to "there has been a col..." Me, too, if I understood the discussion correctly it is conceptually what we do in MoBIE: https://mobie.github.io/specs/mobie_spec.html#view-metadata 01:36:52 Josh Moore: If you are going to disappear and want to get added to the LinkML list, let me know. 01:37:33 Christian Tischer: Replying to "there has been a col..." ..there can be, e.g., viewA, where volume1 is red and volume2 green, and viewB where they have other colors a.s.o. 01:38:13 Norman Rzepka: Replying to "there has been a col..." yes. i think a mobie view and webknossos dataset are models for that 01:39:09 Christian Tischer: Replying to "there has been a col..." cool, if you want to work on this for ome-zarr I’d be happy to join 01:39:23 Norman Rzepka: Reacted to "cool, if you want to..." with 👍 01:39:30 Thomas Pengo: add 👍 01:42:51 Will Moore (OME): Gotta run, thanks all! 01:44:31 John Bogovic: add 👍 01:48:42 sasker: I think Yokagawa is using a form of OME-tiff 01:48:50 Josh Moore: Thanks for the reminder, Wouter-Michiel. Adam Taylor asked the very good question of whether anyone is missing from this meeting. (Vendors were on that list). If anyone knows of anyone else whom we should reach out to, please say the word. 01:49:26 Josh Moore: Replying to "I think Yokagawa is ..." Support for export of OME-TIFF has been one of the most substantial buy-ins, yes. 01:49:28 sasker: But the Yokagawa OME-tiff was non-standard so we couldn’t import in Omero 01:49:33 Josh Moore: Reacted to "cool, if you want to..." with 👍 01:50:21 sasker: Would be good to reach out to companies like Visiopharm, Halo, Patholytix etc 01:50:50 Josh Moore: Reacted to "But the Yokagawa OME..." with 🤯 01:50:55 Josh Moore: Reacted to "Would be good to rea..." with 👍🏽 01:51:12 Christian Tischer: Thank you all for the meeting. See you next time. 01:52:34 Ziwen Liu: There’s also open acquisition software (e.g. Micro-Manager) that’s writing raw data in various TIFF formats including OME-TFF. 01:52:57 Caterina Strambio De Castillia: I will have get off the call soon to go to the next meeting 01:53:15 Caterina Strambio De Castillia: Great to see you all. Thanks for organizing 01:54:39 Kenneth Ho: Sorry got to go. Thanks. 01:54:42 Caterina Strambio De Castillia: @Ziwen: MicroManager is actually (and ironically) not great with Metadata unfortunately 01:55:02 Caterina Strambio De Castillia: +1 ELMI gathering 01:55:09 Ziwen Liu: Yes… There’s non-standard behavior that gives us tons of headache 01:55:44 Dave Mellert: I have interest 01:57:12 Dave Mellert: okay have to run for real now. Thanks so much for organizing Josh (and all others involved in organizing) 01:58:30 Davis Bennett (he/him): I think we might need to have a chat about the tables issue 01:58:57 Caterina Strambio De Castillia: OK got to go. I am interested in being involved in the LinkML conversation so that I can learn more about it 01:59:20 Caterina Strambio De Castillia: Bye for now 01:59:21 Thomas Pengo: Sorry need to jump off - great conversation and such a great community! 02:00:47 Josh Moore: https://github.com/ome/ngff/pull/177 02:01:44 sasker: Thanks everyone!