SPEC Committee meeting running notes

[Zoom link](https://ucsc.zoom.us/j/98441337155?pwd=Mjl6TjdRTTNBc3ZGQzE3MTQyRVkyZz09) Notes: https://hackmd.io/@stefanv/SPEC-meetings/edit ## July 1, 2025, 9am Pacific / 4pm UTC ### Attendees - Kirstie - Stefan - Juanita - Matt H - Brigitta - Mridul - Aditi - Madicken ### Agenda - SPEC0 Long Term Support modification to accommodate scikit-learn - Idea is to also include a different, slightly delayed schedule for dropping dependencies - sklearn had their own SLEP, but seemed better to coordinate centrally via SPEC0 - Changes that need to be made: - When timer starts - Support times (LTS) - How does this impact previously "published" dates? - [Interactive docs SPEC proposal](https://github.com/scientific-python/specs/pull/388) - add tutorial doc to [learn](https://learn.scientific-python.org/development/tutorials/) or [development guide](https://learn.scientific-python.org/development/guides/docs/) - create blog post that refers to that content - let SPEC play out after, 1) right now there isn't much to coordinate as projects either do it one way or not 2) the tech is still somewhat experimental, not having a SPEC but as a guide docs would limit the churn we need to go through with the SPEC - How do the executable tutorials (https://github.com/scientific-python/executable-tutorials/) relate? - it doesn't, I at least try to keep the concepts separate: executable is "computational" content while the interactive could come or not as part of those tutorials. (We currently have JupyterLite for the template repo, but I try focus on the extensive testing angle for the tutorials) [name=bsipocz] - [PR to clarify SPEC process](https://github.com/scientific-python/specs/pull/395) - Add examples to (https://scientific-python.org/specs/purpose-and-process/#new-spec-proposals) like 1. scipy BoF, 2. start discussion in core projects developer or community meetings - Mridul's post in chat referred to PEP guidelines (https://peps.python.org/pep-0001/#start-with-an-idea-for-python). "Posting to the Ideas category of the Python Discourse is usually the best way to go about this, unless a more specialized venue is appropriate, such as the Typing category (for static typing ideas) or Packaging category (for packaging ideas) on the Python Discourse." - Concerns about SPEC 7's proposed use of `transition_to_rng` decorator - signature immediately replaces old parameter name with new parameter name - this causes typing issues (https://github.com/scipy/scipy/issues/23216) - We should consider adding a note to the SPEC but, unless someone with typing interest volunteers to dig into the issue more deeply, we're unlikely to "fix" the example decorator for such a specific requirement ### Open floor ## June 3, 2025, 9am Pacific / 4pm UTC ### Attendees - Stéfan van der Walt - Agriya Khetarpal - Matt Haberland - Brigitta Sipőcz - Paul Ivanov - Matthew Feickert - Mihai Maruseac - Melissa Mendonça - Kirstie Whitaker - Ralf Gommers - Dan McCloy - Matthew Brett - Sanket Verma - Juanita Gomez - Virgil Chan - Serge Rey - Lars Grüter - Nabil Freij - Mridul Seth - Erik Welch - ### Agenda [Summit whiteboard photo](https://photos.google.com/share/AF1QipO0RpKRZNfjtMHsKTXGA1DTKRtItWCNbzQUZGIwudOqjCztG-RtWVKnE7QeY58ylA?key=NGhPWFBjRUVkRWZzN1ZjMnJkZXVHZnhZMTNMSVJn) [Summit worklog](https://hackmd.io/36MAdbEHQseMJU4ibzIyvA) - Interactive docs SPEC draft ([name=Agriya]) ([PR #388](https://github.com/scientific-python/specs/pull/388), [deploy preview](https://deploy-preview-388--scientific-python-specs.netlify.app/specs/spec-0014/)) - [Add link to demo page build on JupyterLite] - Funded, work at Quansight, part of: - [Scientific Python communications grant](https://scientific-python.org/doc/scientific-python-community-and-communications-infrastructure-2022.pdf) - [Example](https://numpy.org/devdocs/reference/generated/numpy.vecdot.html#numpy.vecdot) - (Live on all numpy developer docs) - Sphinx docs can include "try in your browser" button - Currently long compared to other SPECs - Open questions: - where should all the content live (should it all be in a SPEC) - what does it mean to endorse this SPEC (except for NumPy etc) - [From zoom chat]: And then I envision endorsing would mean that a project endorses having interactive docs, rather than endorsing to have jupyterlite flavour of it or pyodide version. - Recommend separating into the generic principles into a SPEC and then put the specific information and implementation documentation somewhere else... - Question: would it be helpful to have demos from other projects (beyond NumPy & SciPy) - The Scikit-HEP domain stack has a few projects that have done this already: - [Awkward](https://awkward-array.org/doc/main/) [pyodide only] - [pyhf](https://pyhf.readthedocs.io/en/v0.7.6/#try-out-now-with-jupyterlite) - From zoom chat: - Security update [name=Mihai Maruseac] - Two new SPECs for [vulnerability disclosure](https://github.com/scientific-python/specs/pull/391) and [security policy](https://github.com/scientific-python/specs/pull/390) - SPEC 8 update: [trusted publishing and better practices](https://github.com/scientific-python/specs/issues/392) - :white_check_mark: Mihai, I think you just answered it RE: supporting projects that aren't using trusted publishers yet, but if you have additional thoughts on https://github.com/scientific-python/specs/issues/392#issuecomment-2884831143 that would be great - [name=Juanita] submitted a BoF at SciPy to work on these and more - Post Summit check in and follow-up (add topics below) - AI discussion update (but not in-depth discussion) - https://github.com/scientific-python/summit-2025/issues/35 - https://github.com/matthew-brett/sp-ai-post/blob/main/notes.md - [name=Stéfan] issue that became clearer for me through this discussion is not only license *violation* but also *attribution* - [name=Ralf] would be helpful to link to actual examples and legal cases - Noted challenge of reading all the comments in the github issue. - Option for a town hall discussion, alternative is to continue in written format eg in discord - Beginner installer update (Dan M, Matthew B) - https://github.com/scientific-python/installer - Considering updating the learn page on the different options, and clarifying which installer is best for which category of user/developer - Pretty close to "shipping", once Apple Certificate is sorted - [name=Ralf] Would be helpful to add target audience to README - executable-tutorials: generalize the takeaways into a Tutorials best CI practice SPEC. (Brigitta S) - SPEC would be recommendations around CI for these tutorials - Workshop tutorials vs narrative documentation for a package etc. - User stories: assumptions for different users (running beginning to end in new environment, or are they copy-pasting snippets, etc.) - Executable tutorials repo as a template for some of the implementations of these recommendations - Start *Tools* meeting - [name=bsipocz] I suggest we start using every second SPEC focusing meeting to focus on Tools, and use the latter to set priorities with the tools repos, or work out some topics that would need coordination. - Meta: [out of AI discussion] [name=Kirstie] what is a SPEC, and what does it mean to endorse a SPEC - [name=Matthew Feickert] from [discussion with Stéfan in the SPEC 8 update issue](https://github.com/scientific-python/specs/issues/392#issuecomment-2884632184) > Since endorsement means agreement with the SPEC concept, and is not tied to implementation, I think it would be fine to endorse on the higher level and implement as the domain stack sees fit. which I take as "endoresment means coneptually a core project team views this as a good for the ecosystem" - Meta [name=Kirstie] Needs of the SPEC meeting that are not currently being met - (Came out of Tools Meeting discussions; comment that SPEC meeting agendas are not always "full") - Should we clarify where discussions on SPECs are happening; where do we update after meetings? - You don't *need* to participate in meetings in order to develop a SPEC - [name=Brigitta] I would even go as far as to say that SPECs are not developed at the meeting, it's more just a checkin from projects to comment on the drafts that are in an advanced stage if they see any blockers with them, etc. - Meta, from tools topics: what is `installer`, is it a tool, or else? - Yes, classifying our repos/tools/apps would be part of the tools committee's job, as well as ensuring that all of these have enough maintainers/access, listed at a reasonable section on the website, etc. [name=Brigitta] ### Open floor ## May 6, 2025 ### Attendees * Dan McCloy * Sanket Verma * Brigitta Sipőcz * Jarrod Millman * Juanita Gomez * Kyle Sunden * Sebastian Berg * Nabil Freij * Kirstie Whitaker * Stéfan van der Walt * Kira Evans * Tyler Reddy * John Kirkham * Lars Grüter * Jake Bowhay * Erik Welch ### Announcements ### Agenda Summit planning issues: https://github.com/scientific-python/summit-2025/issues #### Pre-Summit plans: SPECs * CI best practices (SPEC5) * minimal dependencies * oldest dependencies * nightlies * all OS * efficiency -- caching/etc * tools: check out the [refactored scikit-image CI PR](https://github.com/scikit-image/scikit-image/pull/7756), some of the refactorings may be generic/useful enough for others to use ([example build dependency graph](https://github.com/scikit-image/scikit-image/actions/runs/13961049071)) * Tutorial CI SPEC * Brigitta to lead / work on this * treat the tutorials as a package, not enough to just dump in a container and make sure they run, need to test against old/minimal/nightly * some discussion of when to (not) have things like lockfiles * Release documentation / changelog SPEC * feels lower priority than the rest of the topics, but if there is a champion driving this fwd that would be nice * [Vulnerability disclosure SPEC](https://github.com/scientific-python/summit-2025/issues/8) * Jarrod to lead / work on this * Seth Larson (PSF) may be available remotely * SPEC 0 coordination * [sklearn has its own flavor of SPEC 0](https://github.com/scikit-learn/scikit-learn/issues/30888); can we incorporate their flavor into the SPEC language (maybe as a "LTS" variant?) * need some narrative tweaking around what "endorsement" means (this is especially prominent for SPEC 0, but applies to all SPECs) * SPEC2 - dispatching * Possible SPEC for documentation? * seems maybe not the best fit for a SPEC? * SPEC gives good visibility, but maybe just build the shared content first and "they will come"? #### Pre-Summit plans: Tools * reducing (code) duplication * property-based testing (hypothesis) * minimum dependencies: discuss how other packages handle this (xref to SPEC5) * efficient CI: * run only subset of tests on PRs (see scikit-image, or tutorials) * [tach](https://github.com/gauge-sh/tach) seems promising for preserving ongoing constraints on cross-submodule imports, but may not be powerful/stable/accurate enough to use as a basis for pruning tests * meeseeks/lumberbot: * @drammock to contact Mattias to hopefully schedule a remote session during summit, to discuss finalizing the renaming. Should loop in Madicken, Brigitta, and maybe Eric Larson. * Tutorial infrastructure * most realistic outcome during the summit is to collect ideas, and references for tools or template pieces to be used * the Hugo theme * probably doesn't have long-term support * what instead? MyST / MystMD / JupyterBook * myst * Tracking issue: https://github.com/scientific-python/summit-2025/issues/25 * myst and JB developers will be at the summit, we try to tackle a lot of the blockers that would allow us to move towards using mystmd-JB2 * currently doesn't handle API documentation or blogs * load bearing walls * maybe leverage devstats to identify the under-staffed packages, then publicize them to drive contributions? ## April 1, 2025 ### Attendees - Jarrod Millman - Brigitta Sipőcz - Matthew Feickert - Madicken Munk - Sanket Verma - Stéfan van der Walt - Guillaume Lemaitre - Serge Rey - Juanita Gomez ### Announcements - https://scientific-python.org/summits/developer/2025/ - 11-14 May 2025 – Seattle, WA - https://github.com/berkeley-ospo/safe-ose-2025 April 15th! - reviewers / issues - 67 stars https://github.com/scientific-python/specs - 307 followers https://github.com/scientific-python/ - 30 public members (79 total) https://github.com/orgs/scientific-python/people - 468 members https://discuss.scientific-python.org/ - 661 members https://discord.com/invite/vur45CbwMz ### Agenda - Safe OSE discussion - Repo: https://github.com/berkeley-ospo/safe-ose-2025/issues - Please add input on issues :pray: - For stars etc stats I (Brigitta) would include the most starred/used ones hosted/developed at Scientific Python. All of these have some security aspects, making libraries in the ecosytstem more secure by testing/tooling. E.g.: - 333 for cookie - 158 for lazy loader - 108 spin - 104 pytest-doctestplus - I would also use the "fork" stats to show the community involvement (maybe we need to say something about the fork based contributing workflow here so it makes sense as a meaningful stat). In my book that is a more useful stat than the number of public members in the org etc (but also I'm not the NSF reviewers) - More Core Projects (domain stacks) - The initial idea was to provide high-quality signals to the larger community - scikit-hep [name=Matthew] - [name=Matthew]: As [`scikit-hep`](https://github.com/scikit-hep/scikit-hep) is a metaproject, maybe makes more sense to have [Awkward](https://github.com/scikit-hep/awkward), which is probably our core/flagship library that could actually be useful outside of the HEP context - Context: - Scikit-HEP community organization: https://scikit-hep.org/ - Less mature in terms of governance than other project ecosystems: Lacking steering committee beyond the Scikit-HEP admin board (which is somewhat similar). General interest to adopt more mature org structures. - Awkward itself has its own CSSI supported Awkward ecosystem ("the Awkward family") which has more formal development structures: - [Awkward](https://github.com/scikit-hep/awkward) - [Dask-Awkward](https://github.com/dask-contrib/dask-awkward/) - [Ragged](https://github.com/scikit-hep/ragged) - pysal https://pysal.org/ - proposal: add them to core projects / lightly edit https://scientific-python.org/specs/core-projects/ - More SPEC0 tools - Brigitta proposes to keep the tools separate from the SPECs repo, or at least to the bare minimum and have anything else in their own repo (we already do this in practice, see the nightly wheel tools are separate. Motivation: we need to deal with this PR: https://github.com/scientific-python/specs/pull/374 - [SPEC 0](https://scientific-python.org/specs/spec-0000/) [name=Guillaume] - Reminder that SPECs are aspirational recommendations; perhaps not the best name to signal that, but still the case. - Projects sometimes therefore feel reluctant to endorse, if they don't follow a SPEC to the T themselves - SPEC0 is a good example: some projects, such as sklearn, have more resources and can afford a longer support period. - In an ideal world, we would not need an entirely new recommendation for that, such as a SLEP, but could modify SPEC0 to accommodate various modes of support. - Another example is `lazy_loader`, where we don't drop dependencies so early, because longer-term support of that library is necessary for *other* projects to be able to do longer-term support. - https://github.com/scikit-learn/scikit-learn/issues/30888 - In person participation difficult from Europe this year, but sklearn has a representative in US who can attend - Thoughts on expanding SPEC0 with extended support option? - Will NumPy be able to align with an LTS plan? :shrug: - Larger collaborations may consider LTS an option, but also rely on frozen lockfiles for replicating full environments. - Python updates, in particular, can be challenging for users. - S: Note that you can always install older versions that used to run; you just don't get access to new features on old versions of Python. - Might make migrations harder, when you have to switch Python version to update libs. - Science use cases may need to use some small domain specific packages that are more research code than well maintained libraries. These very likely don't work with new versions of core libraries, thus put very old upper limits on the requirements causing headaches for the domain stack. - [name=Matthew], [name=Brigitta]: The Astronomy and particle physics ecosystems have two communities inside of them: Developers and Users. The Users have expectations of their old codes to work and being supported for very long periods of time. The don't follow new feature developments in e.g. numpy or even in cpython. While the the Developers are building new tools and working in "newer" (than traditional batch systems with preset environments) computing environments (e.g. cloud, k8s) where new features from new versions of libraries are being used. - [name=Matthew]: Something else I forgot to mention when discussing with [name=Matt] about environment constraints is that in particle physics there are lots of users that are not setting up their environments themselves. They are using currated base environments that are globally distributed through a system called [CVMFS](https://cernvm.cern.ch/fs/), and so if they are using a Scikit-HEP tool in their anlaysis that might have to fit into the constraints of this base environment (this is a base environment that the actual person using might have no direct control over). This isn't ideal for obvious reasons, but this comes from a history (15+ years ago) of Python being used in large frameworks (more more application like) rather than our modern concept of pulling in bespoke libraries or ecosystems of libaries for your analysis. - Summit SPEC topic (discussions) - Focuses on community & governance SPECs, new SPECs, infrastructure & tools - Tooling around existing SPECs - Landing additional SPECs - Outreach - How to engage before the summit: - Consider existing SPEC issues: https://github.com/scientific-python/specs/issues - File specific work items at https://github.com/scientific-python/summit-2025/issues ## March 3, 2025 ### Attendees - Jarrod Millman - Kirstie Whitaker (UC Berkeley, she/her) - Matt Haberland - Guillaume Lemaitre - Sanket Verma - Brigitta Sipőcz - Madicken Munk - Stéfan van der Walt - Juanita Gomez ### Announcements - Developer Summit 2025: Seattle, 5/11 - 5/14 (5/15?) - invites will be sent out next week - website stub will go up this week - focus on statistics, optimization, sparse data, plotting, dataframes? - Kirstie Whitaker, ED at Berkeley Institute for Data Science - https://cdss.berkeley.edu/news/meet-kirstie-whitaker-new-executive-director-bids - https://book.the-turing-way.org - SPEC endorsements - scikit-learn - [SPEC 1](https://github.com/scientific-python/specs/pull/376) - [SPEC 6](https://github.com/scientific-python/specs/pull/377) - [SPEC 8](https://github.com/scientific-python/specs/pull/378) ### Agenda - SPEC0 and sklearn - https://github.com/scikit-learn/scikit-learn/issues/30888 - Can we adapt SPEC0 to accomodate? - Proposal: standard and extended support - [Script for querying lower bounds](https://github.com/scientific-python/specs/pull/374/) - We can provide data as part of https://scientific-python.org/specs, but tools should live externally - Action could, e.g., be part of https://tools.scientific-python.org - [SPEC 9 — Governance](https://scientific-python.org/specs/spec-0009/) - Time to review and move toward adoption - Would like to get more authors / reviewers - Sanket will update to include comments from previous meeting, including Onion Governance - [SPEC 2 — API Dispatch](https://scientific-python.org/specs/spec-0002/) - NX dispatch is getting mature ([3.5 planned for April 30, 2025](https://github.com/networkx/networkx/milestones)) - [dispatch merged in scikit-image 🎉](https://github.com/scikit-image/scikit-image/pull/7520) ([0.26 planned for April 14, 2025](https://github.com/scikit-image/scikit-image/milestone/30)) - Potential long term goal: combine approaches into `spatch` - [SPEC 3 — Accessibility](https://scientific-python.org/specs/spec-0003/) - Isabella, Tania working on [some material submitted as PR to learn.scientific-python.org](https://github.com/scientific-python/learn.scientific-python.org/pull/232) - https://book.the-turing-way.org/community-handbook/accessibility.html - Can link to these resources from the SPEC, with a very high-level overview of the topics that are covered - Discussed best places for material to live, how to link to it, process to engage with it - [Anyone want to summarize?] <-- ❓ - Establish and grow the learn team - [SPEC badges](https://scientific-python.org/specs/purpose-and-process/#badges) - let's try to get some projects to use them - NX (Jarrod) - [scikit-image](https://github.com/scikit-image/scikit-image/pull/7737) (Jarrod) - ... - [Core projects](https://scientific-python.org/specs/core-projects/) - engaging pandas - any other projects to invite (more domain specific?) ## February 4, 2025 ### Attendees - Jarrod Millman - Juanita Gomez - Brigitta Sipőcz - Inessa Pawson - Lars Grüter - Stéfan van der Walt - Sanket Verma ### Announcements - Developer Summit 2025: Seattle, 5/11 - 5/14 (5/15?) - Kirstie Whitaker, ED at Berkeley Institute for Data Science ### Agenda - SPEC comments / discussion cleanup - ~~Brigitta mentioned an issue, which she will file~~ It was a red herring, it seems to only affect the netlify rendering and not the deployed version - Zarr endorsed SPEC4 - [SciPy discussion on AI use for code contributions](https://discuss.scientific-python.org/t/a-policy-on-generative-ai-assisted-contributions/1702/14) - https://runllm.com/ - Can do, e.g., fine-tuned searches on project documentation - See, e.g, https://dspy.ai/ - [PR: dump lower bounds for packages in schedule.json](https://github.com/scientific-python/specs/pull/374) ## January 7, 2025 ### Attendees - Jarrod Millman - Matthias Bussonnier - Stéfan van der Walt - Sanket Verma - Juanita Gomez ### Announcements - [NSF Safe-OSE](https://new.nsf.gov/funding/opportunities/safe-ose-safety-security-privacy-open-source-ecosystems): Securing the Scientific Python Ecosystem - Jan. 14 preliminary proposal - Apr. 22 full proposal ? - Pending endorsements - Zarr update: SPEC4? - [Zarr Python 3.0](https://bsky.app/profile/zarr.dev/post/3lew6qrwjda2i) ### Agenda - Developer Summit 2025 (Seattle 5/11 - 5/14? 5/15?) funding - SP may be able to fund 1-2 devs - NX may be able to fund 1-2 devs - skimage may be able to fund 1-2 devs - nvidia may be able to fund 1-2 devs - google may be able to fund 1-2 devs ? - psf funding ? - ... - SPEC comments / discussion cleanup - https://scientific-python.org/specs/spec-0000/ - https://discuss.scientific-python.org/t/spec-0-minimum-supported-versions/33 - https://discuss.scientific-python.org/t/scientific-python-spec-0-minimum-supported-dependencies/1274 - https://scientific-python.org/specs/purpose-and-process/ - issue template - only have one - Review Core Projects - https://scientific-python.org/specs/core-projects/ - important domain stacks communities / projects - compilers / build infrastructure - other data structures - other plotting - https://anaconda.org/scientific-python-nightly-wheels - https://github.com/mesonbuild/meson-python/ - https://github.com/nucleic/kiwi - ... others suggestions? ... - SPEC-like: Matt's document: [deprecating `scipy.stats.mstats`](https://github.com/scipy/scipy/issues/22194) - review open issues: - https://github.com/scientific-python/specs/issues - what we want to accomplish next year TODOs: - Sanket: Zarr endorsing SPEC4 - Jarrod & Stéfan: developer summit dates - Stéfan: unify discussion links in SPECs / purpose & process - Jarrod: propose reaching out to astropy, geopandas, bokeh, cython, ... - Jarrod & Matthew: propose new projects for nightly wheels, e.g. kiwi, meson, meson-python ## December 3 ### Attendees - Sanket Verma - Inessa Pawson - Brigitta Sipőcz - Jarrod Millman - Stéfan van der Walt - Matthias Bussonnier (a bit late) - Madicken Munk (a bit late) ### Agenda - Safe-OSE proposal: https://new.nsf.gov/funding/opportunities/safe-ose-safety-security-privacy-open-source-ecosystems - what we want to accomplish next year - review open issues: - https://github.com/scientific-python/specs/issues - https://github.com/scientific-python/specs/issues/332 - giscus or host on discourse like with blog? - [PR for discourse (https://github.com/scientific-python/scientific-python.org/pull/591) - Endorsements - Zarr update: In the recent conversation, Zarr-Python is ready to endorse SPEC4 - NumPy endorsed SPECs 1,6,7,8 ## November 5 ### Attendees - Stéfan van der Walt - Juanita Gomez - Brigitta Sipőcz - Guillaume Lemaitre - Sanket Verma - Matthias Bussonnier - Ghee the cat 😺 - Lars Grüter - Madicken Munk ### Agenda - Welcome Matt - Endorsements - NetworkX: [SPEC6](https://github.com/scientific-python/specs/pull/358), [SPEC7](https://github.com/scientific-python/specs/pull/356), [SPEC8](https://github.com/scientific-python/specs/pull/357) - SciPy: [SPEC 7](https://github.com/scientific-python/specs/pull/351) - [NumPy thread](https://mail.python.org/archives/list/numpy-discussion@python.org/thread/KDP7EA2DBLZ7M7IQCPYWULI66H43NPKX/) with endorsement concerns - scikit-learn - SPEC 0 is a dead-end - Got endorsement for SPEC 1 - Make PR on https://github.com/scientific-python/specs; get core devs to thumbs-up; link to discussion - Working on endorsement of SPEC 6 - SPEC 7, some issues to deal with in scikit-learn before can endorse - Zarr - Discussion on endorsements (brought up in the community meetings) - [SPEC8](https://ossci.zulipchat.com/#narrow/channel/423692-Zarr/topic/SPEC8.20.28.22Securing.20the.20release.20process.22.29) - [SPEC6](https://ossci.zulipchat.com/#narrow/channel/423692-Zarr/topic/SPEC-6.20.28.22Keys.20to.20the.20castle.22.29) - Nightly wheels (SPEC4)? - In progress - [SPEC 3: Accessibility](https://scientific-python.org/specs/spec-0003/) - [PyData Sphinx Theme accessibility](https://pydata-sphinx-theme.readthedocs.io/en/stable/user_guide/accessibility.html) - [SPEC 9: Governance](https://scientific-python.org/specs/spec-0009/) - Having a section for "Onion Consensus" — a governance model that has worked well so far across multiple projects - Where should we discuss "ownership": noone can make the whole project disapear (github org, not personal account) - Can we identify the *principles of good governance* for this SPEC? - A table of projects vs governance model could be interesting, *if* they're not all using roughly the same one. - Having a section in the SPEC on how to ingest/accept donation of repositories would be a good addition - [name=Guillaume] Question regarding SPEC/recommendations regarding anaconda default channel vs. conda-forge channel - Confusion around - [pixie.dev vendor lock-in blog post](https://prefix.dev/blog/towards_a_vendor_lock_in_free_conda_experience) - We could update [the development guide](https://learn.scientific-python.org/development/tutorials/dev-environment/#option-2-using-conda) by filing an issue on [cookie](https://github.com/scientific-python/cookie) - [name=Guillaume] I'll open an issue and draft an additional resource - [name=Matt] SPEC 12 is complete enough to be merged as draft, although I expect specific rules to be controversial - [Discussion](https://github.com/scientific-python/specs/pull/326) - [Last comment from Ruff team member](https://github.com/scientific-python/specs/pull/326#issuecomment-2338149213) ## General - [name=Jarrod] NumPy matrix deprecation; sparse array API; please push on adoption of sparse array API - [name=Jarrod] topics: https://github.com/topics/spec-0 - Use badges instead - [Badge example](https://github.com/scikit-image/scikit-image/pull/7536/files#diff-b335630551682c19a781afebcf4d07bf978fb1f8ac04c6bf87428ed5106870f5R7) (skimage) - [name=Jarrod] Potential transitioning of theme from Hugo to [MyST JS](https://mystmd.org/guide) ## October 1, 2024 (9:00AM PDT / 4PM UTC) ### Attendees - Lars Grüter - Stéfan van der Walt - Brigitta Sipőcz - Pamphile Roy - Sanket Verma - Guillaume Lemaitre - Madicken Munk - Matthias Bussonnier - Inessa Pawson - Paul Ivanov ### Agenda - [SPEC 12: Formatting Mathematical Expressions](https://github.com/scientific-python/specs/pull/326) ready to merge as draft - Discussing with maintainers from ruff and black - Linting (detecting something is not right) vs Formatting (fixing error automatically) - Please go ahead and merge as draft - Status check on PRs: - [SPEC 10: Changelog](https://github.com/scientific-python/specs/pull/321) No current update; but planning to keep on working. - [SPEC 13: Targets and naming](https://github.com/scientific-python/specs/pull/324) Don't want to fix things that may be moving targets in the Python ecosystem. Let's hold off on developing SPEC now until we have better consensus. May consider closing SPEC? - Endorsements - [SPECS 6, 7, 8](https://scientific-python.org/specs/) are ready for core project endorsement - sklearn update: - SPEC 7 (RNG): some bigger concerns they need to address; no problem with the SPEC itself, which suggests renaming and gives us an opportunity to change behavior; will be having smaller meetings soon where, hopefully, we can converge on path forward; not a very high priority right now - Discussion regarding scikit-learn specific issue with randomness ambiguity: https://github.com/scikit-learn/enhancement_proposals/pull/88 - SPEC 1 (lazy loading): some discussion on sklearn - Discussion going on in this new PR: https://github.com/scikit-learn/scikit-learn/pull/29793 - skimage: - endorsed (and following) SPECs 6, 7, and 8; only unendorsed SPECs are the ones that are not ready for consideration - good exercise to write up permissions, find unused tokens, people with permissions they shouldn't have, etc. - looking into using `teams` action for setting up permissions on gh - ipython: Matthias will take a look - TODO: Improve `teams` action to work with user IDs, not usernames which can be changed (see issue 13). - Badges now available - Example: [![SPEC 6 — Keys to the Castle](https://img.shields.io/badge/SPEC-6-green?labelColor=%23004811&color=%235CA038)](https://scientific-python.org/specs/spec-0006/) (see https://scientific-python.org/specs/spec-0006/#badges) - New members to serve on SPEC committee - [Member list](https://scientific-python.org/specs/steering-committee/) - Welcome Guillaume and Matt! - General discussion - Tools update: bring on [batchPR](https://github.com/astrofrog/batchpr/issues/33)? > The aim of this package is to provide an easy way to do automated issues or pull requests to a selection of repositories and make specific changes to them. - Lars: https://tools.scientific-python.org/ --> should we specify some other minimum requirements, such as testing, e.g. - SPEC9: governance, merged in draft stage - Please review SPEC9, and make suggestions - Within Scientific Python, there can be a security manager role - NumPy triage meeting tomorrow; on agenda, endorse SPEC-1 ## September 3, 2024 (9:00AM PDT) ### Attendees - Jarrod Millman - Brigitta Sipőcz - Lars Grüter - Juanita Gomez - Mridul Seth - Pamphile Roy - Stéfan van der Walt ### Agenda - Membership - [Guillaume Lemaitre as new member](https://discuss.scientific-python.org/t/suggestion-for-new-committee-member/1401/5) - [Old members stepping down](https://discuss.scientific-python.org/t/spec-committee-active-membership/1402/14) - We agreeded to go ahead and add Guillaume, move non-active or per members request to emeritus status, update to reflect updated process. - https://scientific-python.org/specs/steering-committee/#how-do-you-add-a-member - https://scientific-python.org/specs/steering-committee/#how-do-you-remove-a-member - Update SPEC purpose and process to explain how to update existing SPEC [name=@stefanv] - Small adjustment to purpose and process to [move SPECs from Draft to Published](https://github.com/scientific-python/specs/pull/339) - SPEC statuses - Ready for endorsement (move from draft -> published): - SPEC6: Keys to the Castle - SPEC7: Seeding pseudo-random number generation - SPEC8: Securing the release process - Actively being worked on: - [SPEC2: API Dispatch](https://github.com/scientific-python/specs/pull/334) - [SPEC12: Formating mathematical expressions](https://github.com/scientific-python/specs/pull/326) — please take a look, close to ready - Needs work: - SPEC3: Accessibility - ping Tania [name=@stefanv] - SPEC9: governance - SPEC10: changelog and release documentation - Will probably be retired: - SPEC13: Recommended targets and naming conventions - get everything discussed in and then publicize (tweet) - finish comments PR [name=@stefanv] - https://github.com/scientific-python/scientific-python.org/pull/591 - blog post for developer summit - Write collaboratively, between Matt Feickert, Stéfan, Pamphile, Brigitta, anyone else who wants to help! - Try to follow the purpose and process for starting new SPECS - Keep SPECs short and crisp - Merge early, continue editing ## July 2, 2024 (9:00AM PDT) ### Attendees - Jarrod Millman - Juanita Gomez - Sanket Verma - Stéfan van der Walt - Matthias Bussonnier - Lars Grüter - Brigitta Sipőcz ### Agenda * Welcome * using read.ai * Membership * Emeritus status * https://scientific-python.org/specs/steering-committee/ * New members * Guillaume Lemaitre * PRs * https://github.com/scientific-python/specs/pulls * Governance (Jarrod) * 7, 8, and 9 likely merged drafts * badges PR * https://github.com/scientific-python/scientific-python.org/pull/591 ## May 7, 2024 (9:00AM PDT) ### Attendees - Jarrod Millman - Juanita Gomez - Sanket Verma - Brigitta Sipőcz - Matthias Bussonnier - Lars Grüter - Kristen Thyng - Mridul Seth - Stéfan - Pamphile Roy ### Agenda * Welcome * Developer Summit * https://scientific-python.org/summits/developer/2024 * https://tools.scientific-python.org * https://github.com/scientific-python/summit-2024 * https://github.com/orgs/scientific-python/projects/8 * https://github.com/scientific-python/summit-2024/issues * [SPEC: API observability](https://github.com/scientific-python/summit-2024/issues/1) * there was a scipy project ~10 years ago, it may be worth trying to dig up the details. Main concerns: * how to ask consent -> consent broker * https://github.com/takluyver * https://github.com/Carreau * [SPEC-?: Parallel API](https://github.com/scientific-python/summit-2024/issues/13) - Collecting various approaches would be good to do before the summit - Also see Thomas Fan's blog post, linked in the issue - S wrote a CZI proposal to work on this that was not funded; happy to share * [SPEC-7: Seeding pseudo-random number generation](https://github.com/scientific-python/summit-2024/issues/11) - General problem: libraries that invest early in figuring out an API, and have it implemented, will be less likely to want to move - However, still valuable to have general guidance for other libraries in the ecosystem, and that library may even endorse it - Writing up a SPEC also makes the arguments clear, which is helpful for future conversation * [SPEC-8: Supply-Chain Security](https://github.com/scientific-python/summit-2024/issues/9) - https://openssf.org/ - Establishing community trust (e.g., Debian) ## April 2, 2024 (9:00AM PDT) ### Attendees - Jarrod Millman - Matthias Bussonnier - Juanita Gomez - Madicken Munk - Sanket Verma - Brigitta Sipőcz - Lars Grüter ### Agenda * Welcome * Wasn't it Stefan's birthday ? * Review agenda / anything to add? * Main focus today * https://scientific-python.org/summits/developer/2024 * Make the dates more visible. * https://tools.scientific-python.org * SPEC meta stuff * [Core projects](https://scientific-python.org/specs/core-projects/) * Update on sympy * [SPEC Badges/GH Topics](https://github.com/scientific-python/specs/issues/240) * [WIP PR](https://github.com/scientific-python/specs/pull/305) * How does it look? * Jarrod needs to implement shortcode ([tabs refactored](https://github.com/scientific-python/scientific-python-hugo-theme/pull/579)) * SPECs * [SPEC 2 — API Dispatch](https://scientific-python.org/specs/spec-0002/) * [WIP PR](https://github.com/scientific-python/specs/pull/307) * [SPEC 4 — Using and Creating Nightly Wheels](https://scientific-python.org/specs/spec-0004/) * pyarrow added * Jarrod will start inviting build tools (cython, pythran, etc.), dependencies (pillow, etc.), common tools, core packages for important subdomains * [SPEC 8 — Parallel terminology](https://discuss.scientific-python.org/t/terminology-for-parameters-controlling-parallel-computation/1016) * SPEC website updates * [v0.13](https://github.com/scientific-python/scientific-python-hugo-theme/releases/tag/v0.13), [v0.14](https://github.com/scientific-python/scientific-python-hugo-theme/releases/tag/v0.14), [v0.15](https://github.com/scientific-python/scientific-python-hugo-theme/releases/tag/v0.15), [v0.16](https://github.com/scientific-python/scientific-python-hugo-theme/releases/tag/v0.16) * s-p-org PRs [509](https://github.com/scientific-python/scientific-python.org/pull/590), [599](https://github.com/scientific-python/scientific-python.org/pull/599), [601](https://github.com/scientific-python/scientific-python.org/pull/601) * specs PRS [304](https://github.com/scientific-python/specs/pull/304), [306](https://github.com/scientific-python/specs/pull/306), [309](https://github.com/scientific-python/specs/pull/309) * [WIP discussion](https://github.com/scientific-python/scientific-python.org/pull/591) * Planning discussion * https://scientific-python.org/summits/developer/2024 * https://tools.scientific-python.org * repo review (GH agnostic?)? * [meeks](https://github.com/meeseeksBox/MeeseeksDev) resurrection * SPECs * best community practices * parallel terminology * repo review (GH agnostic?) * declarative Sphinx config? * https://sphinx-pyproject.readthedocs.io/en/latest/ * security best practices * accessibility ## March 5, 2024 (9:00AM PDT) ### Attendees - Jarrod Millman - Matthias Bussonnier - Stefan van der Walt - Sanket Verma - Ross Barnowski - Lars Grüter - Kristen Thyng - Brigitta Sipőcz - Mridul Seth - Pey Lian Lim ### Agenda * Welcome * Review agenda / anything to add? * SPEC meta stuff * [Core projects](https://scientific-python.org/specs/core-projects/) * invite [SymPy](https://discuss.scientific-python.org/t/spec-0-include-sympy/975/2), ... * [x] Matthias will DM Aaron Meurer to gage interest and wether to move forward * [ ] Matthias will replay Aaron thoughts and reply. * [SPEC Badges/GH Topics](https://github.com/scientific-python/specs/issues/240) * for badges see [internal link](https://hackmd.io/WPTkRnNtSOeF7OhUTRHDwA?both#Badges) and [![SPEC 1 — Lazy Loading of Submodules and Functions](https://img.shields.io/badge/SPEC-1-blue)](https://scientific-python.org/specs/spec-0001/) and [![](https://img.shields.io/badge/SPEC_0-Compliant-green?labelColor=%23004811&color=%235CA038)](https://img.shields.io/badge/SPEC_0-Compliant-green?labelColor=%23004811&color=%235CA038) * Consider some additional language to emphasize that endorsing is not compelling projects to do anything * Concern with badges: badges as "award" * Bear in mind: badges would also be voluntary * One can make badges that refer to multiple specs at onece: [![Scientific Python Ecosystem Coordination](https://img.shields.io/badge/SPEC-1,3,5-blue)](https://scientific-python.org/specs/) * [name=jarrodmillman] - Will open PR related to badges, review over course of week * for topics see [spec-1](https://github.com/topics/spec-1), [spec-2](https://github.com/topics/spec-2), [spec-4](https://github.com/topics/spec-4) * limitations e.g. [only 20](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/classifying-your-repository-with-topics) * Discussion: is the topics section the right place for this? * promote in ecosystem adoption section (e.g., add text [here](https://github.com/scientific-python/specs/blob/main/spec-0000/index.md#ecosystem-adoption))? * SPECs * [SPEC 2 — API Dispatch](https://scientific-python.org/specs/spec-0002/) - [ ] [Mridul/Jarrod to redraft based on NX](https://github.com/scientific-python/specs/issues/229#issuecomment-1978045914) - Remove old text and authors - [NetworkX Dispatch Meeting notes](https://hackmd.io/@networkx-dispatching/meeting) - [Backends](https://networkx.org/documentation/stable/reference/utils.html#backends) - NXEP coming - potential dispatching summit * [SPEC 4 — Using and Creating Nightly Wheels](https://scientific-python.org/specs/spec-0004/) * requested 2 TB (currently using 11.8GB / 40 GB -- 6.4GB openblas and 1.6GB pandas) * current: awkward, awkward-cpp, contourpy, dipy, h5py, ipython, matplotlib, networkx, numpy, pandas, scikit-image, scikit-learn, scipy, scipy-openblas32, scipy-openblas64, shapely, statsmodels, sunpy, tables, xarray * pending: [pyarrow](https://github.com/apache/arrow/issues/40216) * potential: - build tools (cython, pythran, etc.), dependencies (pillow, etc.), common tools, core packages for important subdomains - pure python (pytest, sphinx, etc.) - Matthias: even pure python, wheels are usefull for pip install --no-binary that is used in some project to make sure that all wheels are working on all platforms. - link to existing collections (e.g., astropy) and maybe also publish just the latest version - Matthias: we can also just have a cron that pull from pytest repo and reupload to the scientific-Python one. * [SPEC 7 — Seeding PRNGs](https://github.com/scientific-python/specs/pull/180) * [SPEC 8 — Parallel terminology](https://discuss.scientific-python.org/t/terminology-for-parameters-controlling-parallel-computation/1016) - njobs, num_workers, workers, etc. - 3 use cases - nondiscriminant - cpus - threads * SPEC website updates - [Use PST styling for spec meta styling](https://github.com/scientific-python/scientific-python.org/pull/576) see [example](https://deploy-preview-576--scientific-python-org.netlify.app/specs/spec-0002/) * What to do for core projects links * Project Gallery * one toml file * [new format example](https://github.com/scientific-python/scientific-python-hugo-theme/blob/main/doc/content/about/theme-team.toml) * Project pages * [current source](https://raw.githubusercontent.com/scientific-python/specs/main/core-projects/ipython.md) and * [current look](https://scientific-python.org/specs/core-projects/ipython/) and the [work in progress](https://deploy-preview-576--scientific-python-org.netlify.app/specs/core-projects/ipython/) * Remove history section or fix links? * Where should discussions happen? - make [issue template](https://github.com/scientific-python/specs/tree/main/.github/ISSUE_TEMPLATE) - make [PR template](https://github.com/scientific-python/specs/blob/main/.github/PULL_REQUEST_TEMPLATE.md) - [Discourse commments](https://github.com/scientific-python/blog.scientific-python.org/pull/177) see [example](https://blog.scientific-python.org/numpy/numpy-rng/) * Miscellaneous * https://scientific-python.org/summits/developer/2024 see [preview](https://deploy-preview-584--scientific-python-org.netlify.app/summits/developer/2024/) * https://tools.scientific-python.org see [preview](https://deploy-preview-36--tools-scientific-python-org.netlify.app/) * https://fosstodon.org/@scientific_python ## February 6 9:00AM PDT ### Attendees - Jarrod Millman - Madicken Munk - Juanita Gomez - Kristen Thyng - Brigitta Sipőcz - Stéfan van der Walt - Sanket Verma - Mridul Seth - Pey Lian Lim - Matthias Bussonnier ### Agenda * Updates on existing SPECs * SPEC0 * What does 2 year support mean for endorsers and adopters * https://github.com/scientific-python/specs/pull/297 * SPEC 4 activities * New SPECs ready for final comments / merging * [SPEC6](https://github.com/scientific-python/specs/pull/168): Keys to the castle * [SPEC7](https://github.com/scientific-python/specs/pull/180): Random API * S: Please merge after final review :pray: * [Update SP group membership / teams via PR](https://github.com/scientific-python/teams) * [sync-teams-action](https://github.com/scientific-python/sync-teams-action) * [Tools](https://scientific-python.org/tools) * Scientific Python developer summit * 3-7 June 2024, eScience Institute, Seattle * Summit topics * Continued sparse work (WG is still active after previous summit) * Community management SPECs * Storage formats * Parallel APIs * Open discussion * SPEC 4 storage * Self-hosted Anaconda server? (on views, e.g.) ## October 3rd 9:00AM PDT **Topic:** SPECs ### Attendees - Jarrod Millman - Madicken Munk - Juanita Gomez - Pamphile Roy - Matthias Bussonnier - Lars Grüter - Kristen Thyng - Brigitta Sipőcz - Stéfan van der Walt ### Relevant Links - [Scientific Python - Scientific Python Ecosystem Coordination](https://scientific-python.org/specs/) - [GitHub - scientific-python/specs: Scientific Python Ecosystem Coordination (SPEC) documents](https://github.com/scientific-python/specs) ### Agenda 1. Updates - [spec-0 and spec-4 endorsements by xarray](https://github.com/scientific-python/specs/pull/270) - [endorse spec 4 by IPython](https://github.com/scientific-python/specs/pull/268) - [add MNE to list of SPEC001 adopters](https://github.com/scientific-python/specs/pull/271) - [Generate gallery from GH teams](https://github.com/scientific-python/specs/pull/276) 2. Teams - [How do you add a member?](https://deploy-preview-276--scientific-python-specs.netlify.app/specs/steering-committee/#how-do-you-add-a-member) - make a PR to https://github.com/scientific-python/teams - make a PR to https://github.com/scientific-python/scientific-python.org/ after running `make teams` - add them to https://discuss.scientific-python.org/g/spec-steering-committee - this requires that they have accounts on both GitHub and discuss.scientific-python.org. - Investigate [Invite discourse users to a group](https://meta.discourse.org/t/invite-users-to-a-group/15544) 3. Pending PRs - [SPEC-0000: Soften explicit drop schedule](https://github.com/scientific-python/specs/pull/273) - [Clarify schedule is guideline](https://github.com/scientific-python/specs/pull/272) - Having github topics for specs ? 4. Affiliated projects / Domain stacks / Cards / Badges / etc. - [Question: Is there any core library missing?](https://github.com/scientific-python/specs/issues/241) - [Domain stacks](https://scientific-python.org/summits/domain-stacks/) - [Add badges](https://github.com/scientific-python/specs/issues/240) 5. Theme - [Dark theme](https://theme.scientific-python.org/) - [Project pages](https://scientific-python.org/specs/core-projects/ipython/) - other issues - reviewers / feature requests 6. CZI EEOS 6 - [LOI due October 17](https://chanzuckerberg.com/rfa/essential-open-source-software-for-science/) Next meeting: Notes from NumFOCUS Summit: https://hackmd.io/32sQNc0STMCEHCF6q3P_4g # September 5, 2023 **In attendance:** Jarrod Millman, Brigitta Sipőcz, Ralf Gommers, Mridul Seth, Stéfan van der Walt, Juanita Gomez, Inessa Pawson, Sanket Verma, Madicken Munk - Review / approve / merge open PRs - [#266 — Stop dependabot from updating requirements](https://github.com/scientific-python/specs/pull/266) - [#264 — SPEC4 language improvements](https://github.com/scientific-python/specs/pull/264) - Increasing project outreach: matplotlib - Jarrod (?) will reach out to Kyle Sunden / attend community meeting - NF Project Summit 2023 - Juanita, Brigitta, Sanket, Madicken, and Inessa are attending. Maybe, they could host a break-out session about new SPECS. - Visualization - Napari - What could SPECs in that space look like? - Element of other SPECs? Accessibility / colorblind-friendliness, e.g. - [Altair](https://altair-viz.github.io/gallery/index.html), [Grave](https://github.com/networkx/grave) - Madicken / Matt McCormick: thinking about (pre- the formation of Scientific Python) a standardized API for viz libraries # August 1, 2023 Previous meeting notes available [here](https://github.com/scientific-python/scientific-python-archive). Please feel free to add topics for discussion and news items below. **In attendance:** Jarrod Millman, Brigitta Sipőcz, Ralf Gommers, Kristen Thyng, Mridul Seth, Pamphile Roy, Kira Evans, Stéfan van der Walt, Juanita Gomez, Lars Gruter, ## Monthly meetings Every month on the First Tue, until Dec 5, 2023, 5 occurrence(s) - Aug 1, 2023 09:00 AM - Sep 5, 2023 09:00 AM - Oct 3, 2023 09:00 AM - Nov 7, 2023 09:00 AM - Dec 5, 2023 09:00 AM Should we - post video recordings? (yes, starting next meeting) - add https://scientific-python.org/calendars/ @jarrodmillman Do we need to revisit structure - chair ? to lead meetings ? - Jarrod to chair the meetings for the initial ones. ## [Endorsements](https://scientific-python.org/specs) - NumPy is endorsing SPEC 4 (PR coming!) - SciPy endorsed [SPEC 0](https://github.com/scientific-python/specs/pull/244), [SPEC 1](https://github.com/scientific-python/specs/pull/245), and [SPEC 4](https://github.com/scientific-python/specs/pull/246) - what other core projects are ready to endorse @jarrodmillman will ping steering committee members involved with other core projects - xarray dev meeting tomorrow ## Existing SPECs - [Clarify distinction between SPEC and implementation](https://github.com/scientific-python/specs/issues/247) @tupui - Clarify what endorsement, implementation means / looks like - [Remaining TODOs on SPEC 2](https://github.com/scientific-python/specs/issues/229) @jarrodmillman @rgommers - [Review SPEC3](https://scientific-python.org/specs/spec-0003/) - It's a very high-level description of the aspects of accessibility that should be considered - Any aspects that are missing? - Maybe recommendations for accessible communication channels (e.g. e-mail as a fallback option to reach out and report bugs) - Q: does pydata-sphinx-theme already address all these aspects? - Some of it is planned work, also on Hugo theme - May be useful to know which sphinx themes are compatible ## [Core projects](https://specs.scientific-python.org/specs/core-projects/) - [Missing core libraries?](https://github.com/scientific-python/specs/issues/241) - [zarr](https://zarr.dev/) - Should we consider a participation-based core criterium? — [name=stefan] - Maybe there could be a lesser level of participation in SPEC that is more passive for communication. ## Team membership We are implementing [a system](https://github.com/scientific-python/sync-teams-action) for adding GH org members and team membership - https://github.com/scientific-python/teams/blob/main/teams.yaml - https://github.com/scientific-python/specs/tree/main/steering-committee - Should this be generated from GH team? ## New SPECs - I started writing up these documents without following [process](https://specs.scientific-python.org/specs/purpose-and-process/); how should I get them onto the "standards track"? — [name=stefan] - [SPEC 6: Keys to the Castle](https://github.com/scientific-python/specs/pull/168) - Includes two compatible workflows now: bitvault, git/gpg based - Missing principles: - Do not create passwords, if possible to avoid - Record *who* has access / is responsible for a specific token / service - Related to nightly wheels; did we come up with a good process for that? - [SPEC 7: seeding pseudo-random number generation](https://github.com/scientific-python/specs/pull/180) - Captures status of recommendation by NumPy and what is in scikit-image - scikit-learn is busy overhauling their interface - SciPy is [discussing](https://github.com/scipy/scipy/issues/14322) what to do (that discussion is what inspired the SPEC) - Re-connect with scikit-learn @stefanv - Generally, a more useful path for moving forward a SPEC: do the work in the project first, then ask to endorse the SPEC they're already following # Badges - https://github.com/scientific-python/specs/issues/240 - Henry did an example implementation: was on the Zoom chat, do we still have that Juanita? - https://img.shields.io/badge/SPEC-1-blue - https://img.shields.io/badge/SPEC_0-Compliant-green?labelColor=%23004811&color=%235CA038 - It would be nice to have a little tree logo next to it :) - [Add image to badge](https://github.com/badges/shields/issues/3612)

Syntax	Example	Reference
# Header	Header	基本排版
- Unordered List	Unordered List
1. Ordered List	Ordered List
- [ ] Todo List	Todo List
> Blockquote	Blockquote
Bold font	Bold font
Italics font	Italics font
~~Strikethrough~~	~~Strikethrough~~
19^th^	19^th
H~2~O	H₂O
++Inserted text++	Inserted text
==Marked text==	Marked text
[link text](https:// "title")	Link
![image alt](https:// "title")	Image
`Code`	`Code`	在筆記中貼入程式碼
```javascript var i = 0; ```	`var i = 0;`
:smile:		Emoji list
{%youtube youtube_id %}	Externals
$L^aT_eX$	L^aT_eX
:::info This is a alert area. :::	This is a alert area.