Absorb the Phosphor Project

[blink1073]

The Phosphor project has been archived by the author, and the name is no longer usable. Here we propose to bring the existing code base as a new repository under the JupyterLab GitHub org, and give it a new name. I say we give a week for folks to give suggestions on a new name.

[GordonSmith]

I would suggest using @jupyterlab-phosphor/algorithm (etc.) - While I suspect "phospor" can't be copyrighted, I would still seek written permission from @sccolbert first and if not given, then consider switching - but it is/was his baby and keeping the name will help preserve his legacy on the project.

I have a few housekeeping PRs I use on my fork which may be of interest:

• Add typescript types to published package (already there, not sure why I thought they were missing)
• UMD / es6 Support
• Updated to latest TS Compiler
• Use NPM instead of YARN (somewhat opinionated, but Yarn was masking some missing dependency issues between the projects)
• Removal of unofficial dom functions (setImmediate / clearImmediate)
• Some minor enhancements to the DockPanel
• etc.

Plus there is a whole ecosystem of visualizations which play well with the DockPanel (Shameless plug):

• https://raw.githack.com/hpcc-systems/Visualization/master/demos/gallery/gallery.html?./samples/layout
• https://raw.githack.com/hpcc-systems/Visualization/master/demos/gallery/gallery.html

[bollwyvl]

Thanks for starting this discussion. Good feature ideas, but I think with all the things in motion right now, what we mostly need is clear, official messaging that Jupyter can provide (even if we haven't done it yet): - re-released packages of the dependencies in question of current lab heads (presumably, 1.1.*, 1.2.*, 2.*) that downstreams can use confidently with an identical API. - a review/ci infrastructure ready to accept contributions and react to them in a timely manner - a readiness (indeed, enthusiasm) to respect the needs of other contributing downstreams - some tooling which can help with the transition The worst things that could happen to lab adoption until all that is in place is users and downstreams, due to fork FUD, rather than on a rational assessment of the work and the investment it represents... - stopping work - starting more, private, or otherwise incompatible-for-no-reason forks of the libraries in question or lab - picking something else to base their work on - not starting new work at all, hurting users PR (ha) aside: There are obvious advantages to moving all the libraries in question into the jl/jl monorepo, and this can be done in a way that preserves history while still respecting the wishes of all involved... e.g. git filter branch. I hope we do maintain (therefore test) a hard separation of deps, though, as allowing react/blueprint/d3/etc to infect the otherwise unopinionated dependency tree would reduce its adoptability. Because of the collisions with @jupyterlab (e.g coreutils), it seems we need a namespace, irrespective of where it lives. Ideally, something that is short, isn't too clever or derivative, on-brand, without being too lab-centric might have the greatest chance of getting outside contribution. Haven't don't the brand/seo drive-though but perhaps some things like @-on-npm: - jupyterkit - jupytoolkit - jupytypes - jupycore - jupycleus - jptk - jpui - jpyde - jig I dunno. It may also be worth opening this up to the various mailing lists, which get some more eyes. I'd be happy help with the rebrand, once we get to that point, having done some work on the OG stuff back in the day. We likely will still want dedicated docs, examples, etc. but can be a bit more liberal with cross linking to lab for examples, which may well be the only in-the-wild use of a particular api.

…

On Sat, Nov 9, 2019, 02:18 Gordon Smith ***@***.***> wrote: I would suggest using @jupyterlab/phosphor - While I suspect "phospor" can't be copyrighted, I would still seek written permission from @sccolbert <https://github.com/sccolbert> first and if not given, then consider switching - but it is/was his baby and keeping the name will help preserve his legacy on the project. I have a few housekeeping PRs I use on my fork which may be of interest: - Add typescript types to published package - UMD / es6 Support - Updated to latest TS Compiler - Use NPM instead of YARN (somewhat opinionated, but Yarn was masking some missing dependency issues between the projects) - Removal of unofficial dom functions (setImmediate / clearImmediate) - Some minor enhancements to the DockPanel - etc. Plus there is a whole ecosystem of visualizations which play well with the DockPanel (Shameless plug): - https://raw.githack.com/hpcc-systems/Visualization/master/demos/gallery/gallery.html?./samples/layout - https://raw.githack.com/hpcc-systems/Visualization/master/demos/gallery/gallery.html — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#28>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALCRDLUSJRICUYKXS23FLQSZP2TANCNFSM4JLDJKNA> .

[timkpaine]

I think its important to keep it as separate as possible, there are plenty of applications that rely on phosphor that don't want to pull in a bunch of jupyter-specific code.

[timkpaine]

Please avoid antagonistic comments, @sccolbert's contributions to this project are extensive and appreciated, and any questions/comments about reasonings for archiving the phosphorjs project should be taken up individually.

[bollwyvl]

Sure, thanks for cleaning my comment up @timkpaine. We've all invested a lot in this work over the last n years, and I'm personally grateful to everyone I've had the opportunity to work with during that time around the code.

Just trying to point out that a potential downstream would interpret the current guidance as a 🔴 🎩 -style-trademark-thing, rather than a BSD-conformance-thing, underlining the need to get something relatively official out soon which fully complies with the guidance.

[ian-r-rose]

Here are few possible ideas on renaming/rebranding. I mostly went down a chemistry-related allusion hole, so take these as seriously/not-seriously as you like:

• Phosphate is a component of the mineral group apatite, which forms a major structural element of vertebrate skeletons.
• Phosphates form a structural part of DNA and RNA. Perhaps one of the nuclide bases could be a suitable name. Many of them have a sort of medicinal flavor to them, but I kind of like the idea of adenine.
• In a similar mineralogical vein: monazite is a phosphate mineral which is used by geochronologists to determine the age of various rocks. Perhaps not as nice of a metaphor, but a fun word.
• The Great Red Spot is a storm that has been storming on Jupiter (note the i) for centuries, and is big enough to swallow several Earths. We don't know for sure why it is red, but one theory involves photochemistry with phosphine. So something like red spot could be a nice name, with the downside that the connection could be proven incorrect with further data from the Juno mission.

[timkpaine]

@ian-r-rose i was thinking in the same vein, phosphor is an old name for phosphorus, so I was looking at other old chemical names.

Could also go for a moon https://en.m.wikipedia.org/wiki/Moons_of_Jupiter

[jasongrout]

More thoughts on the connection with light:

• lumenjs
• flourescent (hard to spell)

More elements close to it on the periodic table:

• xenon
• argon
• neonjs

[bollwyvl]

Picking something in the same group as would reflect compatibility:

- nitrogen, n (taken)

- arsenic, as (taken) - antimony, an (open) - used in batteries, 🔥 retardant, micro electronics - css classes would be funny, an-Widget - not in it for the mon(e)y - bismuth (taken) - moscovium (open but meh) There are archaic forms of the above as well...

[telamonian]

"What's Jupiter made of?"

Roughly 90% hydrogen and 10% helium. So I would say we should hydrogen as the new name, but someone beat us to that...

[fperez]

@jasongrout already pointed out (we had this conversation elsewhere) that the name candela is already taken in npm, so it doesn't work. But I'm still going to write down the idea below just in case the direction of thinking is useful to someone to come up with another one... Naming is tricky and it's all about spurring ideas, so even ones that don't work can be useful :) My original note follows:

In Colombia, 'fósforo' is both the element "phosphor" and the colloquial name for match (as in the little wooden sticks with flammable bit at the end), b/c the flammable tip of matches typically has a lot of phosphor in it (or used to, I have no idea how they're made today). And the name for a fluid lighter is --again colloquially-- "candela", which more precisely is a synonym for fire.

The "proper" Spanish words are cerilla (match) and encendedor (lighter), but in Colombia everybody uses fósforo and candela.

The word candela actually exists in English, it's the the SI unit for luminous intensity (meaning it's physical radiance but integrated with the human visual model of wavelength-dependent sensitivity).

[jasongrout]

Some more suggestions from our meeting today:

• lumino
• sol
• solar
• lightstick
• glowstick
• glow
• lumo (esparanto)
• lux (latin)
• luce (italian)
• marama (maori)
• lumio

[jasongrout]

When we were discussing it, "lumino" kept coming back as a favorite. What do people think about calling the new project Lumino? (feel free to use thumbs up/down, or leave a comment for a nuanced response)

Reasons for Lumino:

• It's pleasant and easy to say, sort of rolls off the tongue
• It's available on npm as an org
• It's Italian for 'a small lamp or light'
• It's the Latin verb for:
  1. brighten (w/color)
  2. illuminate, give light to
  3. light up
  4. reveal/throw light on

And as one person said, it's cuter than "lux"

[jasongrout]

Also, someone suggested the proper Spanish word for "bright": luminosa

[tgeorgeux]

Lumino also comes from "Luminophore", which has two types, Phosphor and Fluorophore. Kind of a cool callback to phosphor as well :) https://en.wikipedia.org/wiki/Luminophore

…

On Wed, Nov 13, 2019 at 3:34 PM Jason Grout ***@***.***> wrote: Also, someone suggested the proper Spanish word for "bright": luminosa — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#28>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFIVZQGVKKJSQRXPXXNAXS3QTSFJBANCNFSM4JLDJKNA> .

[bollwyvl]

👍 to lumino, but @lumino is taken on npm. Here's a tool for checking availability, since npm makes it kind of hard through the UI. Of the above, @marama is available. If we throw a y at some of them, it opens up a bit, and likely better SEO, anyhow. lumyno kinda sounds like a dynamo, which is cool.

[bollwyvl]

(sry, just didn't look hard enough, you can check by URL, e.g. https://www.npmjs.com/org/lumino

[jasongrout]

but @lumino is taken on npm.

By @blink1073, this morning, just in case :)

[choldgraf]

[tgeorgeux]

I am +1 on Lumino. Good looking out @blink1073 ! —

…

You are receiving this because you commented. Reply to this email directly, view it on GitHub <#28>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFIVZQCVUAFB7XSU5JXJ77DQTS463ANCNFSM4JLDJKNA> .

[jasongrout]

More bike-shedding: this is what the package list would look like:

@lumino/algorithm
@lumino/application
@lumino/collections
@lumino/commands
@lumino/coreutils
@lumino/datagrid
@lumino/datastore
@lumino/default-theme
@lumino/disposable
@lumino/domutils
@lumino/dragdrop
@lumino/keyboard
@lumino/messaging
@lumino/properties
@lumino/signaling
@lumino/virtualdom
@lumino/widgets

[bollwyvl]

🚵‍♂️ 🏠 ed some font ideas

it's nice that https://raw.githack.com is picking up for rawgit... i got 503ed vs the mothership

some criteria:

• open source (OFL or equivalent)
• has an l that doesn't look like a 1, or just look cool

On individual fonts:

• I particularly like the interplay between the serifs on Flamenco: those could be expanded into boxes suggesting dockpanel

[bollwyvl]

As for the inhabitants of the bikeshed:

I'd love to bring in a jo-like treatment for this... unsurprisingly lu. Lu's probably a bit more boxy than jo, maybe a little more transforming-pet-like, and helps jo get things done. I can imagine a team of lus:

• delivering messages
• sending signals
• interacting with algorithms
• shopping at the data store
• playing hide-and-seek/brady-bunch/hollywood-squares in a datagrid
• plugging together apps
• typing on a keyboard like ghost in the shell
• dragging/dropping around DOM with a claw game
• looking at virtual DOM in a visor
• listening for properties
• disposing of stuff (responsibly)
• painting with a command palette

[bollwyvl]

When rebooting the docs, it would perhaps be a good time to revisit sphinx-js for the autogenerated parts, which should likely be considered for lab as well.

Also, it would be very interesting to make a mini-jyve for dogfooding the docs with interactive examples... might as well go ahead and ship the typescript interepreter, which basically has all the nuts and bolts for a language server experience, especially with monaco.

[JohanMabille]

+1 on Lumino.

[blink1073]

I think the consensus if for lumino. I'll give this another 24 hrs for anyone to disagree and then go ahead and start the migration. Cheers!

[jasongrout]

I think the consensus if for lumino. I'll give this another 24 hrs for anyone to disagree and then go ahead and start the migration. Cheers!

Sounds good to me. +1.

[jasongrout]

IIRC, we were going to release packages with exactly the same commit as the last released phosphor packages to make the transition for anyone literally a name change (with no code changes)?

[blink1073]

Yes, that sounds right

[blink1073]

I created the repo and made a release of all packages. I did the following:

• Updated naming, contributing guidelines, and licensing
• Removed auto-generate docs for now - I think we should target RTD

[vidartf]

@blink1073 Thanks for handling the logistics here! Would you now say that the process is completed, or are there any remaining tasks?

[blink1073]

Sounds good, I created jupyterlab/jupyterlab#7534 to track the rest of the transition.

[meeseeksmachine]

This issue has been mentioned on Jupyter Community Forum. There might be relevant details there:

https://discourse.jupyter.org/t/is-phosphorjs-no-longer-supported/3451/2