More appealing Ecosystem page. #246
Conversation
|
in class, just saying that I see this and will review it! |
|
Love the little sections here. As far as accessibility is concerned, it's not completely obvious these are expandable sections. In addition, since this is one of our main pages, if it's going to be a static html element that has the event bound it should also have a tabIndex. The other way to handle a11y for this is to use a button instead of a collapsible section (or both!) Generally, I love this yaml definition and generation approach! Looking great here! |
No hurry, I know you have plenty to do and this is not were I'd like to have your feedback first.
Yeah, right now it's just hiding using css opacity 0 – which has its drawbacks as the elements are still there, on medium screen with the tiled layout for example you can get : Unfortunately I'm a bit limited by Jekyll and don't want to try mixing a heavy JS framework with a templating engine.
I'm unsure I can understand what you mean here. |
|
Oh my gosh I read this so wrong @Carreau! You are using a search form control to filter elements. You can disregard everything I said. I was viewing this on mobile and thought it was a combo of search + expandable sections. The interaction I was hoping to make sure we weren't doing was applying interactivity to the individual |
|
Quite a bit of work and design additions. |
|
Sidenote, if we want to iterate on that we can deploy the page without listing it in the headers and footers, so it should not be referenced. |
|
I would like to review this PR before it is merged. |
Would you be ok (at some point) with merging this in non-listed in the nav bar, and then to reach to authors to edit the Also having actual data from what people want on the page, can help fleshing out the design. |
|
OK, getting to the review of this. First set of questions/comments are more about the content than visual design or code. First, I love to idea of having a community/ecosystem page. I like the word ecosystem for a webpage, as I think it captures what we are trying to communicate. Second, while commercial products that are based on jupyter of integrate with Jupyter are certaintly part of the overall ecosystem and are important to mention in this context. However, I think we should think about the criteria for adding links on this page, as in this context we are providing a detailed description of the product or service. Also, I want the folks looking at this page to understand very clearly what things are open-source and which things are commercial. I like the topical organization of this page (rather than organized by open-source versus commercial). So possible actions items related to this point are:
In the long run, it may be better to charge these companies to be featured here as "sponsored" links, but I am fine with it moving forward for now, as long as as we are not biased towards particular companies overly. Third, I think we need to clearly distinguish between things that are jupyter open-source projects from those that are in the broader ecosystem. This is important from multiple angles:
There are different ways of handling this. We could remove all the Jupyter things from this page and put a link to our documentation page (and work on improving it). We could also create another badge that indicate something is "official jupyter" Let's work through this stuff before starting to iterate on the details of the wording or visual design. |
Thanks, but credits goes to Thomas, I was struggling to find a name.
I see your PoV, though I would really like to avoid having both commercial/free core-jupyter/not-core. In my idea this page should list all the things, with an arbitrary order (say alphabetically), with potentially a rough "is still maintained" criteria. While I see the temptation to "feature" either commercially, or "core" components, I believe this should be done in a separate page. One of the reason that lead me to create this page is a single place to list and find jupyter-related project, having a selection criteria make it way less useful, and if we end up with 1 jupyter-core, 1 sponsored, and 1 other page, then there will be the temptation to create a 4th page the regroup all of these and allow to filter. Now I am not against the idea of adding field(s) we can search for, I tend to not like "sponosored" links either, but I see the appeal.
That make sens, though I'd like to keep that simple as a first iteration. I'd like to get adoption first before moving to a more automatized system, if we really get a large list with multiple fields we'll have to move to a non-static website.
For the first iteration I'll go with anyone that bother to submit a PR. I won't name companies that spend more time sending me an email asking me to do things instead of finding an engineer capable of sending a 1 line PR. This feel (IMHO) like an enough threshold.
Yes, I filled some to have an idea of how it would look, still better than random fake brand names and logos. There is of course the question of list order; we could randomize at each page load, but that would be bad for users searching things.
Good, I agree. Would Alphabetical order for now suits you ? We can discuss later what "Sponsored" could mean.
These are some good points, i'd still would like to have this page as open as possible without restrictions on what can be added beyond obviously illegal content. We could of course add a "Controversial" flag to hide by default some items, but for user convenience I would really like to not discriminate projects.
I think this is again going a bit too far. The what is "official jupyter" is already a gray area, like MyBinder for example;
I think we are starting to complicate things too much. While these things would be great, better is the enemy of good. If you start to add "sponsored", or "Core Jupyter", you highly risk to add reason for tension in the team, and dispute, any potential adding of any item in here will end up begin an infinite bikeshed if it worth it and correctly classified. Most repository like NPM, PyPI, conda-forge, ... etc already allow you to register without prior approval, and contain – in an indiscriminate manner – both project from the community and core member. These are hosted on official domain, and there is no confusions that these packages are not vetted by the PyPA. We are not trying to centralize and overwhelmed current developers more, and on the opposite are trying to involved the community. For now I would go for the following rule:
That is IMHO, more than enough, and has been quite battle tested (conda forge, Python3statement...). |
|
* Within a category, I am fine with an alphabetical ordering - it is as
good as any
* Also fine with an informal and liberal process for content to be included
here. I just want us to document it in the README.
* We should not list any organization and use their logo without explicit
permission. For example, we were very careful get to explicit approval to
use all the logos of companies and universities on the home page of this
website. We even signed usage agreements with some companies. Logos and
names of companies are owned by those companies and we can't just use them
without permission.
* In general, the of question of "what is official jupyter" is not a grey
area. We have a formal governance, non-profit affiliation, copyright,
licensing, code of conduct, etc. Those things either apply or they don't.
The specific case of Binder is a special case that needs serious attention.
* I am fine with a merge and iterate approach, but before this is merged I
would like to see *some* way of indicating the open-source/commercial and
jupyter/non-jupyter items. I think this can initially be really minor -
could be text or a small badge by each item.
* Fine waiting on "sponsored" content.
…On Sat, Jan 20, 2018 at 3:18 PM, Matthias Bussonnier < ***@***.***> wrote:
First, I love to idea of having a community/ecosystem page. I like the
word ecosystem for a webpage, as I think it captures what we are trying to
communicate.
Thanks, but credits goes to Thomas, I was struggling to find a name.
However, I think we should think about the criteria for adding links on
this page, as in this context we are providing a detailed description of
the product or service. Also, I want the folks looking at this page to
understand very clearly what things are open-source and which things are
commercial. I like the topical organization of this page (rather than
organized by open-source versus commercial)
I see your PoV, though I would really like to avoid having both
commercial/free core-jupyter/not-core. In my idea this page should list all
the things, with an arbitrary order (say alphabetically), with potentially
a rough "is still maintained" criteria. While I see the temptation to
"feature" either commercially, or "core" components, I believe this should
be done in a separate page.
One of the reason that lead me to create this page is a single place to
list and find jupyter-related project, having a selection criteria make it
way less useful, and if we end up with 1 jupyter-core, 1 sponsored, and 1
other page, then there will be the temptation to create a 4th page the
regroup all of these and allow to filter.
Now I am not against the idea of adding field(s) we can search for, I tend
to not like "sponosored" links either, but I see the appeal.
Some visual indicator (maybe badges or an icon?) that indicate
open-source, versus commercial offerings. It would be good to have
something that we could indicate that a particular things is both (many of
these things have open-source and commercial sides).
That make sens, though I'd like to keep that simple as a first iteration.
I'd like to get adoption first before moving to a more automatized system,
if we really get a large list with multiple fields we'll have to move to a
non-static website.
Some text in the README proposing who can list here and who needs to
approve. Do we accept submissions?
For the first iteration I'll go with anyone that bother to submit a PR. I
won't name companies that spend more time sending me an email asking me to
do things instead of finding an engineer capable of sending a 1 line PR.
This feel (IMHO) like an enough threshold.
Before we post any commercial entities here I want to make sure we are not
favoring some over the others. For example, if we are going to list
commercial entities that provide hosted Jupyter notebooks, we should
mention all of them without bias (Google Cloud DataLab, Colaboratory,
CoCalc, AWS SageMaker, IBM DSE, Dataiku, gryd, etc.) and make it clear to
others that they can add theirs through some process.
Yes, I filled some to have an idea of how it would look, still better than
random fake brand names and logos. There is of course the question of list
order; we could randomize at each page load, but that would be bad for
users searching things.
In the long run, it may be better to charge these companies to be featured
here as "sponsored" links, but I am fine with it moving forward for now, as
long as as we are not biased towards particular companies overly.
Good, I agree. Would Alphabetical order for now suits you ? We can discuss
later what "Sponsored" could mean.
Trademark - we now have a trademark and we need to make sure there is no
confusion about what things are covered by the trademark and which are not
CoC - Jupyter open-source projects are covered by our Code of Conduct and
I think that should be part of our "brand".
These are some good points, i'd still would like to have this page as open
as possible without restrictions on what can be added beyond obviously
illegal content. We could of course add a "Controversial" flag to hide by
default some items, but for user convenience I would really like to not
discriminate projects.
There are different ways of handling this. We could remove all the Jupyter
things from this page and put a link to our documentation page (and work on
improving it). We could also create another badge that indicate something
is "official jupyter
I think this is again going a bit too far. The what is "official jupyter"
is already a gray area, like MyBinder for example;
Let's work through this stuff before starting to iterate on the details of
the wording or visual design.
I think we are starting to complicate things too much. While these things
would be great, better is the enemy of good. If you start to add
"sponsored", or "Core Jupyter", you highly risk to add reason for tension
in the team, and dispute, any potential adding of any item in here will end
up begin an infinite bikeshed if it worth it and correctly classified.
Most repository like NPM, PyPI, conda-forge, ... etc already allow you to
register without prior approval, and contain – in an indiscriminate manner
– both project from the community and core member. These are hosted on
official domain, and there is no confusions that these packages are not
vetted by the PyPA.
We are not trying to centralize and overwhelmed current developers more,
and on the opposite are trying to involved the community.
For now I would go for the following rule:
- Make sure the project author is ok with their project being listed
here.
- Have a quick sanity check that the project does indeed exist, works
and is maintained, on which we trust the merger to check that.
- Accept only PRs to add projects, that is to say no issues with
"pease add..." (to make sure authors care).
That is IMHO, more than enough, and has been quite battle tested (conda
forge, Python3statement...).
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#246 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABr0PR3k-0n1mDX7xISIWefScshZufwks5tMnQ_gaJpZM4RYieo>
.
--
Brian E. Granger
Associate Professor of Physics and Data Science
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
bgranger@calpoly.edu and ellisonbg@gmail.com
|
|
That suits me. I'll work on implementing that later.
…On Jan 20, 2018 16:57, "Brian E. Granger" ***@***.***> wrote:
* Within a category, I am fine with an alphabetical ordering - it is as
good as any
* Also fine with an informal and liberal process for content to be included
here. I just want us to document it in the README.
* We should not list any organization and use their logo without explicit
permission. For example, we were very careful get to explicit approval to
use all the logos of companies and universities on the home page of this
website. We even signed usage agreements with some companies. Logos and
names of companies are owned by those companies and we can't just use them
without permission.
* In general, the of question of "what is official jupyter" is not a grey
area. We have a formal governance, non-profit affiliation, copyright,
licensing, code of conduct, etc. Those things either apply or they don't.
The specific case of Binder is a special case that needs serious attention.
* I am fine with a merge and iterate approach, but before this is merged I
would like to see *some* way of indicating the open-source/commercial and
jupyter/non-jupyter items. I think this can initially be really minor -
could be text or a small badge by each item.
* Fine waiting on "sponsored" content.
On Sat, Jan 20, 2018 at 3:18 PM, Matthias Bussonnier <
***@***.***> wrote:
> First, I love to idea of having a community/ecosystem page. I like the
> word ecosystem for a webpage, as I think it captures what we are trying
to
> communicate.
>
> Thanks, but credits goes to Thomas, I was struggling to find a name.
>
> However, I think we should think about the criteria for adding links on
> this page, as in this context we are providing a detailed description of
> the product or service. Also, I want the folks looking at this page to
> understand very clearly what things are open-source and which things are
> commercial. I like the topical organization of this page (rather than
> organized by open-source versus commercial)
>
> I see your PoV, though I would really like to avoid having both
> commercial/free core-jupyter/not-core. In my idea this page should list
all
> the things, with an arbitrary order (say alphabetically), with
potentially
> a rough "is still maintained" criteria. While I see the temptation to
> "feature" either commercially, or "core" components, I believe this
should
> be done in a separate page.
>
> One of the reason that lead me to create this page is a single place to
> list and find jupyter-related project, having a selection criteria make
it
> way less useful, and if we end up with 1 jupyter-core, 1 sponsored, and 1
> other page, then there will be the temptation to create a 4th page the
> regroup all of these and allow to filter.
>
> Now I am not against the idea of adding field(s) we can search for, I
tend
> to not like "sponosored" links either, but I see the appeal.
>
> Some visual indicator (maybe badges or an icon?) that indicate
> open-source, versus commercial offerings. It would be good to have
> something that we could indicate that a particular things is both (many
of
> these things have open-source and commercial sides).
>
> That make sens, though I'd like to keep that simple as a first iteration.
> I'd like to get adoption first before moving to a more automatized
system,
> if we really get a large list with multiple fields we'll have to move to
a
> non-static website.
>
> Some text in the README proposing who can list here and who needs to
> approve. Do we accept submissions?
>
> For the first iteration I'll go with anyone that bother to submit a PR. I
> won't name companies that spend more time sending me an email asking me
to
> do things instead of finding an engineer capable of sending a 1 line PR.
> This feel (IMHO) like an enough threshold.
>
> Before we post any commercial entities here I want to make sure we are
not
> favoring some over the others. For example, if we are going to list
> commercial entities that provide hosted Jupyter notebooks, we should
> mention all of them without bias (Google Cloud DataLab, Colaboratory,
> CoCalc, AWS SageMaker, IBM DSE, Dataiku, gryd, etc.) and make it clear to
> others that they can add theirs through some process.
>
> Yes, I filled some to have an idea of how it would look, still better
than
> random fake brand names and logos. There is of course the question of
list
> order; we could randomize at each page load, but that would be bad for
> users searching things.
>
> In the long run, it may be better to charge these companies to be
featured
> here as "sponsored" links, but I am fine with it moving forward for now,
as
> long as as we are not biased towards particular companies overly.
>
> Good, I agree. Would Alphabetical order for now suits you ? We can
discuss
> later what "Sponsored" could mean.
>
> Trademark - we now have a trademark and we need to make sure there is no
> confusion about what things are covered by the trademark and which are
not
> CoC - Jupyter open-source projects are covered by our Code of Conduct and
> I think that should be part of our "brand".
>
> These are some good points, i'd still would like to have this page as
open
> as possible without restrictions on what can be added beyond obviously
> illegal content. We could of course add a "Controversial" flag to hide by
> default some items, but for user convenience I would really like to not
> discriminate projects.
>
> There are different ways of handling this. We could remove all the
Jupyter
> things from this page and put a link to our documentation page (and work
on
> improving it). We could also create another badge that indicate something
> is "official jupyter
>
> I think this is again going a bit too far. The what is "official jupyter"
> is already a gray area, like MyBinder for example;
>
> Let's work through this stuff before starting to iterate on the details
of
> the wording or visual design.
>
> I think we are starting to complicate things too much. While these things
> would be great, better is the enemy of good. If you start to add
> "sponsored", or "Core Jupyter", you highly risk to add reason for tension
> in the team, and dispute, any potential adding of any item in here will
end
> up begin an infinite bikeshed if it worth it and correctly classified.
>
> Most repository like NPM, PyPI, conda-forge, ... etc already allow you to
> register without prior approval, and contain – in an indiscriminate
manner
> – both project from the community and core member. These are hosted on
> official domain, and there is no confusions that these packages are not
> vetted by the PyPA.
>
> We are not trying to centralize and overwhelmed current developers more,
> and on the opposite are trying to involved the community.
>
> For now I would go for the following rule:
>
> - Make sure the project author is ok with their project being listed
> here.
> - Have a quick sanity check that the project does indeed exist, works
> and is maintained, on which we trust the merger to check that.
> - Accept only PRs to add projects, that is to say no issues with
> "pease add..." (to make sure authors care).
>
> That is IMHO, more than enough, and has been quite battle tested (conda
> forge, Python3statement...).
>
> —
> You are receiving this because your review was requested.
> Reply to this email directly, view it on GitHub
> <#246#
issuecomment-359209955>,
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/AABr0PR3k-
0n1mDX7xISIWefScshZufwks5tMnQ_gaJpZM4RYieo>
> .
>
--
Brian E. Granger
Associate Professor of Physics and Data Science
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
***@***.*** and ***@***.***
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#246 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAUez755D2nK6D0IK7JEEaSHnIg9RWayks5tMotvgaJpZM4RYieo>
.
|
While the docs page is nice and list (some) of the community related project is is a bit tough to navigate for some one new to Jupyter. This is a proposal to provide a more vibrant page, easier to get contribution to (that is to say you do not have to write html). Allow people to have logos, links, description, as well as a Basic Search. I structured with the Kernels at the top, as I believe this is really what most people will be interested in. I also explicitely list things like services hosting Jupyter.
One narrow screen this will be a single column. Sections now have headers with logo, and items that have no provided logo will use the generic section header. The search is a bit better (space is used as an AND between terms), each item get also tag with the name of the section it's in so you can filter by section as well. The css should allow to pretty easily decide what get shown/hidden when searching. I vent with hiding the sections descriptions and logo (as when you search you want compact results) but keep the sections otherwise they might be hard to discover.
|
@ellisonbg, do you want to ping Tim to have him look on the visual design ? I added the following :
Sponsored/featured so that we can distinguish our projects, the one we like and the sponsored one. The the attached screenshot. The IJulia is dark gray as my (invisible) cursor is hovering.
Some draft of instructions at the end on how to submit a project and the conditions. Once we are happy-ish, I'll squash all and remove any project we don't control. |
|
Hi @Carreau it would be great it Tim (@tgeorgeux) can help do a design review of this. I will add him to the relevant GitHub teams so we can assign him to things, but consider him @ mentioned for now. I will help Tim get a dev build of the website going so he can check it out locally as well. |
|
Great ! Thanks ! I recently updated the instructions to get the website running, so it should be pretty strait forward. The new instructions should automatically include a live reload plugin, so you can "just" write code and see the website updates. |
|
Also @tgeorgeux, nice to meet you ! |
|
@Carreau Tim may need help in getting a dev build of the website going along with the build tools needed. I can help with this on Wed if needed. |
Happy to help if needed, as I said above I recently reviewed and updated he instructions on how to get a dev version of this website up and running on a local machine. The readme section is pretty self explanatory, but if there are hiccups, I'm happy to improve the instructions, and help debug. |
|
@Carreau I'll try to get the dev build running on my machine in the morning
and I'll ping you if I have any troubles.
…On Mon, Jan 22, 2018 at 4:46 PM, Matthias Bussonnier < ***@***.***> wrote:
@Carreau <https://github.com/carreau> Tim may need help in getting a dev
build of the website going along with the build tools needed. I can help
with this on Wed if needed.
Happy to help if needed, as I said above I recently reviewed and updated
he instructions on how to get a dev version of this website up and running
on a local machine. The readme section
<https://github.com/jupyter/jupyter.github.io#quick-local-testing> is
pretty self explanatory, but if there are hiccups, I'm happy to improve the
instructions, and help debug.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#246 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AVFcwDnPrNRlkhArcN_FxShePunYShVzks5tNSvNgaJpZM4RYieo>
.
|
|
Add a "Binder link" where relevant. |
|
Closing this for now, until we figure out the future path forward for the website. |
While the docs page is nice and list (some) of the community related
project is is a bit tough to navigate for someone new to Jupyter.
You also don't always "reach for the docs" . Imho, Docs are useful once you have found your project.
This is a proposal to provide a more vibrant page, easier to get
contribution to (that is to say you do not have to write html).
Allow people to have logos, links, description, as well as a Basic
Search.
I structured with the Kernels at the top, as I believe this is really
what most people will be interested in.
I also explicitely list things like services hosting Jupyter.
It's late so I submit that as a PR and will rework on it tomorrow.