a-cosmology-group/acg#3

‘free’ ACG GitHub-organization must be owned by paid user for functionality

Indeed, all platforms are sliding into higher-cost models. It used to be that public and open source repos were free and essentially full-featured, while private repos were paid. I am continually surprised by the feature erosion of the free tiers 😞.

new ‘redshift’ email @my-emailbox is invalid for .github/FUNDING.yml

As far as I understand, an email is not a valid funding route under the sponsor button (little heart, top right). There are a number of supported funding platforms (see list in FUNDING.yml, perhaps consider one of them?). There is a custom: element for custom URLs. That is where I put a (bit.ly) link to your paypal donation page. All files in the repo are public, and similar to the web in general, repositories can be (are!) scrubbed for personal info. Best is to use URLs and your @ username handles. If you ever post sensitive info by mistake, let me know, there is way to remove stuff from the repo history, with some gymnastics 😄.

Backup of acg on GitHub

The website is “backed up” de facto, since it is just the content of the repo (the html pages can be built offline with similar tools that github uses behind the scenes). Discussions however are an interface perk that are not part of the repo itself. I will check how to connect to the github discussions API to download the content of discussions as a local backup. For EGSnrc we have scripts to preserve Issues and Pull Requests daily as a safeguard if we ever move away from github.

Separate groups

Understood (I think 🤔), let’s try it!

Separate groups

• Create repos acg and cg under organization a-cosmology-group.
• Create corresponding teams acg and cg.
• Tune permissions for teams and repos

@marmetl I put your email address in the Code of Conduct, you may want to remove that and use a throwaway email instead (e.g., report.acg@gmail.com). Let me know, I can remove your email from the repo history at all.

The website is “backed up” de facto, since it is just the content of the repo (the html pages can be built offline with similar tools that github uses behind the scenes).

Ha! That’s because you assume I use ‘GitHub Desktop’! I don’t, yet, but I will install it!

Oh, I only meant there is no need to backup the pages rendered in html format. I don’t use any git Desktop tool, only the terminal and Visual Studio Code. Simple, efficient!

What’s the “terminal”? …Oh you mean the command line interface CLI?

Yes, the CLI. I don’t know what to call it anymore! Linux users call it the CLI or the console, and macOS users the Terminal. 🤔

Great to see you here @sahil5d ! I haven’t graduated to owner, I can’t invite you yet… but coming soon!

@sahil5d I sent you an invite to the acg team.

@ftessier what you suggest above doesn’t seem to be what ACG needs. You are thinking “one group of people who work together”, which ACG is not! What is needed is to reproduce the dual-group discussion structure of Gaggle.email, so that cg members can have discussions on their side, and acg members can have their discussion separately. cg members do not want to receive more than one email/day, and only receive messages from other cg members.

Perhaps this is not possible in a single organization, since GitHub is designed for collaboration - let me know.

Otherwise it may be necessary to create another organization (e.g. “cosmology-group”) with its own members. Then “cosmology-group” members could also ask to join “a-cosmology-group” if they want to work on the acg repo and wiki.

Got it, there are two distinct groups. The confounding factor is the public/private aspects of the repos and discussions. Right now, both repos are public, and so are their respective discussion areas. This means that anyone with a github account can create new discussions, or comment on existing ones.

The only way to reserve the repo and its discussions to a given group is to make the repository private. One compromise alternative is to reserve the right to create new discussions to members with triage permission (in each repo), but anyone can still comment on existing discussions.

Perhaps the way to go is to leave acg open, and make cg private?

Yes, we could create two organizations, but at this point if there is just one repo in each org, then I would just use a repo for each, without the organization overhead. The acg could be public, and cg private. Will you want to store files in the cg repo, or just use the discussions board?

… anyone with a github account can create new discussions, or comment on existing ones

A good feature for acg members.

… we could create two organizations, but at this point if there is just one repo in each org,…

Only one repo is necessary for all members. The current acg repo is perfect since there is only one website for A Cosmology Group and all files can be shared with everybody. Also, only one wiki is necessary for A Cosmology Group.

Will you want to store files in the cg repo, or just use the discussions board?

No, there are no files to store in the current cg repo. We only need the discussion board on cg to keep cg members isolated from everything that’s happening in acg.

Perhaps the way to go is to leave acg open, and make cg private?

This sounds good. Would the two ‘Discussions’ boards be independent? That is, cg members would only receive email messages for discussions on cg, but not for discussions on acg?

I’m going to create a second account soon so we can test how this works.

Yes, the separate repos are independent, each with its discussion thread. People will only be notified of discussions in the repo they are watching. If members of ACG know not to watch or post in the cg discussions, and abide by that, you can leave both open. But you can also turn cg private, so that only the named collaborators (or team) can view and participate in discussion. I presume this is the most straightforward alternative. I will change the visibility of the cg repo, so you can see what that entails.

Done, cg is now private, hence only visible to the cg teams, to which you can add members. This implies that cg people need a github account to view the content.

I have never used a repo just for the discussions, but why not? Plus, in time, if cg members want to store and share files, the repo is already there. I suggest editing the top-level README.md to guide new users towards the Discussions are initially.

There can be more than one owner; I read here

<<Organization owners have complete administrative access to your organization. This role should be limited, but to no less than two people, in your organization.>>

My question is, do both owners need a paid account? If not then you can add me as an owner, but if so then you can transfer ownership later to the next owner, no worries.

Ah, I did not know. I made you an owner, fantastic!

Now let’s hope we haven’t lost any features… ;-)

2FA enabled for my account, go ahead!

Done!

You might want to add people to the cg team and give read access to the team (which it currently has), instead of adding them to the repo itself. I find it a little easier and clearer to manage teams compared to repo permissions for individuals. For example, if you want later to create a team with write access, you can create a new team, which is easier than changing the permissions of each collaborator that should have write access. My two cents.

For example, you can add @marmetl2 (who is that fraud?) to the cg teams, which will provide read access on the cg repo.

I’m having problems understanding the hierarchy. Ok I get it now, I think! I’m just going to play with this a bit more.

I see cg is listed as a member of a-cosmology-group/cg. Does that mean “the team cg is part of the a-cosmology-group/cg repo? Louis Marmet is the admin, and A. Louis Marmet is an “outside collaborator”, that seems good. You (@ftessier ) are not listed, which means you can be the owner or creator of cg but not a member?

I just realized I can: 1- add people to the cg repo, Then the member appears as an “outside collaborator”. 2- add people to the cg team (as a “collaborator”??) If I add people to the cg team, will they have access to the repo? 3- add people to the a-cosmology-group organization (as a “member”??) Then “A.L. Marmet” will get email notifications for every message in the acg repo (and team)? We don’t want that.

@marmetl2 received an email notification about @doctorlard ’s post on acg - not the desired behaviour. I will add @marmetl2 to the cg team instead…

Organization

For the record, a github organization contains (amongst other things):

• Repositories (public or private)
• People (github accounts, invited by the owner, that are part of the organization)
• Teams (groups of People, by invitation by the owner)

Each of these is a tab near the top of the organization page.

Think of nrc-cnrc/: People is a pool of all staff (that requested access in this case). Then there are separate repos, and there is a team of developers for each repo (my choice). Or in corporate parlance:

• the organization is the company
• the people are all the staff
• the teams could be: Sales, Marketing, IT, Finance, etc.
• and the repos could be different portfolios.

Repositories

At the level of each repository, the admin can grant access permissions to both individuals or teams. The way I use this in an organization is that I grant access based on teams for organization people, and reserve invitations to individuals for external collaborators, that are not part of the org. This clarifies the permission list in the repo settings.

Notifications

Email notifications are set up for each user, and via the Watch button near the top right of the repo page

Under this button menu is where one can set the granularity of notifications for this repo. Each user can set overarching settings under their account settings:

I don’t know what are the default settings there, but I presume by default one does not receive notifications for repos they are not part of (as a team or individual).

So yes, it may be that members will have to click on the Watch button.

Thanks Jonathan, would you know how to configure Jeckyll to have webpages with a side column that would list the content of the site? (This would imitate the current ACG website, or something like GitBook does…)

That should just be a matter of choosing a theme that does that, and configuring it so it has a consistent “menu” and displays it appropriately (separation of form and content), c.f. http://jekyllthemes.org/

(note to self more than anything) it should also be possible to choose/enable latex math formulæ using mathjax for display, if it isn’t enabled already.

Another problem is that even if @marmetl2 is only a member of cg-team, the previous message was sent as an email-notification to this non-member of acg-team (even if @marmetl2 is not mentioned in the message).

It seems that @marmetl2 was automatically made a member of the a-cosmology-group organization when he was added to the cg-team (unless I made a mistake), and so automatically receives notifications for the discussions on this acg repo. I can’t find settings to stop email-notfications for @marmetl2 when discussions happen outside cg-team.

It seems that @marmetl2 could stop those notifications by changing the settings in their profile, but this is already way beyond what most cg members will be willing to do…

I would disable organization discussions, which allow members to broadcast message to the entire organization. I would also disable teams discussions in the organization, to keep it simple. Discussions in each repo only, and @mention people by name. Both can be disabled in the organization settings:

What happens to marmetl2 is probably that he was initially a member of acg team, so the Watch was turned on for that repo. Each user decides what repos to watch, this has nothing to do with membership or teams. But it might have been turned on by default when joining the acg team initially. More testing required! 😄

I recall a mail from github a few months ago that announced they were canning Teams discussions. Perhaps that is why the tab does not show anymore. Under our EGSnrc team, I see that the tab is gone, and that our previous exchanges are now available in JSON as an archive.

I guess github decided, as I did, that teams discussions were not very useful and confusing. I never used them, aside from test message to see if team members received them 😆

Yep, here is the sunset notice from February:

https://github.blog/changelog/2023-02-08-sunset-notice-team-discussions/

And finally, yes, joining an org team makes you a member of the organization! So everyone in the cg-team or acg-team teams will be a member of the organization a-cosmology-group.

Thank you 🙏🏻 for renaming the teams with a -team suffix, I was going to suggest that this morning, lest our discussion becomes unintelligible!

Interestingly, https://github.com/watching shows the watched repositories for the current logged in github user.

Agreed. I was starting to feel that this was getting out of hand fast, in the direction of square peg, round hole and all that. 🤓

Using GitHub discussions as a forum writ large was pushing it. It is more appropriate for discussions about the project itself, as in the files that are stored in the repo.

So yes, use the repo to edit and manage the website content (that makes sense), and just remove the cosmology discussion section to keep discussions about the website files, about the repo. If all cosmology discussions happen via Gaggle, then that is simplest.

I think using GitHub and markdown for the website is nevertheless big progress, enabling true collaborative editing.

At this point you could do away with the organization, and just have an a-cosmology-group repository. Unless you foresee wanting to host under content under a common org.

On second thought, best.to keep the org, otherwise the repo is under an individual account…

More tasks (for @marmetl ):

• delete both teams,
• delete the cg-repo,

Planned structure:

• A Cosmology Group will have only one organization on GitHub, a-cosmology-group, with:

  • one a-cosmology-group/acg-repo for the website files (e.g. members will want to upload their papers to the website),
  • Discussions about the website and cosmology, (this is a bit different from your understanding @ftessier, see below)
  • Wiki for whatever reference text that is not on the website.

• The 46 members on the discussion platform forum@cosmology.info (which is an alias for a.cosmology@gaggle.email) will join the 27 members on the low-traffic cosmology@gaggle.email platform (moderated). (The cosmology@gaggle.email forum is only for announcements to all members via email, there are virtually no discussions happening there.)

• Members who wish to discuss cosmology (high-traffic with multiple replies, questions, etc…) will have to sign up for a GitHub account and post here. An advantage of using GitHub Discussions is that equations can be nicely rendered for everyone to see (which is not the case for email).

I may be out of line, here, but it seemed to me that the original reason for a new platform was the incapacity of Gaggle to render equations adequately. Keeping Gaggle for discussions appears to me a non sequitur…

Hi Pierre $\approx -{\rm i} \ln(-1) \times 10^5$ ! Not out of line, but you may not have seen my message to @ftessier posted 10 minutes before yours:

Members who wish to discuss cosmology will have to sign up for a GitHub account and post here. An advantage of using GitHub Discussions is that equations can be nicely rendered for everyone to see.

The Gaggle.email platform is mostly used for one-way-notifications, for example:

A new paper that may be of interest Measurement of parity-odd modes in the large-scale 4-point correlation function of SDSS

or

The May ACG Newsletter has been posted.

A relevant comment might be posted on cosmology@gaggle.email, but any discussion will happen on GitHub’s a-cosmology-group/acg-Discussions.

Hi Pierre ≈−iln⁡(−1)×105 ! Not out of line, but you may not have seen my message to @ftessier posted 10 minutes before yours:

Right! My apologies. There are large components of this architecture that are still obscure to me.

So, what you want then is a mailing list to avoid subscribers the need for a GitHub id, right? That still leaves me with the impression that two platforms is one too many, but anyways…

$F=GMm(\frac{1}{r^2}+\frac{6GM}{c^{2}r^3})$

The equations rendering in discussions will certainly be a substantial improvement.

A Cosmology Group has two mailing lists on Gaggle.email

• Current members of cosmology@gaggle.email will remain in the same mailing list.
• Current members of a.cosmology@gaggle.email will be moved to the cosmology@gaggle.email mailing list.

Any member who wishes to create GitHub Discussions with $equations$, or edit the repo (new website) will need to create a GitHub id.

There are large components of this architecture that are still obscure to me.

That’s one of the reasons why two platforms are necessary: all members of A Cosmology Group will remain on the Gaggle.email mailing list since most of them are only watching the discussions and won’t be bother to learn to navigate the GitHub system just to read a few messages.

The most important reason for two platforms is that the professionals currently on cosmology@gaggel.email (physicists, philosophers, etc.) do not want to see their mailbox cluttered with public discussions (e.g. Han and Bud discussing “dimensional analysis”, something any scientist learns as a teenager; or e.g. Olli’s indecipherable philosophy)

@Pierre314159 I will invite you to join a-cosmology-group on GitHub. You will need 2FA enabled to join the Organization.

Once you are a member it will be easier to use GitHub. You will see the `a-cosmology-group’ organization listed in your personal organizations (click the OctoCat at the top-left of your screen)

Hi Louis,

I’m in I guess. It seems to me that GitHub’s primary usage is as a collaborative tool for coders, which raises the question, will any original post (as opposed to a comment) be subject to collaborative revision (as opposed to commentary)?

Regards, Bud

Hi Bud, if by “original post” you mean discussions, it’s not possible for members to modify someone else’s post. You can only add comments to the discussions and edit/delete your own.

If by “original post” you mean files in the repo (website, code, papers, articles, etc.), these files can be added/edited/deleted by everyone. They are both subject to collaborative revisions and comments. See HOWTO.md for how to edit files.

By the way, when editing on github becomes cumbersome, eventually, you can clone the repo on your computer and work on your local copy, pushing changes up to the origin repository on github. But that’s another chapter… 😉

@budrap00, git is a versioning system to end the chaos! Anyone can suggest changes through pull requests, but they must be approved before moving on. If the review honour system is not sufficient, it is possible to create code owners, required multiple reviews, etc. as gatekeeping mechanisms. The honour system means, for example, to add the original author as a reviewer, for example.

The world’s software is built on git and github, with 100 million + repositories. It is very apt at ordering the chaos 😄

In Git version control, he “add, edit or delete” operations are done in a separate branch until the changes are accepted by a GitHub maintainer. Then and only then are changes “merged” with the main branch of the repo (i.e. accepted as part of the system).

This prevents destructive mistakes (e.g. deletion), mischief (e.g. modify files) or chaos (e.g. moving files and breaking the website). Even if the maintainers make mistakes, the Git version control always remembers previous versions and the repo can be returned to an earlier version.

(And @ftessier just answered as I was typing… ;-)

Well said. In fact, it is quite difficult in git to remove anything from a repository once it gets going. Try! 😉

Nice work Louis!

Beautiful! I tuned out a bit while teaching our course this week. I will continue to follow development, and help out when possible.

Success ! MathJax runs on the website (see the $\sigma_8$ tension, second line…)

This is a bit complicated: the .md file runs through the .html converter Jekyll, and then when the .html file is displayed it runs through the MathJax converter. As a result,

inline equation are written $$E=mc^2$$ with two $$-signs

in the .md file.

Looks good indeed!

I used https://www.drlinkcheck.com/ to scan for broken links, maybe deserving attention:

404 Not found https://a-cosmology-group.github.io/acg/essays/traunmueller/embarrassing-blunders-in-big-bang.pdf Linked from: https://a-cosmology-group.github.io/acg/essays/

404 Not found https://a-cosmology-group.github.io/acg/essays/.sorli_as/big-bang-cosmology-funeral_sorli.pdf Linked from: https://a-cosmology-group.github.io/acg/essays/

404 Not found https://a-cosmology-group.github.io/acg/essays/.lerner_ej/concordance-cosmology-challenged.pdf Linked from: https://a-cosmology-group.github.io/acg/essays/

404 Not found https://a-cosmology-group.github.io/acg/essays/.lerner_ej/new-results-challenge-big-bang-at-ccc-1.pdf Linked from: https://a-cosmology-group.github.io/acg/essays/

404 Not found https://a-cosmology-group.github.io/acg/essays/.lerner_ej/alternatives-to-big-bang-at-ccc-1.pdf Linked from: https://a-cosmology-group.github.io/acg/essays/

404 Not found https://a-cosmology-group.github.io/acg/conferences/ccc2/assets/ccc2.gif Linked from: https://a-cosmology-group.github.io/acg/conferences/ccc2/

404 Not found https://cosmologyscience.com/ref.htm Linked from: https://a-cosmology-group.github.io/acg/resources/

404 Not found https://a-cosmology-group.github.io/acg/essays/traunmueller/assets/e404.jpg Linked from: https://a-cosmology-group.github.io/acg/essays/traunmueller/embarrassing-blunders-in-big-bang.pdf

404 Not found https://a-cosmology-group.github.io/acg/essays/.sorli_as/assets/e404.jpg Linked from: https://a-cosmology-group.github.io/acg/essays/.sorli_as/big-bang-cosmology-funeral_sorli.pdf

404 Not found https://a-cosmology-group.github.io/acg/essays/.lerner_ej/assets/e404.jpg Linked from: https://a-cosmology-group.github.io/acg/essays/.lerner_ej/concordance-cosmology-challenged.pdf and 2 more

404 Not found http://www.spaceref.com/news/viewpr.html?pid=16668 Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

404 Not found http://nautil.us/issue/15/turbulence/do-we-have-the-big-bang-theory-all-wrong Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

404 Not found http://www.spaceref.com/news/viewpr.html?pid=16665 Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

404 Not found http://www.spaceref.com/news/viewpr.html?pid=14524 Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

404 Not found http://www.spaceref.com/news/viewpr.html?pid=17420 Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

Host not found http://bourabai.kz/universum.htm Linked from: https://a-cosmology-group.github.io/acg/redshift/

Host not found https://www.noao.edu/outreach/press/pr05/pr0508.html Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

Timeout https://www2.ling.su.se/staff/hartmut/Well-Founded_Cosmology.pdf Linked from: https://a-cosmology-group.github.io/acg/resources/marmet_l/reviewed-publications.html

Interesting statistics out of link checkking:

Thanks! Xenu’s Link Sleuth also found many broken links.

I fixed some ‘internal’ links that were typos in my translation to .md.

As for the external links that have disappeared, one has to google search the webpage to see where it has been moved… a lot of work!

I know that https://cosmologyscience.com/ref.htm is David Dilworth’s website, and it’s off-line temporarily, so we’ll leave the link as is.

Oh, my list is exhaustive, as far as I know, so there is only 15 404s.

Not sure what you mean. The list is exhaustive, but does not tell us if the Error 404 is caused by a webpage that has disappeared, or if it simply has been moved!

For example, the 12th Error 404 link http://nautil.us/issue/15/turbulence/do-we-have-the-big-bang-theory-all-wrong produces an Error 404.

However, this page may have been moved since this link worked in the past. With a bit of searching I found that the page has indeed been moved to https://nautil.us/do-we-have-the-big-bang-theory-all-wrong-234989/.

I’ll fix that one tomorrow…

I get it. I thought you meant “a lot of work”, as in there might be many more dead links. I will quickly check if I can recover some of them.

@ftessier 80% of the broken links are now fixed!

Indeed @doctorlard, I looked into this a few weeks ago to be sure it could be done (before going through the trouble of building the site on GitHub).

It’s just easier to forward the browsers to the GitHub webpage for now, but the final step will be to serve the GitHub page from cosmology.info, and I have access to the CNAME A-record in the c-panel configuration (on Hostmonster.com). The DNS records for cosmology.info can be set to a-cosmology-group.github.io/acg/, and the browser will show cosmology.info in the address bar.

The way it’s setup is that the website a-cosmology-group.github.io/acg/ is associated with the acg repo through an IP address generated by GitHub that I copy to the A-record.

Hostmonster.com is currently hosting the ACG website/email for ~$110/year. Once the DNS record is changed we won’t need them to host the website. I just hope they will accept to host the two URLs cosmology.info (at $30/yr) and cosmologystatement.org (at $35/yr) without charging to host the website on their servers…

You are right, but A-records also deal with IP addresses. GitHub generates an IP address for a-cosmology-group.github.io/acg/ which is copied to A-record. See here.

Yes, I understand how A and CNAME records work. What I’m saying is that you will need something else to handle the /acg URL path, because when someone arrives at the GitHub-hosted site from a cosmology.info redirect, the DNS will only get you as far as a-cosmology-group.github.io which is currently a 404, whereas the site is at a-cosmology-group.github.io/acg

I believe github handles that, because you can publish a website from an organization repository.

GitHub Pages is supposed to handle that using the Pages configuration of the acg repo.
We’ll know soon if it works…

To do:

• forward cosmology.info to https://a-cosmology-group.github.io/acg,
• install BIND on my computer for diagnostics,
• configure GitHub custom subdomain, configure CNAME on HostMonster, etc.,
• configure GitHub enforce HTTPS encryption.

It took a while for the changes to propagate through the servers, but it seems to work now! 😅

This can be tested by clicking the first link at the top of this discussion, and end up with https://cosmology.info/ in the URL bar of the browser. (www.cosmology.info also works.)