Documentation points to mathics.net which is hopelessly out-of-date

[TobiSchluter]

Per @wolfv 's request, I'm spining this off #754

The most visible site running mathics is http://www.mathics.net. It is hopelessly out-of-date.

The canonical site seems to be https://mathics.angusgriffith.com/. This site has broken certificates and a hard-to-remember domain name. Moreover, if you open the documentation on it, you are pointed to the "bad" site, see the screenshot below.

I think there are a number of issues which I vaguely sort in terms of ease-of-addressing (note that I write 'should', but please don't take that as more than a way of expressing potential aciton items, I'm definitely not trying to tell anybody what they should do):

• the documentation should not point to mathics.net
• certificates of angusgrifftih.com should be fixed
• a beginner-friendly domain name should be found (actually mathics.org appears available)
• somebody should volunteer to host mathics using that domain (someone very nice :) )
• mathics.net should be reclaimed for the project

It would be important to make sure that there is an easy way for potential users to not run into an old version of mathics when they first try it, because simple bugs (such as my #754 ) are sure to turn potential users away quickly from this very good piece of software.

[GarkGarcia]

Yeah, this is a real problem indeed. @wolfv and @poeschko are members of the Mathics organization, I think they have the authority to invite us to becomes members too. I think creating another organization isn't the most ideal path, but I understand why you guys felt the need for doing so. I'd like to keep our efforts focused in this repository, but this project can't go forward if the people maintaining it don't have the authority necessary to release the code. As such, I'll accept the invitation to be a part of this new organization, but I'd prefer if the organizations eventually got merged.

@GarkGarcia, the idea is -at least by now- to have a common space to work together in PRs without breaking the main project. If once the project is ready for release, we do not have an answer from the owners of this project, we could analyze to publish the release from there.

And as for day-to-day work - I would prefer making PR's reviews and commits in Mathics3, since we now all have equal access.

It is possible to issue PRs from there back into Mathics. Whether we do that or how often it is done or whether the project owners even want that, it is undecided and up for discussion. (Silence will be taken to mean go with the other organization.)

Fair enough. It would be cool @wolfv or @poeschko could make us part of the Mathics organization as soon as possible, but if they can't I'll start pushing my commits to the Mathics3 fork.

@GarkGarcia, the idea is -at least by now- to have a common space to work together in PRs without breaking the main project. If once the project is ready for release, we do not have an answer from the owners of this project, we could analyze to publish the release from there.

Makes sense too. I don't see how we could resolve our governance issues in here without the help of someone who is part of the Mathics organization (so we don't have another option, at least until someone invites us to be a part of the organization).

[GarkGarcia]

This issue has been mostly addressed by https://github.com/mathics/Mathics/pull/921. @mmatera @rocky I guess we can close the issue?

We really shouldn't discuss our governance issues in here. Let's continue the discussion in https://github.com/mathics/Mathics/issues/841.

[rocky]

What would you like to get working before tackling the Django issue?

For myself, it is just a matter of seeing that more or less everything else that is easily covered is covered and that the main thing blocking us is having this resolved.

I may only have for a while short amounts of time to work on this. So for me and right now I have this backlog of little stuff that had been accumulating that I'd like to sort out first.

[rocky]

This issue has been mostly addressed by #921. @mmatera @rocky I guess we can close the issue?

This would have been fine if we any of us had access to mathics.github.io. The the final bit about going to http://mathics.github.io/ and seeing somethat is current and works would be solved quickly. Right now that that is not the case.

We can make this work using the new organization in short order though.

As for the process about closing the issue (even though in my opinion right now this is still broken) is up to you.

[mmatera]

I'm sorry for not coming today. I'd be happy to arrange another meeting if you guys are available some other time this weekend. I'm available today.

' I can't this weekend, but we can plan on another one next weekend. Suggest a time.

What about next Saturday at 5 pm in the UTC-3 timezone?

OK. Let's schedule then a new meeting on https://meet.jit.si/Mathics1.1milestone for October 3rd, 17.00hs (GTM-3)

[rocky]

I think I can make it but if I can't I'll let you know.

On Sat, Sep 26, 2020 at 3:07 PM Juan Mauricio Matera < notifications@github.com> wrote:

I'm sorry for not coming today. I'd be happy to arrange another meeting if you guys are available some other time this weekend. I'm available today.

' I can't this weekend, but we can plan on another one next weekend. Suggest a time.

What about next Saturday at 5 pm in the UTC-3 timezone?

OK. Let's schedule then a new meeting on https://meet.jit.si/Mathics1.1milestone for October 3th, 17.00hs (GTM-3)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/mathics/Mathics/issues/762#issuecomment-699535936, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAACFEZRVBGOMCTPDLR3QXLSHY3YZANCNFSM4F3MBMRQ .

[poeschko]

@GarkGarcia @rocky I added you to a new team @mathics/maintainers which has Maintainer access to the repos Mathics and mathics.github.io. Let me know if you need anything beyond that.

Regarding domains, I'll point mathics.org to https://mathics.github.io – any suggestions what should happen with mathics.net?

[mmatera]

In any case, maybe it would be a good idea to work on the "fork" organization, in a way to do not break the main repository. Then we can merge everything here.

[slel]

Suggestion: point mathics.net to https://mathics.gighub.io/try and put some words there explaining ways to try mathics online. If we find a way to revive something like mathics.net we can either make it live at that "try" page or at least link from there.

[rocky]

Many thanks!

I will be testing this out. The first order of business will be to try to add a PDF for the 1.1 dev branch.

On Sat, Sep 26, 2020 at 3:14 PM Jan Pöschko notifications@github.com wrote:

@GarkGarcia https://github.com/GarkGarcia @rocky https://github.com/rocky I added you to a new team @mathics/maintainers https://github.com/orgs/mathics/teams/maintainers which has Maintainer access to the repos Mathics and mathics.github.io. Let me know if you need anything beyond that.

Regarding domains, I'll point mathics.org to https://mathics.github.io – any suggestions what should happen with mathics.net?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/mathics/Mathics/issues/762#issuecomment-699536638, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAACFE2C7S4DKFXIFY6RLFTSHY4Q3ANCNFSM4F3MBMRQ .

[rocky]

Suggestion: point mathics.net to https://mathics.gighub.io/try and put some words there explaining ways to try mathics online. If we find a way to revive something like mathics.net we can either make it live at that "try" page or at least link from there.

@slel I am not exactly sure what you mean here. Would you fork the code and put in a PR for this. If I have access, I'll merge that. Thanks.

[rocky]

@poeschko Many thanks for doing this. Would you also consider adding @mmatera ?

I have just updated https://mathics.github.io so that it no longer has broken links.

[poeschko]

Yes, of course. I added @mmatera to @mathics/maintainers.

I'm happy to see some increased interest in Mathics again!

It would be nice if mathics.net was a place where you could directly use it, not just a description of how to do it (that would go on mathics.org). If you manage to set that up somewhere, let me know and I'll change its DNS.

[rocky]

Absolutely! Will do.

It would be nice if mathics.net was a place where you could directly use it, not just a description of how to do it (that would go on mathics.org). If you manage to set that up somewhere, let me know and I'll change its DNS.

Right now it feels for me like we are fighting too many small fires. #916 is a plan for doing this. Hopefully someone will volunteer in the short term to make it happen.

[slel]

It would be nice if mathics.net was a place where you could directly use it, not just a description of how to do it (that would go on mathics.org). If you manage to set that up somewhere, let me know and I'll change its DNS.

Can you describe how you had it set up at mathics.net when it was active? With a description I can ask my department or university about hosting this.

[GarkGarcia]

This issue has been mostly addressed by #921. @mmatera @rocky I guess we can close the issue?

This would have been fine if we any of us had access to mathics.github.io. The the final bit about going to http://mathics.github.io/ and seeing somethat is current and works would be solved quickly. Right now that that is not the case.

We can make this work using the new organization in short order though.

As for the process about closing the issue (even though in my opinion right now this is still broken) is up to you.

Fair enough. I'll update the information in mathics.github.io and then I'll close the issue. Does that work to you?

[GarkGarcia]

Regarding domains, I'll point mathics.org to https://mathics.github.io

Actually, I think using the mathics.org domain for mathics.github.io would be nicer, don't you guys agree? GithubPages allows users to use custom domains.

any suggestions what should happen with mathics.net?

For now I think our best option is to point it to mathics.org or mathics.github.io. We can set it up again if we ver menage to get it running as before.

[GarkGarcia]

I have just updated https://mathics.github.io so that it no longer has broken links.

Great! Thanks! I guess we can close this issue then?

[GarkGarcia]

Can you describe how you had it set up at mathics.net when it was active? With a description I can ask my department or university about hosting this.

That would be amazing!

[rocky]

I guess we can close this issue then?

This issue is kind of a tough one with many levels. The top link seems to refer to the online-help. As far as I can tell, that relies on the Javascript $doc variable getting set and I have no idea what that should be. I downloaded the 1.0 tarball from PyPI and I don't see docs in that.

When I currenlty open mathicsserver and go to docs I see blank info, not obsolete info.

So in a sense, yes, that is fixed. Maybe a "better" fix would be to disable the doc ability altogether. I don't know.

Again this is the data, I leave to you what you want to do with it or whether you want to close this issue.

[GarkGarcia]

Also, @rocky @mmatera I guess the other organization is kinda irrelevant now? Is a reason to keep it?

[poeschko]

I also updated the repo details of https://github.com/mathics/Mathics so it points to https://mathics.org (again), since that's equivalent to (and the canonical place for) https://mathics.github.io now.

Can you describe how you had it set up at mathics.net when it was active?

That was a while ago and that server is long gone, unfortunately... I think it was a fairly standard Django webserver. The biggest complication was getting Sage to work on the server, but that's no longer a required dependency for Mathics, anyway (right?). It did support logging in and saving "notebooks" (using a MySQL database). As far as I remember, there was a way to turn off any file-related operations (or other potentially unsafe operations) in the Mathics interpreter, and that was basically the security mechanism.

There was definitely a stability issue, so every now and then (perhaps if somebody evaluated something "special"), the server became unresponsive. Maybe this is not an issue anymore with all the technologies involved (most importantly the Mathics code base) having evolved in the meanwhile, but it's certainly something to watch out for.

Good luck! :-)

[rocky]

I guess the other organization is kinda irrelevant now? Is a reason to keep it?

Yes. Although I have removed whatever exists here, it appears that I can't create new repositories under Mathics and I assume @mmatera can't either. I have a new repository for splitting out a terminal CLI. mmatera added one from his namespace as well.

Also in the future recall we want separate Python modules for LoadModule imports. Specificaly one for natural language stuff, one for graphs, and so on. This would naturally go right under Mathics.

@poeschko if you could make that happen to I would be eternally grateful and will remove the github Mathics3 Organiation as soon as we copy the remaining repositories to Mathics. Many thanks in advance.

[GarkGarcia]

I guess we can close this issue then?

This issue is kind of a tough one with many levels. The top link seems to refer to the online-help. As far as I can tell, that relies on the Javascript $doc variable getting set and I have no idea what that should be. I downloaded the 1.0 tarball from PyPI and I don't see docs in that.

When I currenlty open mathicsserver and go to docs I see blank info, not obsolete info.

So in a sense, yes, that is fixed. Maybe a "better" fix would be to disable the doc ability altogether. I don't know.

Again this is the data, I leave to you what you want to do with it or whether you want to close this issue.

Humm, I see. Indeed there are still references to mathics.net in the code (something related to authentication) and I don't understand how the "online help" stuff you mentioned works. I'll keep the issue opened while we figure this things out.

[rocky]

As far as I remember, there was a way to turn off any file-related operations (or other potentially unsafe operations) in the Mathics interpreter, and that was basically the security mechanism.

Alas I recently made this worse by adding ProcessId, ParentProcessId, MachineId and so on. But assuming you get everything else worked out, I'd be happy to figure out a hack for removing these functions.

Probably there might be an attribute added that indicates these are not suitable for online live servers.

[GarkGarcia]

As far as I remember, there was a way to turn off any file-related operations (or other potentially unsafe operations) in the Mathics interpreter, and that was basically the security mechanism.

Alas I recently made this worse by adding ProcessId, ParentProcessId, MachineId and so on. But assuming you get everything else worked out, I'd be happy to figure out a hack for removing these functions.

Probalby there might be an attribute added that indicates these are not suitable for online live servers.

That's a great idea! We should probably add a switch to disable IO functionalities altogether (reading an writing to disk, but only the "user-controled" stuff, like saving images to disk).

[poeschko]

Shall we move the discussion about new repositories to https://github.com/orgs/mathics/teams/maintainers/discussions? Let me know there what repositories you'd like to have created.

[GarkGarcia]

Shall we move the discussion about new repositories to https://github.com/orgs/mathics/teams/maintainers/discussions? Let me know there what repositories you'd like to have created.

Yes! I try to make a summery of our current discussion in there.

[GarkGarcia]

Shall we move the discussion about new repositories to https://github.com/orgs/mathics/teams/maintainers/discussions? Let me know there what repositories you'd like to have created.

Yes! I try to make a summery of our current discussion in there.

See https://github.com/orgs/mathics/teams/maintainers/discussions/3 for the summery.

[rocky]

I just spent some time with git bisect to figure out which commit broke the docs. It appears to be:

ee1677d1a528ce1633109b59674d08fb1c62eede is the first bad commit
commit ee1677d1a528ce1633109b59674d08fb1c62eede
Author: Angus Griffith <16sn6uv@gmail.com>
Date:   Sun Mar 12 22:32:53 2017 -0700

    basic Django 1.11 support

:040000 040000 79dd6a61dd4e9a948c97a5eef2128255b4f52d9e 1c3028f3560da350c3e272d8fd37c1f8f6cab05f M  mathics
:100644 100644 048e619d9d4763e36681186f69af9efad84a4c3a f6c0463c7783069648fea1ecb5a825d538994204 M  setup.py

[rocky]

I believe this is now fixed in bc51171fc9b45d56f9aed217e85293b5b42cc4b4

Of course there's always more to do. but my take is this should be follow-on isssues.

For example going over the content to ensure it has everything. (Im too tired to do right now).

Right now, I am not seeing docs on newly-added functions.

Where exactly is this information coming from? The information comes from mathics/doc/tex/data and docs are updated when that file is updated.

[GarkGarcia]

I believe this is now fixed in bc51171

That's great! Awesome work!