Sat 19 Jan 2019 12:28PM

Stable, well-supported platforms we can for this group instead of corporate datafarms

DS Danyl Strype Public Seen by 128

*TL;DR *: this thread is a place to share our experiences with community-hosting organizations, both good and ... not-so-good. The primary goal is to identify reliable community hosts for storing the inputs and outputs crucial to our work in this group, as a replacement for using corporate datafarms. The secondary goals are to help individual members and our projects identify reliable community hosts for their needs, and to help community-hosting organizations to be better understand how to serve their users well.

Body: I've been refusing more and more to use corporate datafarms since I quit FarceBook into 2010. I don't use Skype and other tools that I'm not already locked into, in ways that require research on data export and replacements (still working on breaking my last ties with Google, GH, and others). I will never use Slack, Discord, WhatApp or the various other datafarms that have become popular since I started this journey.

This is literally the only reason I'm putting effort into this group. Because I understand it as being about identifying, and where necessary creating, ethical replacements for the datafarms, and making them more accessible, first to ourselves (developers and geeks), and then to the general public.

In the Housekeeping thread, Greg C said:
> I don’t think it’s very important for me to abandon ‘using the tools of the enemy’ compared to my many urgent priorities.

I don't quote this to pick on Greg, but rather to ask that we all acknowledge something important. Having the option to not worry about the practical implications of using corporate datafarms is a privilege, one that not everyone shares. For example, I cannot read GoogleDocs without a working VPN, nor can anybody else in China, and a number of other countries. I want this group to be inclusive of people who may live in less ... liberal regimes than others, and for whom having their activities on the net tracked by corporations, and sometimes made available to governments, may be more compromising than it is for others. For these reasons, and many others, I'm not comfortable with obliging members of this group to expose ourselves to corporate datafarms to fully participate in the discussions here.

For sharing documents in this group, I suggest avoiding third-party sites unless they're really needed.
* If you want to explain something in too much detail for a normal comment, type it in a text editor or word processing program, save it in a free format (eg .txt, .md, or .odt), and attach it as a TL;DR comment.
* If you want to share something that isn't text (eg an image or diagram), if you can, attach that to a comment too.
* If you want to create a text collaboratively use the context box at the top of each Loomio thread.

But yes, sometimes we will want to collaborate on documents in ways that editing a Loomio context doesn't allow, and we do need to identify good tools to use for this.

In that same thread, Greg raised some valid concerns about the reliability and longevity of self-hosted or community-hosted platforms. I'm sure we've all been stung by having a site disappear on us, I certainly have. But when this happened to me, it was mainly because I didn't the have reliable information I needed about software, and hosting organizations, to make good choices. The better I've got at collecting this information, the safer my data has been.

Let's share the huge amount of collective knowledge held on this by the members of this group, so we can all make better hosting choices. Both for this group itself, and in our own work.

If we need to collaborate on texts as a group, my suggestion is to create an Open App Ecosystem project on a GitLab instance run by a cooperative or non-profit that we can trust to be competent sysadmins, to stick around for the long haul, and to give us plenty of warning if anything changes that might require us to change hosts. On GitLab, almost everything is a Git repository, including issues trackers and wikis, so anyone who wants to can maintain a synchronized backup of all our data using Git tools. Those who want to, could even contribute to collaborative documents using GIT and a text editor in their terminal, while the rest of us, who don't feel confident using those tools, can just use the web interface as we would on GH.

I would suggest we use either:
* use git.nzoss.org.nz, run by a non-profit called the NZ Open Source Society, "We are happy to provide this service for any Free and Open Source Software licensed projects with a New Zealand connection. You may create up to 10 projects". Enspiral/ Loomio are a NZ connection, and I'm confident the NZOSS would thoroughly support this project. GH accounts can be used to login here.
* ask to use git.feneas.org, run by a non-profit called the Federated Networks Association. We're not strictly a federated network project, but I suspect that federated networks will play a big part in anything that deserves the name Open App Ecosystem. This instance allows login using accounts from GH, GitLab.com, BitBucket, so it wouldn't necessarily require folks to set up a new account to use.
* ask to use Git.coop, run by a cooperative in the UK called Web Architects. Used by social.coop and various other projects folks here are involved with.

I have been building up a list of community-hosted GitLab instances on the P2PF wiki:


Danyl Strype Sat 19 Jan 2019 1:36PM

Loomio itself, obviously, is not a corporate datafarm, but we trust it's here for the long haul. I chose RiseUp to host my email, and CoActivate to host the Disintermedia blog and wiki. More than a decade later, they are still going, and providing everything I need (if not always everything I want ;)). With more help from folks with a working knowledge of devOps, Python, Zope and its web frameworks like Plone and Django, we could probably improve both the UX and performance of CoActivate. But my data has always been safe there, and any serious performance issues have been dealt with promptly (edit: CoA went dark during the COVID-19 pandemic and it's hard to say if it will return).


Simon Grant Sat 19 Jan 2019 2:42PM

Thanks for starting this thread, @strypey . As you say, it's a tricky discussion sometimes. I applaud and admire your dedication to not using corporate datafarms. This is one of the places that I sense most strongly that coordination and collaboration between us is vital, so that we can focus our efforts on a small selection (that can change) of services that align with our values. Otherwise we simply can't compete. I'm ready to jump when the services reach a good workable standard of reliability and usability, but may need to be told when. Then there's the challenge of moving the next round of people who are one stage more reluctant (or less informed) that we are...


Chris Croome (Webarchitects Co-operative) Sat 19 Jan 2019 6:15PM

Loomio itself, obviously, is not a corporate datafarm

Agreed, but last time I looked it was hosted on one... it is now behind Cloudflare so I don't know if it is still hosted on AWS.

But it is Free software so a copy of the code could be hosted anywhere.


Greg Cassel Mon 21 Jan 2019 10:19PM

This is an important subject; thanks @strypey for seriously approaching it.

Before I go further, please note Strypey that I found it unnecessary and distracting for you to create context by directly quoting me outside of the other thread's context, and indirectly but clearly identifying me with a broad problem. I.e. you focused on the idea that using Google is a privilege. That's true in some important ways; however, privilege isn't a simple yes/no binary. It's a deeply complex multidimensional variable, and difficult if not impossible to measure comprehensively within any specific shared activity. With that in mind, I share some related statements I made in the thread which you quoted me from:

Most cultural & tech revolutionaries I know, myself included, live precariously on limited resources. This includes many people & teams who provide open source alternatives to multinationals like Google; and I think it's important to be mindful to not multiply our points of potential failure.
For example, I use GH and GDocs extensively, because I literally can't afford to depend on alternatives which might vaporize either suddenly or gradually. Nor can I even really afford to do thorough research to ensure that specific alternatives are practically as stable (both technically, and in terms of attentive human support) as GH and GDocs.

I referred above to some very real risks, which are related to specific (but non-binary; dimensional) privileges which I lack. Privilege occurs in personal resources, such as knowledge/familiarity with specific tools, and also of course in economic resources-- such as users' ability to financially buffer themselves against potential tech failures.

With that said, I hope we can move forward focused on subjects other than yes/no perceptions of privilege, and I hope I can attend sufficiently below (in a new comment) to the important ideas you've raised.


Greg Cassel Mon 21 Jan 2019 10:29PM

For example, I cannot read GoogleDocs without a working VPN, nor can anybody else in China, and a number of other countries. I want this group to be inclusive of people who may live in less ... liberal regimes than others, and for whom having their activities on the net tracked by corporations, and sometimes made available to governments, may be more compromising than it is for others.

Important issues @strypey . However, I wonder:

  • How big of a hardship it is to often or always use a VPN? I use one intermittently, for specific purposes related to unfair copyright laws.
  • Wouldn't using a VPN all of the time be wise within any especially repressive regime, regardless of whether or not a person's signals went through any centralized corporate servers (like Google's)?

As far as I can tell, my listed uncertainties mitigate (but do not eliminate) your stated concern. I guess I would always use a VPN in China, etcetera.

I'm not comfortable with obliging members of this group to expose ourselves to corporate datafarms to fully participate in the discussions here.

Well, people also need to know English to fully participate here, and learning English is surely a bigger 'ask' than our other tech dependencies. If we wanted this Loomio group to be as globally inclusive as possible, I expect we'd not only allow but actively encourage participation in languages other than English, and relying upon auto-translate technologies to reduce our discussions to some reasonably simple and mutually clear level of complexity.

^ I'm definitely not trying to be snarky with that, although it might seem that way! My point is that privilege happens in every social context; it can't be avoided; thus it comes down IMO to a mindful cost/benefit analysis of each choice which we willfully support. All choices either intentionally or unintentionally create privilege.

On a practical level: I expect that participants here may link media in various formats, including GDocs, which require others to opt in or out to a number of technical and social dependencies. (Using loomio.org itself includes some such dependencies, which could be criticized.) I recognize your discomfort with GDocs links, Strypey, but I think I'd have similar discomfort with some other techs which I know little or nothing about.

For now, I suggest that everyone create & link media in whatever format they prefer. (Hopefully most if not all of it will be persistently readable for free, without creating new accounts with any services.) To move forward inclusively, however, I hope Strypey to support your valuable effort here:

… to identify reliable community hosts for storing the inputs and outputs crucial to our work in this group, as a replacement for using corporate datafarms.

I'd definitely prefer to avoid funneling any activity through Google, or any for-profit corporation, or even a nonprofit centralized system-- for many reasons.

In discussing specific possibilities for data storage & signaling, I hope we can inclusively unpack assumptions, and technical and social dependencies, wherever they are.

For example, Strypey, you would suggest that we as a (small) community invest in depending upon Gitlab software hosted by some other group. Well, I get that Gitlab isn't Github and isn't owned by the 'evil empire' of Microsoft, but I don't find that particularly persuasive in itself. As far as I can tell, Gitlab is a not-nonprofit corporation. Yes their open source software can be independently used, and that's a potential plus. On the other hand, you suggest that we use that feature by creating another dependency, by relying upon NZ Open Source Society, Federated Networks Association or Web Architects.

Please understand I have no specific reasons (so far) to doubt the integrity and sustainability of anything which Gitlab, NZOSS, FNA or WA does! However, I also know next to nothing about these potential dependencies.

I'm generally unlikely, right now, to default toward directly depending on Gitlab plus some hosting organization instead of Google. Shared values and intentions only go so far for me. However, I feel receptive to any concise and easily-processed info which could develop my confidence in committing to depend upon Gitlab + a specific app hosting organization. I don't know yet if such concise and easily-processed info exists.

BTW none of this addresses the fact that I, for example, use GDocs for some purposes and GitHub (another proprietary empire!) for other purposes, because the toolsets are significantly different. GDocs and distributed versioning have different positions in my current creative workflow; however, I do feel flexible on that subject. (I.e. I'm not deeply committed to any specific features of GDocs.)

I don't expect my attitude here to be easy to deal with. I wish the world were simpler in some ways. Thanks in advance for any dialogue which may advance our shared goals.


Danyl Strype Mon 4 Mar 2019 8:37AM

I've already said a lot here, and so has Greg, so I'll wait for others to contribute before saying much more. However, I just want to made a few points about the VPN situation.

1) VPNs are costly. There's no guarantee that everyone in China (or other countries with net restrictions) who may be interested in learning about open app ecosystems can afford the ongoing cost.

2) VPNs are unreliable. The VPN we use stopped working on my GNU/Linux laptop for a few months at the end of last year. All VPNs can be compromised within China for days or weeks because of crackdowns when important events happen, as has been happening for the past few days.

EDIT: 3) VPNs are illegal in China, and while ex-pats are unlikely to get into serious trouble for using one, the same isn't necessarily true for Chinese nationals.

I have no objection to folks using datafarms like GDocs as a backup, to make sure the failure of a more ethical service doesn't result in catastrophic data loss (although making sure it's being archived by a service like web.archive.org is a more ethical way to achieve the same thing). But I see it as an act of exclusion to post datafarm links here, when there are so many other ways of sharing the same work (as mentioned in the OP), especially when it's just text.

I don't expect my attitude here to be easy to deal with.

I will admit to finding it both confusing and frustrating.


Greg Cassel Wed 6 Mar 2019 8:25PM

@strypey If much interest/ momentum develops among others here to adopt any of your suggestions such as git.nzoss.org.nz , git.feneas.org or Git.coop, I will consider it my responsibility as an OAE member to study their technical and social dependencies more closely.


Danyl Strype Tue 5 Mar 2019 10:15AM

The WikiMedia Foundation are launching a "global consultation about communication", so any insights shared here about how folks communicate within and between projects (both processes and tools) could be somehow fed into that. I encourage anyone involved in collaborative wiki projects to take part in this consultation, especially if you use MediaWiki.


Danyl Strype Tue 14 Nov 2023 4:30AM

Intriguing to revisit this thread 4 years later, after this topic came up again recently in another Loomio group. I wonder if @Greg Cassel still feels the same way about these issues?

I can't think of anything much I'd change if this conversation were starting today. Other than acknowledging that both CoActivate and Feneas turned out to be less durable than I gave them credit for. But RiseUp, Loomio, NZOSS and Git.coop are still going strong, as are a growing number of platform cooperatives and other community-hosting services.