What if we killed the mailing list altogether

Chef Infra (archive)
21

Let's rephrase then :slight_smile:

*We're moving to Google Groups shortly, that's absolutely the right

decision. Is it worth moving to Discourse later?*

Discourse has mailing list support, according to its website, so we need
not lose anything at all by using it. I've used Google Groups before, but
not so much Discourse. Here's the relevant advantages over Google Groups I
see just looking at the feature list:

  • Markdown support, with consistent, rich formatting
  • Embeddable on the Chef website, so new visitors can feel they are
    still under the same roof
  • Single Sign-On
  • Sticky posts
  • @ summoning
  • Trust system to identify people who consistently help
  • Muting, flagging, etc.

I haven't figured out any disadvantages here--if there are any Discourse
features that make it worse, or if there are any Groups features that
Discourse is missing or does worse at. Help me out?

--John

On Wed, Nov 26, 2014 at 8:13 AM, Adam Jacob adam@getchef.com wrote:

I'm not blocking groups :slight_smile: I was just asking. :slight_smile:
On Nov 26, 2014 8:09 AM, "Brian Hatfield" bmhatfield@gmail.com wrote:

I'm also perfectly happy with a standard mailing list. In particular,
other open-source projects that I care about also have mailing lists in a
similar form, which makes it easy for me to filter and follow projects that
I am interested in.

In my opinion, there is a surprising difference between
tools-that-send-emails and an honest-to-goodness mailing list.

Your project requirements sound sane, however. Google groups is a common
answer, as mentioned earlier; in fact, I thought we were already using
Google groups because I wasn't paying attention.

On Wed, Nov 26, 2014 at 11:04 AM, John de la Garza <
john.garza@rallyhealth.com> wrote:

I'm a fan of keeping the mailing list.

On Wed, Nov 26, 2014 at 10:19 AM, Nathen Harvey nharvey@getchef.com
wrote:

On Tue, Nov 25, 2014 at 7:39 PM, Adam Jacob adam@getchef.com wrote:

And just moved all of our conversation over to Discourse?

If only because it's questions like this that have killed the
migration in
the past. We get into analysis paralysis about which tool should we
migrate
to. We analyze the pros and cons of each. Try some of them out and
then
give up on the project because, while a priority, it doesn't have a
high
enough priority to maintain our attention span.

Our goals in this project include:

  • Migrate off of an in-house managed mail list system
  • Provide "mail in" and "mail out" on a reliable, easy-to-use platform
  • Provide an archive of previous mail list content

I'm all for debate and looking at alternate solutions but this will
prolong
the migration process and may even stall it out completely again.

There is a proposal on StackExchange 0 for a Chef group. As of this
morning, it still needs 30 more questions with a score of 10 or more
to move
on to the next phase. Help get behind that if you think it's
appropriate.

Google groups, with all of it's challenges, is stable, reliable, and
easy-to-use.

My vote is to move forward with the approved RFC [1] and I think we're
on
track to complete that work by the end of the calendar year.

http://www.discourse.org/

Adam

[1]:

https://github.com/opscode/chef-rfc/blob/master/rfc028-mailing-list-migration.md

22

On Nov 26, 2014, at 11:16 AM, Daniel DeLeo dan@kallistec.com wrote:

With this group, I think you’re gonna get a fair bit of survivor bias. That said, I like email as an interface, and I think GGroups has decent enough usability for a casual user (subscribe to a single thread, etc.).

I have friends who work at Google, including the former Gmail SRE. While some of them use Google Apps For Your Domain, none of them use Google Groups, and they all recommend to avoid it. It's unsupported. Use at your own risk, and be aware that if/when it dies, they most likely will not even notice that it's gone, much less likely to be bringing it back.

IMO, RFC28 is misinformed and misguided.

--
Brad Knowles brad@shub-internet.org
LinkedIn Profile: http://tinyurl.com/y8kpxu

23

On Wed Nov 26 10:03:36 2014, Brad Knowles wrote:

On Nov 26, 2014, at 11:16 AM, Daniel DeLeo dan@kallistec.com wrote:

With this group, I think you’re gonna get a fair bit of survivor bias. That said, I like email as an interface, and I think GGroups has decent enough usability for a casual user (subscribe to a single thread, etc.).

I have friends who work at Google, including the former Gmail SRE. While some of them use Google Apps For Your Domain, none of them use Google Groups, and they all recommend to avoid it. It's unsupported. Use at your own risk, and be aware that if/when it dies, they most likely will not even notice that it's gone, much less likely to be bringing it back.

IMO, RFC28 is misinformed and misguided.

That still sounds like better support than we're giving to
lists.opscode.com.

24

On Nov 26, 2014, at 12:06 PM, Lamont Granquist lamont@opscode.com wrote:

That still sounds like better support than we're giving to lists.opscode.com.

If you want to pay someone to run the system, the folks at EMWD.com are well-respected in the Mailman community, and they can provide hosted mailing lists, with searchable archives, USENET gateways, and the whole shebang.

I'd offer to run all that for free on the python.org infrastructure, but I don't have enough volunteer time to do what needs to be done to help them with the python.org infrastructure so I wouldn't want to commit them to any further work in this space.

If I could manage to find a job somewhere that was more open-source friendly, then I would have available time to help provide whatever the Chef community wants in this space, as well as python.org, and ntp.org, etc....

--
Brad Knowles brad@shub-internet.org
LinkedIn Profile: http://tinyurl.com/y8kpxu

25

JK's analysis here of Discourse here is on-point. I'd consider
supporting an RFC that detailed these as benefits over GGRo for a
future-switch.

The badge system in discourse is pretty sweet, if you contribute
different types of topics, replies, etc, you can gamify that shit.
Trust your users, whatever. Fits with the LT stuff we're pushing.

Here's some info re: Discourse' anti-spam:
-- User trust levels:
https://meta.discourse.org/t/what-do-user-trust-levels-do/4924
-- https://meta.discourse.org/t/what-about-the-spam-problem/2724/8
-- Spam, bacon, sausage and blog spam: a JavaScript approach

My prime considerations are around: anti-spam, moderation, muting. I'd
like to be able to banhammer users so I don't see 'em (mute) or their
unfortunately curiosity-inspiring troll-threads. I expect that this
behavior could be channeled into more effective thread management or
moderation, but I'd settle for an individual mute for now :slight_smile:

--aj

On Thu, Nov 27, 2014 at 6:56 AM, John Keiser jkeiser@getchef.com wrote:

Let's rephrase then :slight_smile:

We're moving to Google Groups shortly, that's absolutely the right
decision. Is it worth moving to Discourse later?

Discourse has mailing list support, according to its website, so we need not
lose anything at all by using it. I've used Google Groups before, but not
so much Discourse. Here's the relevant advantages over Google Groups I see
just looking at the feature list:

Markdown support, with consistent, rich formatting
Embeddable on the Chef website, so new visitors can feel they are still
under the same roof
Single Sign-On
Sticky posts
@ summoning
Trust system to identify people who consistently help
Muting, flagging, etc.

I haven't figured out any disadvantages here--if there are any Discourse
features that make it worse, or if there are any Groups features that
Discourse is missing or does worse at. Help me out?

--John

On Wed, Nov 26, 2014 at 8:13 AM, Adam Jacob adam@getchef.com wrote:

I'm not blocking groups :slight_smile: I was just asking. :slight_smile:

On Nov 26, 2014 8:09 AM, "Brian Hatfield" bmhatfield@gmail.com wrote:

I'm also perfectly happy with a standard mailing list. In particular,
other open-source projects that I care about also have mailing lists in a
similar form, which makes it easy for me to filter and follow projects that
I am interested in.

In my opinion, there is a surprising difference between
tools-that-send-emails and an honest-to-goodness mailing list.

Your project requirements sound sane, however. Google groups is a common
answer, as mentioned earlier; in fact, I thought we were already using
Google groups because I wasn't paying attention.

On Wed, Nov 26, 2014 at 11:04 AM, John de la Garza
john.garza@rallyhealth.com wrote:

I'm a fan of keeping the mailing list.

On Wed, Nov 26, 2014 at 10:19 AM, Nathen Harvey nharvey@getchef.com
wrote:

On Tue, Nov 25, 2014 at 7:39 PM, Adam Jacob adam@getchef.com wrote:

And just moved all of our conversation over to Discourse?

If only because it's questions like this that have killed the
migration in
the past. We get into analysis paralysis about which tool should we
migrate
to. We analyze the pros and cons of each. Try some of them out and
then
give up on the project because, while a priority, it doesn't have a
high
enough priority to maintain our attention span.

Our goals in this project include:

  • Migrate off of an in-house managed mail list system
  • Provide "mail in" and "mail out" on a reliable, easy-to-use platform
  • Provide an archive of previous mail list content

I'm all for debate and looking at alternate solutions but this will
prolong
the migration process and may even stall it out completely again.

There is a proposal on StackExchange 0 for a Chef group. As of this
morning, it still needs 30 more questions with a score of 10 or more
to move
on to the next phase. Help get behind that if you think it's
appropriate.

Google groups, with all of it's challenges, is stable, reliable, and
easy-to-use.

My vote is to move forward with the approved RFC [1] and I think we're
on
track to complete that work by the end of the calendar year.

http://www.discourse.org/

Adam

[1]:

https://github.com/opscode/chef-rfc/blob/master/rfc028-mailing-list-migration.md

26

Am 26.11.2014 um 19:06 schrieb Lamont Granquist lamont@opscode.com:

On Wed Nov 26 10:03:36 2014, Brad Knowles wrote:

On Nov 26, 2014, at 11:16 AM, Daniel DeLeo dan@kallistec.com wrote:

With this group, I think you’re gonna get a fair bit of survivor bias. That said, I like email as an interface, and I think GGroups has decent enough usability for a casual user (subscribe to a single thread, etc.).

I have friends who work at Google, including the former Gmail SRE. While some of them use Google Apps For Your Domain, none of them use Google Groups, and they all recommend to avoid it. It's unsupported. Use at your own risk, and be aware that if/when it dies, they most likely will not even notice that it's gone, much less likely to be bringing it back.

IMO, RFC28 is misinformed and misguided.

That still sounds like better support than we're giving to lists.opscode.com.

Yes!

After some years of abstinence I re-subscribed to this ML but didn’t get any response mail. Turns out you have at least two systems that send an lists.opscode.com hostname while EHLO-ing but only one has the right reverse-delegation.

I guess you already lost many users who have a decent anti-spam mx configuration at this stage. I reported that through various channels (twitter, irc) but didn’t get an answer.

E.g.
https://twitter.com/rmoriz/status/512553752482807808

(Beware, lots of OT following:)

Despite the decision you are going to make, please consider providing more love to it. Many chef users I saw in the last couple of months don’t even know that there is a mailing list or irc channel (or couldn’t get a response/use out of it). IMHO chef has an adoption/community problem: If you know the ecosystem, the right cookbooks, the good parts: it’s brilliant! If you don’t know what’s going on/where to look for best practices and support, you’re lost.

Take - for example- the strategy of the cheap hosting provider DigitalOcean - They provide lots of nice and easy tutorials that not only rank quite well in Google but also slightly tie the users to their (commodity/kvm-vps) services. When someone searches for „chef best practices“ or „chef tutorials“, he will find https://learn.getchef.com/ which is just a tutorial about chef, not about solving real world problems (e.g., install and maintain a decent mailing list service in the cloud).

You’ll only reach people that are already sold to chef (or forced to use it) but not the unbiased user that want’s to level up in DevOps. Such users usually have a „real world problem they need to solve“ and will find „copy+paste“ tutorials or trending „how I solved everything in 5 minutes with docker, go and nodejs“ blogposts somewhere else and will never learn the benefits of chef and its toolchain.

IMHO this is an excellent occasion to write a real-word application-cookbook for a mailing list/discourse-combination and use it as an example in your tutorial/docs/blog :wink:

best
Roland