Proposal: Moving from lists.opscode.com to googlegroups… any concerns/objections?


#1

Proposal:
We’re thinking about migrating the emails lists we’ve been hosting at lists.opscode.com to googlegroups, and I’d love your feedback if you think this is a terrible idea.

Rationale:
The reason for this is that I’d rather Opscode put resources into new community.opscode.com features & fixes than maintain and enhance the listserv. We don’t currently have so many lists or special security requirements that justify our own listserv infrastructure. We would all benefit from better search, easier interfaces, and lower friction.

Scope:

Impact:
Current subscribers will be imported to the new list, so all you will have to do is change where you send mail to.
We’ll make a static archive of the old posts, and I’ll attempt to have MarkMail do an ingest.

History:
When I originally insisted (in 2008) we host our own listservs, it is because I felt that we would have a proliferation of private lists which would require special authentication… and I didn’t want to rely on google. Almost everyone I have ever talked to about this has told me that was dumb and stifling community growth.

Request:
Please let me know either privately or on list here if you have objections, suggestions, etc. I’ll consider them, and adjust as reasonable/appropriate.

Time:
I’ll make a final decision next Friday, June 22nd.

Thanks,
Jesse


Jesse Robbins | Cofounder - Opscode, Inc.
e: jesse@opscode.com | +1-206-508-7449 | http://www.opscode.com


#2

On Jun 14, 2012, at 2:34 PM, Jesse Robbins wrote:

Rationale:
The reason for this is that I’d rather Opscode put resources into new community.opscode.com features & fixes than maintain and enhance the listserv. We don’t currently have so many lists or special security requirements that justify our own listserv infrastructure. We would all benefit from better search, easier interfaces, and lower friction.

Speaking only for myself, I’ve been involved in Mailman mailing list administration for a number of years, and I also have experience with other tools like Majordomo and Sympa, as well as being on a couple of Google Groups lists.

Frankly, of these four programs, Google Groups does not compare well. Modern versions of Majordomo have greatly improved since the one I worked with, but I made the switch to Mailman a long time ago, and have never looked back.

From what I’ve seen, Sympa is somewhat better than Mailman 2.1.x in some ways but Mailman 2.1.x is somewhat better than Sympa in others, and overall it seems to be a wash. Sympa does seem to be more targeted towards the International and educational communities, whereas Mailman is more targeted towards the Linux/Open Source communities. As a business, you’d want to figure out what parts of those descriptions fit you better.

Of course, work on Mailman 3 is well underway, and work on the mail archive handling tools is planned for Google Summer of Code 2012, among many other things. There are some people who are already using the early access alpha and the first beta (see http://www.mail-archive.com/mailman-developers@python.org/msg12551.html) and doing productive work in hosting their lists.

I’ve supported Mailman-hosted mailing lists for a variety of open source projects, and this is definitely something that can be done on a volunteer basis. For small sites (such as yours), I would honestly expect no more than an hour or two per week that would need to be spent in managing the mailing list system.

History:
When I originally insisted (in 2008) we host our own listservs, it is because I felt that we would have a proliferation of private lists which would require special authentication… and I didn’t want to rely on google. Almost everyone I have ever talked to about this has told me that was dumb and stifling community growth.

Stifling community growth? A goal of reducing friction? I’m at a loss to understand how these issues would be resolved by switching to Google Groups. You’d still have to have people designated to monitor the mailing list system and to be the POCs to contact Google when there are problems, but then it would be entirely out of your hands as to when (or if) the problems ever actually get resolved.

We know that offshoring doesn’t really work, not when you compare total lifecycle costs. Same for many types of outsourcing. I would honestly like to see some TCO and ROI numbers that would prove that switching to Google Groups would actually do anything beneficial for you.


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


#3

±-----------------------------------------------------------------------------
| On 2012-06-14 20:26:58, Brad Knowles wrote:
|
| We know that offshoring doesn’t really work, not when you compare total lifecycle costs. Same for many types of outsourcing. I would honestly like to see some TCO and ROI numbers that would prove that switching to Google Groups would actually do anything beneficial for you.

Outsourcing can work fine. It’s how most businesses actually get paid. Are you
suggesting that Hosted Chef is bad business, for instance? :wink:

Google Groups has a subjectively awful interface, and requires a Google
account. Beyond that, it generally works. Spice generally flows. My experience
with Google paid support has been pretty poor, however.

Puppet moved to Google Groups a while ago, and as far as I know there wasn’t a
huge backlash from the community there.

I have no personal issue with the lists moving, because at the end of the day
it doesn’t actually affect anything: mail gets delivered to my inbox. Huzzah.
Just be aware of possible support problems.

The issue Jesse (I’m guessing) is trying to solve is that managing email sucks,
and he’d rather have his guys working on his actual product than dealing with
their network provider dropping PTRs again (lots of chef mail held in my
discards for require_ptr, sadly.)

Brad and I are on some of the same super secret email admin lists, so we
breathe this particular brand of sewage, but it may be he’s forgotten that
managing email sucks if you don’t have your email respirator handy.

There’s a huge win to be had from outsourcing email if your core competency is
not dealing with deliverability, antispam, and so forth. We outsource it at my
current dayjob to Google, and while I hate gmail, I am often very happy to
never have to fix email myself.

Cheers.

(FD: I ran infra for pobox.com and listbox.com, both long-standing ESPs, for
six years.)

bdha
cyberpunk is dead. long live cyberpunk.


#4

We have moved out of lotus notes and switched to google apps. Its has been
a great experience, especially for end users. I dont see any major concern,
maintaining in house mailing system brings additional pains, not only
related to the mail solution itself, but also the backup and e-discovery
requirments. Considering those options gmail (or gmail+postini) might be a
better choice if you can afford it (cost, and compliance can be a
bottleneck).

When we migrated to gmail/google groups (3 years back) pushing the archived
data inside gmail was difficult and we had to do lot of tooling around
that, i think the situation has changed now,

regards
ranjib

On Fri, Jun 15, 2012 at 11:41 AM, Bryan Horstmann-Allen <
bdha@mirrorshades.net> wrote:

±-----------------------------------------------------------------------------
| On 2012-06-14 20:26:58, Brad Knowles wrote:
|
| We know that offshoring doesn’t really work, not when you compare total
lifecycle costs. Same for many types of outsourcing. I would honestly
like to see some TCO and ROI numbers that would prove that switching to
Google Groups would actually do anything beneficial for you.

Outsourcing can work fine. It’s how most businesses actually get paid. Are
you
suggesting that Hosted Chef is bad business, for instance? :wink:

Google Groups has a subjectively awful interface, and requires a Google
account. Beyond that, it generally works. Spice generally flows. My
experience
with Google paid support has been pretty poor, however.

Puppet moved to Google Groups a while ago, and as far as I know there
wasn’t a
huge backlash from the community there.

I have no personal issue with the lists moving, because at the end of the
day
it doesn’t actually affect anything: mail gets delivered to my inbox.
Huzzah.
Just be aware of possible support problems.

The issue Jesse (I’m guessing) is trying to solve is that managing email
sucks,
and he’d rather have his guys working on his actual product than dealing
with
their network provider dropping PTRs again (lots of chef mail held in my
discards for require_ptr, sadly.)

Brad and I are on some of the same super secret email admin lists, so we
breathe this particular brand of sewage, but it may be he’s forgotten that
managing email sucks if you don’t have your email respirator handy.

There’s a huge win to be had from outsourcing email if your core
competency is
not dealing with deliverability, antispam, and so forth. We outsource it
at my
current dayjob to Google, and while I hate gmail, I am often very happy to
never have to fix email myself.

Cheers.

(FD: I ran infra for pobox.com and listbox.com, both long-standing ESPs,
for
six years.)

bdha
cyberpunk is dead. long live cyberpunk.


#5

On Jun 14, 2012, at 8:41 PM, Bryan Horstmann-Allen wrote:

Outsourcing can work fine. It’s how most businesses actually get paid. Are you
suggesting that Hosted Chef is bad business, for instance? :wink:

I didn’t say that no outsourcing works, just that many types don’t work. At the very least, many types of outsourcing don’t provide anything remotely close to the level of cost savings as was used to justify the action.

Where outsourcing can work well is when you look at increasing the value of the resulting product, as opposed to simply cutting costs. There’s a limit to how much you can cut costs, and every time you add an administrative barrier between any two sets of resources, that barrier comes with some additional costs of its own, many of which are unlikely to be known in advance. On the flip side, there is no limit to how much you can increase the value of a product.

IMO, Hosted Chef falls on the “increased value” side of that equation.

The issue Jesse (I’m guessing) is trying to solve is that managing email sucks,
and he’d rather have his guys working on his actual product than dealing with
their network provider dropping PTRs again (lots of chef mail held in my
discards for require_ptr, sadly.)

Brad and I are on some of the same super secret email admin lists, so we
breathe this particular brand of sewage, but it may be he’s forgotten that
managing email sucks if you don’t have your email respirator handy.

I’m willing to admit that possibility. I guess it all comes down to the actual day-to-day problems that Jesse is trying to fix. I certainly didn’t see anything in his previous message that alluded to outsourcing the “managing e-mail sucks” problem, but maybe I wasn’t reading between the lines correctly.

There’s a huge win to be had from outsourcing email if your core competency is
not dealing with deliverability, antispam, and so forth. We outsource it at my
current dayjob to Google, and while I hate gmail, I am often very happy to
never have to fix email myself.

I certainly got out of the business of managing e-mail, and I did manage to draft Ralf Hildebrandt and Patrick Koetter to help with managing the e-mail and mailing list services for python.org – as the authors of “The Book of Postfix”, they certainly knew a lot more about that program than I did, and they were able to easily pick up what little advantage I had over them with regards to managing the Mailman part.

I have known for many years how hard it is to properly manage a mail system, and that’s why my vanity domains have been hosted at Heller Information Systems (his.com) since the day they were created – I know Paul Heller and his team, and I know that they can do a better job of managing my e-mail as one of their customers than I could do for myself. At least I was able to avoid that “Cobblers Children” problem for my own personal e-mail during the times I was doing large-scale mail system administration/engineering for my various employers.

If that’s the problem Jesse is trying to solve, then I’ll sit back down and shut up, because even I was smart enough to finally get out of that business.


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


#6

I vote for Google Groups simply for the reason that it makes it much easier
to search through the archive and link to it

iirc, google groups has a feature where a archive url is appended to each
message. That would be very helpful to us in putting together What’s Cookin
for the Foodfightshow

on the negative side, I have seen a lot more spam on google groups than I
have seen on the current chef ML platform.

On Fri, Jun 15, 2012 at 5:59 AM, Brad Knowles brad@shub-internet.orgwrote:

On Jun 14, 2012, at 8:41 PM, Bryan Horstmann-Allen wrote:

Outsourcing can work fine. It’s how most businesses actually get paid.
Are you
suggesting that Hosted Chef is bad business, for instance? :wink:

I didn’t say that no outsourcing works, just that many types don’t work.
At the very least, many types of outsourcing don’t provide anything
remotely close to the level of cost savings as was used to justify the
action.

Where outsourcing can work well is when you look at increasing the value
of the resulting product, as opposed to simply cutting costs. There’s a
limit to how much you can cut costs, and every time you add an
administrative barrier between any two sets of resources, that barrier
comes with some additional costs of its own, many of which are unlikely to
be known in advance. On the flip side, there is no limit to how much you
can increase the value of a product.

IMO, Hosted Chef falls on the “increased value” side of that equation.

The issue Jesse (I’m guessing) is trying to solve is that managing email
sucks,
and he’d rather have his guys working on his actual product than dealing
with
their network provider dropping PTRs again (lots of chef mail held in my
discards for require_ptr, sadly.)

Brad and I are on some of the same super secret email admin lists, so we
breathe this particular brand of sewage, but it may be he’s forgotten
that
managing email sucks if you don’t have your email respirator handy.

I’m willing to admit that possibility. I guess it all comes down to the
actual day-to-day problems that Jesse is trying to fix. I certainly didn’t
see anything in his previous message that alluded to outsourcing the
"managing e-mail sucks" problem, but maybe I wasn’t reading between the
lines correctly.

There’s a huge win to be had from outsourcing email if your core
competency is
not dealing with deliverability, antispam, and so forth. We outsource it
at my
current dayjob to Google, and while I hate gmail, I am often very happy
to
never have to fix email myself.

I certainly got out of the business of managing e-mail, and I did manage
to draft Ralf Hildebrandt and Patrick Koetter to help with managing the
e-mail and mailing list services for python.org – as the authors of “The
Book of Postfix”, they certainly knew a lot more about that program than I
did, and they were able to easily pick up what little advantage I had over
them with regards to managing the Mailman part.

I have known for many years how hard it is to properly manage a mail
system, and that’s why my vanity domains have been hosted at Heller
Information Systems (his.com) since the day they were created – I know
Paul Heller and his team, and I know that they can do a better job of
managing my e-mail as one of their customers than I could do for myself.
At least I was able to avoid that “Cobblers Children” problem for my own
personal e-mail during the times I was doing large-scale mail system
administration/engineering for my various employers.

If that’s the problem Jesse is trying to solve, then I’ll sit back down
and shut up, because even I was smart enough to finally get out of that
business.


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


#7

On Jun 14, 2012, at 8:26 PM, Brad Knowles wrote:

Stifling community growth? A goal of reducing friction? I’m at a loss to understand how these issues would be resolved by switching to Google Groups.

Great point to challenge.

I think it is a lot easier for new people to discover, search, and join google groups than any of the other tools available without making larger investment in integration and maintenance.

I’m not suggesting that we won’t make that investment at a later stage. In particular, as the community grows and new subgroups form, it will likely make sense to move back to our own dedicated/managed lists as part of a bigger community toolchain. (as other large communities did)

We have a huge backlog of features we want to ship for the current community site & resources that go far beyond an email list.

On Jun 14, 2012, at 8:26 PM, Brad Knowles wrote:

I would honestly like to see some TCO and ROI numbers that would prove that switching to Google Groups would actually do anything beneficial for you.

The existing listserv needs to be upgraded, and is running in an old environment that needs to be migrated. In this case, migrating to google groups provides the advantages I outlined above and lets us get back to building new stuff. As we reach the next stage of community growth, we can re-evaluate.

Does that better explain my thinking here?

-Jesse


#8

On Jun 15, 2012, at 12:57 PM, Jesse Robbins wrote:

The existing listserv needs to be upgraded, and is running in an old environment that needs to be migrated. In this case, migrating to google groups provides the advantages I outlined above and lets us get back to building new stuff. As we reach the next stage of community growth, we can re-evaluate.

Does that better explain my thinking here?

Yup. I hereby withdraw all previous questions on this topic. Thanks!


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


#9

On 2012-06-14 14:34:25, Jesse Robbins wrote:

We’re thinking about migrating the emails lists we’ve been
hosting at lists.opscode.com to googlegroups, and I’d love your
feedback if you think this is a terrible idea.

As long as the (email) list continues to flow it makes no
difference who hosts it to me.

I think you have to create a google account to subscribe to
a list. This is bit annoying but not the end of the world.

The vim-users mailing list migrated to googlegroups a while back.
It works well now, but there were problems for a while. Check
their archives. If I recall someone from Google was subscribed,
and took care of things, otherwise support might have been
problematic.

/Allan

Allan Wind
Life Integrity, LLC
http://lifeintegrity.com


#10

On Jun 14, 2012, at 3:20 PM, Allan Wind wrote:

I think you have to create a google account to subscribe to
a list. This is bit annoying but not the end of the world.

Indeed, the last Google Group I signed up for (devops-toolchain) did require use of a Google account. And yes, that is a bit of a pain for those of us who see Google as the next Great Evil.

The vim-users mailing list migrated to googlegroups a while back.
It works well now, but there were problems for a while. Check
their archives. If I recall someone from Google was subscribed,
and took care of things, otherwise support might have been
problematic.

Speaking as one of the postmasters for python.org (albeit not very active), and as a member of the Mailman Steering Group (a.k.a., the Mailman Cabal), I’m happy to do whatever I can to help make the mailing lists at Opscode work better.

I know that searching the archive has been a big point of pain for a long time, but it’s trivially easy to set up a USENET gateway which can then be accessed with all the standard tools, and it’s easy enough to set up subscriptions to mail-archive.com (or any of the other similar free services).


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


#11

Hi,

On Fri, Jun 15, 2012 at 1:41 PM, Bryan Horstmann-Allen
bdha@mirrorshades.net wrote:

There’s a huge win to be had from outsourcing email if your core competency is
not dealing with deliverability, antispam, and so forth.

Indeed. Google groups is more than sufficient and even if you only
lose 2-3 hours a week maintaining the current setup, I would prefer
that was spent improving the community.opscode.com web site.


Cheers,

Peter Donald


#12

All,

On 15 June 2012 00:20, Allan Wind allan_wind@lifeintegrity.com wrote:

I think you have to create a google account to subscribe to
a list. This is bit annoying but not the end of the world.

Last time I checked, one doesn’t need a Google account to subscribe to a
Google Group. If the chef list would be moved to Google Groups as “
chef@googlegroups.com”, subscribing via email is as simple as sending a
mail to:

chef+subscribe@googlegroups.com

Similar for unsubscribing. More info here:

Ringo


#13

I’m currently indifferent on this matter. I’ve noticed no problem with
the Sympa lists, but it makes sense to avoid the allocation of future
resources. That being said, those resources may come into play in the
event google groups are problematic regardless.

Jesse, count me as a +1

–AJ

On 21 June 2012 18:37, Ringo De Smet ringo.desmet@gmail.com wrote:

All,

On 15 June 2012 00:20, Allan Wind allan_wind@lifeintegrity.com wrote:

I think you have to create a google account to subscribe to
a list. This is bit annoying but not the end of the world.

Last time I checked, one doesn’t need a Google account to subscribe to a
Google Group. If the chef list would be moved to Google Groups as
"chef@googlegroups.com", subscribing via email is as simple as sending a
mail to:

chef+subscribe@googlegroups.com

Similar for unsubscribing. More info here:

http://www.mydigitallife.info/how-to-subscribe-or-join-google-groups-without-google-account/

Ringo


#14

On Thu, Jun 21, 2012 at 9:37 AM, Ringo De Smet ringo.desmet@gmail.com wrote:

On 15 June 2012 00:20, Allan Wind allan_wind@lifeintegrity.com wrote:

I think you have to create a google account to subscribe to
a list. This is bit annoying but not the end of the world.

Last time I checked, one doesn’t need a Google account to subscribe to a
Google Group. If the chef list would be moved to Google Groups as
"chef@googlegroups.com", subscribing via email is as simple as sending a
mail to:

chef+subscribe@googlegroups.com

+1 with account it is as simple as two clicks. Needless to say you
can reply to any thread right from the web.

anatoly t.


#15

+1