couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Noah Slater <>
Subject Re: [NOTICE] Create marketing@ lis (Was: Re: Marketing suggestion)
Date Mon, 03 Feb 2014 18:14:46 GMT

Sorry if I gave you that impression. I read your mail in full, and
take on board all of your points. I just don't want to drag this out
into a needlessly long discussion that tires people out.

We do a lot of *talking* in this project. It would be nice if we could
make decisions quicker and get on with *work* though. ;)

So to repeat: I understand your concerns. I'd like to do this as an
experiment. I will take responsibility for it, and am happy to report
back to the PMC at set time periods. If it doesn't work, we can close
the list, and move things back.

I split this thread off like this because it is a separate topic. And
I am using Jan's email subject tags to properly notify the whole list
that lazy consensus is in effect.

If you want to raise a formal objection, this is your opportunity. But
please do so unambiguously. If you do not raise a formal objection, I
will request the list in a few days, and proceed with the plan. If you
do, I will cancel lazy consensus and start a majority approval vote.

One of the problems we've had (IMO) is that we do generate a lot of
discussion. I think we need to make decisions faster. I only have a
certain amount of hours to spend on the project each week. I could be
using them better. :)

That doesn't mean we cannot talk about things. But if you look back at
the big threads over the last few years: we spend weeks and weeks
talking a subject to death, and then nothing happens. This is a
repeating pattern.

So. This is an easier and more efficient way to do it.

1. I propose an action
2. People bring up some concerns
3. I modify my proposal to address those concerns
4. I state that lazy consensus is in effect (where we are now)

Then it goes one of two ways:

a) Nobody objects, the proposal passes

b) Somebody objects and either the idea is canned, discussed a bit
more, or taken to a vote

Either way: the focus is on making progress. Either put the idea to
bed and accept it was a bad idea, or JFDI. :)

On 3 February 2014 18:41, Benoit Chesneau <> wrote:
> Noah,
> I wonder why I took the time to elaborate on that topic to see all my
> concerns and *objections* properly ignored just because you don't care about
> what does the others (which had a known and accepted success in their
> strategy)  or because you just disagree without giving much reason. I would
> have preferred a formal discussion and a more interesting answer that could
> have eventually convinced me. You are talking how unfriendly some people are
> finding this mailing-list, I find your answer particularly unfriendly and
> not very open.
> Anyway let's forget that part and let me quickly answer. You may be not
> remember but I was disagreeing about the creation about the erlang ml, not
> finding it particularly useful. And its emptiness since gives me reason
> somehow. I still don't see any reason to this list, and I am probably not
> registered to it ( I forgot since).
> For the others I wasn't particularly available at the time they were created
> those I have no objections to them since their goal make them apart from the
> current topic. i10n may become really noisy soon (which I wish). And
> replication have to exist if the goal is to create a neutral spec widely
> used in other projects (this the way I understood its creation).
> Telling me about a project I never heard except on this mailing-list doesn't
> help me either to find a good reason for it. Though I will look at it.
> I still have some concerns with an advocacy list (marketing is definitely
> not the right term, this is not a market), since you choosed to ignore it,
> that may not have any sense. I will just say that I agree to disagree then.
> - benoit
> On Mon, Feb 3, 2014 at 3:37 PM, Noah Slater <> wrote:
>> Benoit,
>> This is a notice that I am going to assume lazy consensus on the
>> proposal to create a marketing@ list. If you have a formal objection
>> to raise, please do so now, and I will move this to a vote.
>> We have the following lists:
>> erlang@ - Created specifically to create "safe space" for people to
>> get up to speed on Erlang
>> l10n@ - Created specifically to create a "focused space" for people to
>> do translations
>> replication@ - Created specifically to create a "focused space" for
>> people discussing replication, etc
>> You say this:
>> "Also due to the low volumes of mails on @dev this shouldn't be a
>> problem."
>> But this is not the common perception. In fact, there is a lot of
>> traffic to our dev@ mailing list. Way too much for most non-devs to
>> cope with. I even know current devs who find the traffic from CouchDB
>> hard to deal with.
>> On top of that, our dev@ list can be a bit of a hostile and scary
>> place. I have had direct feedback on this point. So I am worried that
>> there are people who are not participating because they don't want to
>> be on dev@.
>> So my goal here is to create a "safe/focused" place where people who
>> are interested in the "softer" side of marketing and project/community
>> growth to hang out and discuss things without:
>> a) Having to feel put off by devs or dev discussion
>> b) Having to feel like they are wasting people's time/bandwidth/attention
>> "Having a marketing list is also quite uncommon in an opensource
>> projects."
>> I don't care. We find what works for us, not what works for other
>> people. Though, as you mention it, the idea for a marketing@ list
>> comes from Apache CloudStack. They have one, and it is working out
>> just fine for them. They get lots of non-dev participation, which is
>> exactly the sort of thing I am hoping for. You don't have to be able
>> to code to contribute to CouchDB.
>> "When a project starts to have more than 2 lists it starts to be
>> really annoying to track and quite expensive."
>> Expensive in what sense? We already have a number of lists. I think
>> this expansionism is a good thing. If the lists don't work, it's not a
>> problem. We close the list, and we move the discussion back to dev.
>> This is a reversible experiment.
>> "I'd be in favour of keeping the number of lists small until it
>> becomes clear that some topic needs to spin off into its own list."
>> We didn't do this for erlang@, or l10n@, or replication@. In each
>> case, we identified that there might be some discussion which is *not
>> happening yet because the dev@ list is not a good place for it*.
>> "While the volume of marketing emails is low, it's not hard for devs
>> who aren't that interested in marketing to ignore them, just as those
>> who aren't interested in specific dev topics can ignore those."
>> I believe that just like people talking about how to learn Erlang, and
>> people talking about translation, and people talking about third-party
>> apps, the reason the volume is so low is because there is/was no place
>> to talk about it.
>> Our dev list is noisy, can be unfriendly, and is mostly focused on
>> dbcore dev. (Unsurprisingly.)
>> That's fine. But there are other areas to contribute. And I outright
>> reject the idea that you need to know anything about dbcore or Erlang
>> or even how to programme to be able to contribute to CouchDB.
>> I don't want to dwell on this. I appreciate the discussion, but I
>> don't want to get lost in the weeds.
>> Having acknowledged the concerns raised, I will keep a close eye on
>> the marketing@ list and assume responsibility for it. I can provide
>> oversight, and am happy to report on progress in three months, six
>> months, and so on.
>> If this isn't good enough, please raise a formal objection to the
>> proposal. I will then attempt to call a majority consensus vote so
>> that we can get this over and done with.
>> On 3 February 2014 10:46, Benoit Chesneau <> wrote:
>> >
>> >
>> >
>> > On Mon, Feb 3, 2014 at 10:27 AM, Andy Wenk <> wrote:
>> >>
>> >> On 3 February 2014 10:14, Benoit Chesneau <> wrote:
>> >>>
>> >>> On Mon, Feb 3, 2014 at 10:03 AM, Andy Wenk <> wrote:
>> >>>>
>> >>>> On 3 February 2014 08:42, Benoit Chesneau <>
>> >>>> wrote:
>> >>>>>
>> >>>>> On Sun, Feb 2, 2014 at 3:16 PM, Noah Slater <>
>> >>>>> wrote:
>> >>>>>>
>> >>>>>> Ashley,
>> >>>>>>
>> >>>>>> Wrt marketing plans: yes, but half way between my head,
and my
>> >>>>>> private
>> >>>>>> notes. Unfortunately, my private notes also contain things
>> >>>>>> private conversations with people. Major mistake on my part.
>> >>>>>> Apologies
>> >>>>>> to the community.
>> >>>>>>
>> >>>>>> I've just sent an email giving a few people notice that
I plan to
>> >>>>>> start moving things over to the wiki. Hopefully over the
next week
>> >>>>>> or
>> >>>>>> so I can get all of our existing marketing ideas in a communal
>> >>>>>> space
>> >>>>>> so we can start to discuss it.
>> >>>>>>
>> >>>>>> As for the marketing@ list: great. So what we'll do now
is wait
>> >>>>>> another day or two. If nobody objects, we can make the list.
>> >>>>>> is
>> >>>>>> how we make most of our decisions on the project.
>> >>>>>
>> >>>>>
>> >>>>> I am not sure it's a good idea to have a marketing list. Marketing
>> >>>>> should be linked to dev and vice-versa . It's important that
>> >>>>> marketing
>> >>>>> follows dev discussion and that dev follows and interact with
>> >>>>> marketing.
>> >>>>> Having 2 mailing-lists will create a disconnection. Which is
>> >>>>> path to
>> >>>>> the failure in tech. Also due to the low volumes of mails on
>> >>>>> this
>> >>>>> shouldn't be a problem.
>> >>>>>
>> >>>>> - benoit
>> >>>>
>> >>>>
>> >>>> hm ... I understand exactly what you mean and I agree, if we would
>> >>>> speak
>> >>>> of a company with different big departments here. But in our project
>> >>>> I think
>> >>>> it is totally ok that we have two different lists and the people
>> >>>> are
>> >>>> strongly interested in both parts should subscribe both lists. The
>> >>>> advantage
>> >>>> imho is to not flood the dev@ list with unrelated stuff ...
>> >>>
>> >>>
>> >>>
>> >>> Why do you think it would be different because we are an opensource
>> >>> project? If marketing people don't want to follow all devs discussion
>> >>> then
>> >>> there is some perspective problem imo. The same for devs that ignore
>> >>> the
>> >>> users perspectives. Marketing should be elaborated with all the devs,
>> >>> not in
>> >>> a side corner. At least this what we learn in management schools. And
>> >>> this
>> >>> is really true for a **neutral**  opensource project which has no
>> >>> business
>> >>> perspective (and shouldn't have).
>> >>>
>> >>> - benoit
>> >>
>> >>
>> >> I did not mean to see it differently because we are an OpenSource
>> >> project
>> >> but because of the size of the project. I don't think that we will have
>> >> the
>> >> situation, that the marketing activities are going into a different
>> >> direction because of having two lists. I still believe that everything
>> >> is
>> >> very transparent. Having more lists does not lead to in-transparencies
>> >> but
>> >> will lead in more focused discussions. The connection between marketing
>> >> and
>> >> development targets is created by the interest people have - and they
>> >> should
>> >> be interested in both and should therefor subscribe both lists ... if
>> >> they
>> >> don't they are not interested in marketing activities (what is ok for
>> >> me).
>> >> But I agree that if no dev will subscribe the marketing list, we will
>> >> have
>> >> the marketing activities in a side corner ...
>> >>
>> >>
>> >>
>> >
>> > this is the " if they don't they are not interested in marketing
>> > activities"
>> > which is problematic. By marketing in a community project, I often mean
>> > every actions taken to grow the community. I can't imagine a dev not
>> > interested by it. Having a marketing list is also quite uncommon in an
>> > opensource projects. But to be more concrete I often take the zeromq
>> > project
>> > as a template to build a successful community, When you see the
>> > mailing-lists attached to the project [1] you only have 2. If you take a
>> > recent success in communication, the docker project, this is the same
>> > [2].
>> >
>> > Imo this is part of its success. While it's totally fine to multiply the
>> > annonces channels, I do think that a community and its members  should
>> > act
>> > together when it's about core community discussions. Part of these core
>> > discussions are:
>> >
>> > - dev discussions : features/roadmap/status
>> > - community discussions
>> > - users discussions about some features
>> >
>> >
>> > Also lot of peopple are already subscribed to more than XXX list, to
>> > follow
>> > N projetcs daily (customer purpose, survey...). When a project starts to
>> > have more than 2 lists it starts to be really annoying to track and
>> > quite
>> > expensive.
>> >
>> > - benoit
>> >
>> >
>> > [1]
>> > [2]
>> --
>> Noah Slater

Noah Slater

View raw message