On github there is an issue about a ‘community’ repo. I’m moving the discussion here to facilitate writing a survey to get feedback from a larger set of users.
Let’s build this survey!
Paging @dawnfoster
1 Like
I’m all in on gathering data! 
These will need some wordsmithing, but here is a very quick dump of my thoughts on a few initial questions that we could consider …
Demographics:
Are you a
- contributor to OpenSearch
- user of OpenSearch
- both
How long have you been contributing to open source projects:
- this is my first project
- less than a year
- 1 - 3 years
- 3 - 5 years
- 5+ years
Does your employer consider your participation in this community as part of your job?
- Yes
- No
Community Questions
Where would you expect to find common information that applies across most or all of the repos (community meetings, governance, values / principles, etc.)? (check all that apply)
- Website
- GitHub
- Forums
If you were looking for this information in GitHub, where would you expect to find it:
- in every repo
- in a dedicated repo called “community”
- other (please specify)
If you were looking for this information on the website, where would you expect to find it:
- on the home page
- in a section called “community”
- other (please specify)
1 Like
We should probably ask a couple of open ended questions, too. Something along the lines of
- what are we doing well?
- what could we improve?
i’d also add these:
What is your use-case for OpenSearch?
- Private project / playground
- Component in free open-source software
- Component in free closed-source software
- Component in commercial software
- Offering OpenSearch directly as a main component of a SaaS offering
(^- yes, AWS, i mean you with the last point 
)
Do you already have a (productive) Elasticsearch stack deployed?
- Yes
- No, but in the middle of a project
- No, just getting started
What are your plans for the future?
- Decided to move to OpenSearch, waiting for first release
- Undecided: considering to move to OpenSearch, but no decision taken
- Watching the OpenSearch project, but staying on Elasticsearch for the foreseeable future
What do you absolutely need (must-have) in OpenSearch in order to be able to move from Elasticsearch to OpenSearch?
(open-ended question)
What do you want (should-have) in OpenSearch but can live without for the time being in order to be able to move from Elasticsearch to OpenSearch?
(open-ended question)
On the topic of website vs. community, I am running a <sarcasm> highly scientific Twitter poll</sarcasm> which will run for another day, but early reports show that there are a fair number of us who expect to find common content in a community repo, but a lot more people who expect to find it on the website. I’ll be curious to see what we get from this survey for the OpenSearch folks.
1 Like
I wonder if we should have 2 surveys - one for contributors and another for users? Your questions seem more like user questions, and we probably want to make sure that all users can answer those questions, while the questions I proposed really only apply to the subset of users who plan to also contribute.
i presume that all contributors would also be users (the only contributors who aren’t users i could imagine would be some students “forced” to work on FOSS for their studies? though in that case i’d rather contribute to some game - then at least testing my changes would be fun ). having it in one survey would give you an overview of which kinds of users are most likely to be come contributors and maybe also to see why they become a contributor (usually that’ll be some missing feature) and whether they only contribute for the things which directly affect them or also help out with general topics (that might be a question in itself to be asked?).
note that all your “community questions” are user questions as well.
what might also be interesting to know about contributors would be their rough location / timezone; right now this forum seems to be clustered around the american timezones (at least based on the times when posts are being written). this is also visible in the times when community meetings are held, for us europeans this is 7pm (not a time where i particularly enjoy to still be working) which is why i’ve skipped them so far.
@ralph My point was actually that not all users will become contributors, and your questions might also be interesting to learn more about the users that won’t ever contribute to the project. Open source projects typically have huge numbers of users and a relatively small number of contributors. I’m focused on that small number of contributors.
Actually this isn’t true at all. Granted most contributors will also be users, but it’s wrong to assume that all of them will be. For example, I don’t currently use Elasticsearch / OpenSearch personally (I have used Elasticsearch in the past), but my company (VMware) has embedded Elasticsearch in a bunch of our products. I’m contributing to the project to help us decide whether OpenSearch is likely to build a community and innovate in the project in ways that will make it a viable project for us to embrace over the long-term. In the short-term, we’ll probably use OpenSearch, but over the longer-term, we might encourage our products to begin planning to use something else. I’m still on the fence
Also, there are loads of software companies, like VMware, who will contribute to projects like this because we have embedded them into our products. We may or may not actually be users of our own products in production environments, but we are certainly actively involved in their development and testing. As part of this development, we will want to contribute bug fixes / features back into the open source project, so these people would be contributors, but not necessarily users.
You could argue that product teams are using the software as part of their testing, but this use case is way different from the people (our customers and others) who are running OpenSearch in production.
With that said, this is not a hill I’m willing to die on if others things we should include some user questions in the contributor survey, but I do worry about making it too long, which will reduce the participation rate. I have a preference for shorter, more focused surveys.
As someone in the UK, I feel your pain. I’ve resigned myself to knowing that for communities like this, the meetings are going to be more friendly for the US west coast. Personally, I would love to have the meetings recorded with detailed minutes, so that I can catch up if I miss something. Every other open source project I participate in has recorded meetings with minutes, so there is no reason that we can’t do that with this project, too.
2 Likes
I’d also add a background question on the domain, mapping to Elasticsearch’s original domains:
- Enterprise search
- Observability
- Security
Wow - this blew up!
@dawnfoster @ralph @horovits There is a lot of good stuff here. I’ll respond more specifically to some messages but (in a general sense) do we think it would be better to do multiple surveys or one big survey? Survey fatigue is real and can come from frequency or survey length, so it’s really an open question.
My initial thought is we should ask focused questions on a single topic. @dawnfoster’s first reply is a good example.
@ralph and @horovits questions are an amazing start for a more general survey which I think we should absolutely ask. My only questions is if they belong with the other survey which is more focused.
I like it, I have some small feedback
Should this question be about intent vs only current status? So, doing something like “I am (or intend to be) a:”
I have two nits with this:
- ‘employer’ makes an assumption of an employment relationship. You could be a contractor or independent consultant, for example. For sure, we have folks in the community that fall into this category.
- Some people may not have it as a job assignment specifically, but it’s adjacent so it falls under their duties. I’m not sure they would fall into this binary.
Maybe a little more generally:
Is participation in this community as part of your job duties?
- Yes
- No
I’m debating if we should have an ‘other’ option… (slack? wiki? I dunno). Thoughts?
Nothing is set in stone since it’s open source, but I think this is being organized for the website already. I’m not sure I would ask this one since it’s unlikely to change at least the next version of the website.
For this particular topic, I also think that we should name it “Contributor” survey, instead of “Community” survey, since community tends to mean wildly different things to different people
The most common application of this type of software would be supporting a service (not SaaS), so I’m not sure that is captured in these options.
I would swap ‘productive’ to ‘production’ and change ‘stack deployed’ to ‘deployment’ (the word ‘stack’ in this context is confusing - it’s associated with the specific ELK stack which is not always the case).
Also, many folks have many deployments! Perhaps we could capture this?
This question is great! 
Re: Open ended questions. These are tricky. You’re really likely to influence the response by how you write the question. I’m not against them, but I gathering the quantifiable information is first priority then we can dig deeper to get the qualitative second, but maybe someone else know better?
Do we need an ‘other’ category?
[ off-topic ]
(also @dawnfoster) I hear you on this. We have a fair number of people in Europe that show up to the community meetings and it frankly surprises me. I don’t like working at 7pm either. Back around the new year we did a survey on date/times for the community meetings and a majority voted to keep it the same 
.
That being said, I might actually have a more European friendly meeting next time since we have a series of conflicts for the typical time. If this works better, I’m very welcome to the idea of alternating timezone friendlyness.
Here are the final results of my unofficial Twitter poll. While most people look at the website to find common information across repos, a lot of people > 20% look for a community repo.
That is interesting. I think we move forward with gathering the data of this community and see what they think. If we have a corroborating data with your twitter poll for the website, then we’ll have a very clear direction.
With regards to the 21.1% - I wonder if they would find it on the project website on second look. 
They would definitely find it eventually. I tend to err on the side of making it as easy as possible to find things like meetings, governance, and other key info. Removing as many barriers as possible will help us build a bigger and stronger contributor community over time.