[ckan-dev] The unofficial CKAN survey results

Adrià Mercader adria.mercader at okfn.org
Tue Aug 7 15:45:49 UTC 2018


Thanks to all people who spent some time adding their thoughts and of
course thanks Ross for taking on this!

This snapshot of what the community thinks is super valuable. I'd like to
take the opportunity to encourage everyone to have a say in any discussion
regarding the project, don't feel like your opinion is not important.
Whether you are a developer, user or maintainer it's important to get as
many inputs as possible, specially around things like deprecating features,
barriers to contribution or the upcoming discussion about 3.0.


There's a lot to unpack there so I'll have a better read and share my
thoughts soon.

Thanks again!

Adrià


On 6 August 2018 at 18:37, Ted Strauss <ted.strauss at gmail.com> wrote:

> Amazing job on preparing just the right survey and summarizing the results
> for actual action.
>
> Here is my quick take on the results:
> - If "Search, Discovery and metadata management" has the most hits, with
> 3, then that's an important result of the survey deserving recognition.
> It's easy to take for granted that people are interested in the core
> functionality of a software. I think it's reminder to keep things simple,
> and notice if the other questions can be connected to that. For example,
> which of the frustration points relate to stabilizing and maturing the core
> metadata features? Give those a bump.
>
> - Most of the frustration points are the kinds of issues that any software
> project faces, and have no easy solutions. The ongoing conundrum is whether
> to (a) grit teeth and carry on, or (b) completely scrap and rebuild some
> parts of the application. JKAN is an amazing example of how (b) can be
> taken to it's logical conclusion (I had no idea this existed). "Make CKAN
> core *just* be the API" is a more common approach to (b). The point is,
> this is the most difficult decision facing any project and cannot be
> rushed, for technical reasons, and also because the decision could alienate
> contributors who see their hard work thrown out.
>
> Ted
>>
> On Sat, Aug 4, 2018 at 6:42 AM Ross Jones <ross at mailbolt.com> wrote:
>
>> Hello.
>>
>> I thought I should release the results of the Unofficial CKAN Contributor
>> survey so that everyone can see where we are. I've obscured and reworded
>> all of the responses so none of the following is in the submitters
>> own words.
>>
>> I don't believe I've mis-represented anyone's feedback and
>> hopefully some of these points will spur on further discussion (maybe
>> in separate threads?). If you think I have mis-interpreted your answer
>> or have missed it entirely, please ping me off-list and I'll find some
>> way
>> to correct it.
>>
>> As of today, there have been 29 responses - which is great!
>> Thank you for contributing - it might look like things are in a bad place,
>> but I'm confident we can make things better.
>>
>> There's quite a lot here, you may want to read it whilst relaxing with
>> a cup of tea (or relaxing beverage of your choice)
>>
>> ########################################################
>>
>> # Have you ever contributed to CKAN? (29 responses)
>>
>> * 1 person tried (and presumably did not succeed :( - please ping me
>> offlist
>>     and I'll see if I can help)
>> * 7 people have not (hopefully qualified with 'yet' but who knows)
>> * 21 people have contributed.
>>
>>
>> # Contributor - I am paid to contribute (21 responses)
>>
>> * The breakdown here is 11 are paid, 10 aren't paid.
>>
>>
>> # Contributor - On average I work on CKAN core ... (21 responses)
>>
>> * Only 3 people work on core daily
>> * 6 a couple of times a month
>> * 8 very occasionally
>> * 1 response each for the other options
>>     - No code committed
>>     - A couple of times a week
>>     - Rarely if ever, but would like too
>>     - Rarely, mostly outside core
>>
>>
>> # Contributor - Which areas of CKAN interest you the most (20 responses)
>>
>> Some responses mentioned 2 or more areas, so these won't add up.
>>
>> * Data visualisation surprisingly only appeared once
>> * Geospatial features got mentioned twice
>> * The API was mentioned twice in responses
>> * Extensibility made it into the list once
>> * Search, Discovery and metadata management was mentioned three times!
>> * Contributing to something for the public good
>> * The datastore got mentioned twice
>> * Scalability (ftw)
>> * Deployment
>>
>>
>> # Contributor - What areas frustrate you the most? (20 responses)
>>
>> This was much more polite than I thought it was going to be :) A few
>> things appear more than once.  I've paraphrased (but hopefully caught
>> the gist) of each response and broken them down where there is more
>> than one point made
>>
>> * It isn't pleasant to work with.
>> * Logic layer (this is me, I'm not sure what I meant by this, I do think
>>                overall that it's a good thing)
>> * Documentation
>> * Deployment story
>> * There is too much code and it does too much
>> * There is a specific sqlalchemy 'feature/bug' which keeps biting me
>> * There are too many features
>> * I don't know
>> * Deployment story
>> * Quality of some plugins
>> * Pylons (this is me too, I am sooo happy this is being fixed)
>> * No easy way to tell how good a plugin is until you try it
>> * Deployment story
>> * Things that need to change together don't live close together in
>>   the codebase
>> * Plugins break too easily when core changes
>> * Speed of development (again, me. I know why this is, and this is
>>   why we are having an unofficial survey so we can work out how to
>>   increase developer happiness even if we can't increase the number
>>   of people contributing).
>> * Dependency hell (not me, but I wholeheartedly agree)
>>
>>
>> # Contributor - If you could make one change? (20 responses)
>>
>> Again, some with more than one point (which is technically cheating for
>> this question - you should feel bad :P ) - so I've split them out.
>>
>> * Make CKAN core *just* be the API, everything else as 'separate'
>>   apps/repos (okay, okay, I'm sorry - busted).
>> * More people paid to work on CKAN full-time
>> * Why not just use a document store given it's how we handle the data
>> * Infrastructure as code (+ containers) to make deployment much easier
>> * Start again, build on top of JKAN instead ( /me *glares at submitter*)
>> * Can't say, I could just submit a PR couldn't I?
>> * Coordinating the work people are doing around a roadmap
>> * More contributions from those with a vested interest in CKAN's success
>> * More people paid to work on CKAN full-time (again) but funded from
>>   'outside'.
>> * "Why isn't there a tetris game hidden in the data preview?"
>>   (This is an awesome question. I __will__ prioritise review of your PR)
>> * Decouple those areas of the data model that are too tightly bound
>> * Accessibility! and improve the UI
>> * Make data preview betterer
>> * Django! Django! Django!
>> * Improved data management practices (DOIs, versions, etc)
>> * Proper project oversight which improved sustainability
>> * Python 3
>> * Make ckanext-scheming _even_ better/easier
>> * Split the frontend from the backend (I promise this wasn't me again)
>>
>>
>> # Non-contributor - I haven't contributed because (8 responses)
>>
>> * Someone does it for me
>> * I answered this in the next question
>> * I don't have a lot of time and without direction for my efforts
>>   it doesn't feel productive
>> * I have a special use for CKAN and finding company time to make this
>>   useful for others is difficult
>> * Didn't realise you need help!
>> * Haven't really used it yet
>> * Still trying to get it installed
>> * I never got response to my contributions (this is bad ... please ping
>>   me offlist and I'll take a look)
>>
>>
>> # Non-contributor - Are you interested in contributing to CKAN (8
>> responses)
>>
>> * 37% say "Yes"
>> * 50% say "Maybe"
>> * The rest say "No"
>>
>>
>> # Non-contributor - What would make it easier for you to contribute (6
>> responses)
>>
>> As usual, responses split ...
>>
>> * Being able to install it
>> * Better understanding of it
>> * A project that needs it
>> * A clearer community focus and a product roadmap
>> * A guide on how to get started
>> * Nada. No time.
>> * More up to date and better structured documentation
>> * A more 'open' process
>>
>>
>> # Non-contributor - What areas of CKAN interest you the most? (6
>> responses)
>>
>> * The whole project
>> * Documentation
>> * The code
>> * Non-manual metadata creation
>> * All the docs (for users and technical)
>> * It's place in the ecosystem is v. positive
>>
>> Fin.
>>
>> #########################################################
>>
>> As I mentioned above, it'd be great if some of these points got taken up
>> in
>> other threads, but I'd definitely encourage everyone to contribute to any
>> over-riding themes in response to this email.
>>
>> For me, no real surprises - except whoever suggested JKAN ;) I think we
>> all
>> know that the deployment story isn't great, although there are some
>> attempts
>> to improve that happening right now.  We all also know there are some dark
>> dusty corners of CKAN that need some TLC.
>>
>> The hardest thing, is for those wanting to contribute, how do we make
>> 'onboarding' people easier, given the relatively small number of regular
>> contributors, many of whom are doing their work for free.  There are
>> clearly some people interested, for instance, in helping with
>> documentation
>> but this would need time from someone to share their knowledge of CKAN.
>>
>> Whoever posted comments responses questioning where the roadmap is, and
>> where are the stakeholders contributing - I suspect we (as a community)
>> are struggling as the steering group is made up of the busiest people
>> around who expressed an interest - I know they do contribute, but I fear
>> they generally don't have enough time to lead. As far as I am aware,
>> currently, the roadmap is defined by whoever turns up to the technical
>> meetings (which are regular and lined to from
>> https://ckan.org/about/technical-team/). You should attend these (and
>> I should make more time to attend as well ☹)
>>
>> I think it would be amazing if a new steering group could form out of
>> the community, composed of people/companies that had an interest in
>> helping define the roadmap, help find funding for specific pieces of
>> work and help build some cohesion and clearer collaboration into the
>> project.
>>
>> Anyway, enough about what *I* think, what do *you* think the results say?
>>
>> Ross
>>
>> _______________________________________________
>> ckan-dev mailing list
>> ckan-dev at lists.okfn.org
>> https://lists.okfn.org/mailman/listinfo/ckan-dev
>> Unsubscribe: https://lists.okfn.org/mailman/options/ckan-dev
>>
>
> _______________________________________________
> ckan-dev mailing list
> ckan-dev at lists.okfn.org
> https://lists.okfn.org/mailman/listinfo/ckan-dev
> Unsubscribe: https://lists.okfn.org/mailman/options/ckan-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/ckan-dev/attachments/20180807/c4c5430c/attachment-0002.html>


More information about the ckan-dev mailing list