Codes of Conduct

90 views
Skip to first unread message

Christopher Pitt

unread,
Mar 28, 2016, 10:47:51 PM3/28/16
to SilverStripe Core Development
I noticed that Helpful Robot has been scoring modules (by a percent or two) according to whether they have a code of conduct or not. This isn't actually part of the supported module standard, but I wanted to discuss it here and see whether we should include this information on addons.silverstripe.org or not.

The way that I think about this is: it doesn't matter whether you are for X or against X (where X can be anything, and recently has been "diversity in tech"). Having a code of conduct (as an extension of getting a higher module score) is about being clear about the ground-rules of your module's community. I think being clear about those ground-rules is important, so I would personally like to see this be a factor in deciding which modules I should use. But that's just my opinion. We could easily factor that score difference out of the scores we can get from that API...

On the other hand, if the general feeling is that we should recommend clear community guidelines, we should maybe consider adding a recommendation for _any_ code of conduct with each supported module.

Thoughts?

Ingo Schommer

unread,
Mar 28, 2016, 11:34:18 PM3/28/16
to SilverStripe Core Development
I'm a bit concerned about the amount of bootstrapping you have to do these days to get a module to comply with this standard.
For example, CONTRIBUTING.md and CHANGELOG.md are already overkill for smaller modules which rarely change and can easily contain that in a single README.md.
As a module author, if you take your work seriously and are asked to make a public statement about your "code of conduct" (through including this file),
you'll also want to understand what it says and if you stand behind those statements. All of which makes it less likely that you'll publish that module because it all takes more time.
If a module author or community member behaves badly in the wider SilverStripe ecosystem, everybody can point to the "canonical" documentation of "code of conduct" anyway.
If inclusion of that file is the difference of a 96% or 97% rating, I don't mind though.

--
You received this message because you are subscribed to the Google Groups "SilverStripe Core Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to silverstripe-d...@googlegroups.com.
To post to this group, send email to silverst...@googlegroups.com.
Visit this group at https://groups.google.com/group/silverstripe-dev.
For more options, visit https://groups.google.com/d/optout.

Damian Mooyman

unread,
Mar 28, 2016, 11:41:09 PM3/28/16
to SilverStripe Core Development
I think as you say, having a code of conduct does clarify to users how they are expected to behave in a certain sphere, and it's no means limiting to a certain pre-defined set of behaviours. Since of course, you are still free to define your code the way you feel is best for your community!

I think that open source is about freedom to create your own communities or areas where you feel comfortable, and I think it's important to respect that "the silverstripe community" is actually a collection of many communities, each of which has their own traits and preferences, or even different cultures and languages.

With that in mind, and the respect I have for those who wish to declare their support of a particular code, I think it's also important to respect the wishes of groups or individuals who are happy with the freedom of not declaring any code at all. If that works for them, I don't feel that we should penalise them for thinking differently to us.

So I guess my vote is for "non-penalised recommendation", and maybe simply point to our own code of conduct as a drop-in, and we continue to add this as a matter of habit (rather than of law) to our own modules under the silverstripe namespace.

The less we demand the more likely people will adopt the standard. Indeed, we should be demanding as little as possible, so as to avoid frustrating module maintainers. :)

Kind regards,

Damian

Sam Minnée

unread,
Mar 28, 2016, 11:42:01 PM3/28/16
to silverst...@googlegroups.com
As a module author, if you take your work seriously and are asked to make a public statement about your "code of conduct" (through including this file),
you'll also want to understand what it says and if you stand behind those statements. All of which makes it less likely that you'll publish that module because it all takes more time.

The positive corollary of this is that the effort is unlikely to be "boilerplate time" and instead it's going to be effort by the author to consider the social implications of their open source project. It seems like this would be a good thing to encourage people to do.

And, for the "boilerplate time", it's a case of accepting the helpful robot PR, which I guess would be saying "SilverStripe's code of conduct is good enough for me".

PS: for those who haven't seen it, our Code of Conduct is here:
 
If inclusion of that file is the difference of a 96% or 97% rating, I don't mind though.

I think that's the intention. It's one of a large number of checks and might have a smaller weighting (e.g. I'd say that "are travis tests running & passing" would be of more material benefit to a module consumer).

As relates to the "Supported Module Standard" we may want to say "These parts are critical" and "These parts are also really good". It can be hard to wade through a long list of unprioritised must-haves.
--
Sam Minnée
CEO
SilverStripe Limited

Hamish Friedlander

unread,
Mar 29, 2016, 12:24:23 AM3/29/16
to silverst...@googlegroups.com
I echo Ingo's concern. 

I think in general having a formal code of conduct is desirable. 

However as a signal strictly of quality, I don't think it adds additional value beyond the hygiene tests we already have.

I'm also uncomfortable saying "you will never get a 100% score, and this other module is better than your module because you have deliberately chosen not to add a code of conduct". That starts to feel too much like activism to me.

Something that might change my mind is: is the score supposed to be a "quality" score or an "ability for external people to contribute" score? The two are correlated but only fairly weakly, and I can see a stronger argument for including having a formal code of conduct in the score if it is the later.

Hamish Friedlander 

Michael Strong

unread,
Mar 29, 2016, 12:28:44 AM3/29/16
to silverst...@googlegroups.com
If I seen a code of conduct in a SilverStripe module I would think that the author had accepted a PR from helpful robot - not that effort had been put in by the author. 

Ironically, the automation part of this de-values its existence in my opinion and so scoring based on it doesn't really make sense. It doesn't tell me which author has put in effort and which author has pressed the green button without reading it just to get the extra points.

I'd (at most) suggest indicating on addons whether a code of conduct exists but not rating a module base on it.


Cam Findlay

unread,
Mar 29, 2016, 5:09:10 PM3/29/16
to SilverStripe Core Development
Would it be workable to have the indicator of the existence of a code of conduct visible and not contribute to the score specifically? 

Christopher Pitt

unread,
Mar 29, 2016, 5:18:30 PM3/29/16
to SilverStripe Core Development

Given the points that are now showing, it appears these are the weightings for each check:


Simon Erkelens

unread,
Mar 30, 2016, 12:51:32 AM3/30/16
to SilverStripe Core Development
Isn't it better to, instead of making a pull-request, make an issue advising to add a CoC?

Cam Findlay

unread,
Apr 4, 2016, 4:49:00 PM4/4/16
to SilverStripe Core Development
Just a note, we rolled out the score feature on addons website. Adjusted to account for the removal of the code of conduct aspect to the score (for now). Only fair since it's not officially in the standard and it's good that Helpful Robot does measure it. We can consume the API and adjust pretty easy for the module.


On Tuesday, 29 March 2016 15:47:51 UTC+13, Christopher Pitt wrote:
Reply all
Reply to author
Forward
0 new messages