You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@diversity.apache.org by Niclas Hedhman <ni...@hedhman.org> on 2019/07/19 00:17:40 UTC
Open source as employment?
Something struck me yesterday, as I was reading a post (could have been an
old one) from someone who claimed that 50% of contributors work on ASF
projects as part of their $day job. Of course without a citation, but the
commit stats on GitHub, as well as volume of email, seems to corroborate on
that, we have more commits and email during weekdays.
Now, assuming that is a reasonable approximation...
Many companies in USA are heavily vested in D&I efforts, and shouldn't it
therefor be of our interest to look at whether these same companies (those
that are proactive, such as Google) are assigning under-represented groups
to work on open source projects, whether they match their hiring ratios and
if there is a difference, ask why is that? If they are over-represented
there, what can other companies learn, and if under-represented, what can
ASF do about it? How does one even get a job to work on open source in
these giants?
Wouldn't the D&I departments of these companies be interested in sharing
these data? And wouldn't we then be interested in cross-referencing their
data with their share in ASF?
I think I saw somewhere (her own introduction?) that Gris was employed by
Google, so I assume it would be really easy for her to check internally at
work on this, as a starting point. If the data exist, quick answers, if
data don't exist, maybe see if it can be produced...
Reaching out to others, both inside and outside the ASF, in similar
companies would be the next step.
Cheers
Niclas
Re: Open source as employment?
Posted by Justin Mclean <ju...@classsoftware.com>.
Hi,
> Unfortunately, such calculations don't help if there is statistical bias[2] involved.
Sure 100% agree. Hopefully the new survey will take that into account or at least explain what they might be.
BTW tis he analysis of all committer names gave the same percentage for gender and ethnicity as that survey (+- a couple of %), which I think add some weight that it’s representative at least on those two issues.
Thanks,
Justin
Re: Open source as employment?
Posted by Ross Gardler <Ro...@microsoft.com.INVALID>.
Speaking with my Microsoft Hat on, but not in an official capacity...
In summary MS employees come from all over the world already. We are working hard on diversity within our workforce. All of those employees are encouraged to contribute upstream in order to enable 1st and 3rd party products. Some are required to do so, but they are a minority. If the ASF is not ready and able to work with the diversity found in a company like MS then there will be missed opportunities in ASF projects regardless of any efforts inside those companies.
In more detail...
Need to think about two grips of staff 1) product groups who build solutions both in house and with partners. 2) field teams who deliver product, with partners, to customers (pre and post sales).
Employees here in the product groups of MS are not employed to work on external open source, our contributions are a by-product of good software engineering practice (there are, of course, a few exceptions in which specific individuals are employed because of their roles within key open source projects). People are employed to work on a product and the product team as a whole will decide how to build it. Very often that includes contributions to open source, but people are rewarded for the success of their product not the success of the open source components within it. Good engineers will contribute to and rise open source when delivering product.
As a multinational we have engineering centers across the globe. The core is in Redmond, Washington, US. We are working very hard to improve our diversity numbers and have made some good progress in recent years - still a very long way to go though. This effort is independent of open source activities.
ASIDE: When MS is the "owner" of the open source there are plenty of people employed to work on the open source, e.g.. Net, visual studio code, typescript etc. But that's not going to help answer the original question.
As Kenn suggests employees decide which projects they want to work on. This means some employees seek projects where open source opportunities are strongest, but they are self selecting. A lack of diversity in open source may be leading to a lack of diversity in our open source focused employees, or maybe the opposite is true. Most likely both.
We do partner a great deal in delivery of product into the market. Often this involves collaboration in open source projects. In these cases we can have both direct and indirect contributions to the open source components. That is sometimes we work with the partner and they take that code and push upstream as appropriate for their strategy, other times we work directly on the open source to unlock a partner. In these cases we tend to be very much driven by the needs of our shared customers and thus the engineering talent will often be where the customer /partner is, supported by our product groups. In the case of MS this could be anywhere in the world. So, again, a lack of diversity inhibits contribution.
Finally. There are our technical field teams and Systems Integrator partners. These folks deal directly with customers solving very specific problems. Engineering is always where the customer is and thus could be anywhere in the world and thus impacted by diversity issues.
In my opinion, MS has done a really good job of enabling every one of our employees to contribute to open source (especially when one considers the history of the company). There has been massive cultural shift to enable this. But it's about empowering all individuals to contribute as they desire. That is, I'm not aware of D&I initiatives that target open source contribution. All our employees are expected to contribute as they need to.
What this means, from an ASF perspective, is that we need to ensure anyone, anywhere in the world, can easily contribute. Any friction in the community will result in missed contributions of value. Worse still, it means missed opportunities to turn contributors into committers through specialization.
This is not new for the ASF, but the ASF no longer lives in a technology world dominated by startups mostly in the US.
In summary, MS employees come from all over the world already. We are working hard on diversity within our workforce. All of those employees are encouraged to contribute upstream in order to enable 1st and 3rd party products. Some are required to do so, but they are a minority. If the ASF is not ready and able to work with the diversity found in a company like MS then there will be missed opportunities in ASF projects regardless of any efforts inside those companies.
Ross
---
Sent from my phone, likely while waking down the stars and having a conversation. Sorry about my carelessness, I blame the machines.
________________________________
From: Justin Mclean <ju...@classsoftware.com>
Sent: Friday, July 19, 2019 5:04:22 AM
To: dev@diversity.apache.org <de...@diversity.apache.org>
Subject: Re: Open source as employment?
Hi,
This may also be of use [1][2]. There's a section on employment. In short 70% are employed of those 65% contribute to OS, but less are allowed to contribute at work due to IP concerns.
Thanks,
Justin
1. https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fgithub%2Fopen-source-survey&data=02%7C01%7CRoss.Gardler%40microsoft.com%7Cc80065841cac44cbdbfe08d70bfe3b6c%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C1%7C636991058852817091&sdata=JJ6f8pl8L6co%2FH5RNYXfWUCLMX7ncGJCfN2ze2rHewo%3D&reserved=0
2. https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fopensourcesurvey.org%2F2017%2F&data=02%7C01%7CRoss.Gardler%40microsoft.com%7Cc80065841cac44cbdbfe08d70bfe3b6c%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C1%7C636991058852817091&sdata=v9Y%2FMeTj0tMjIhYPQyOgqHzkrnCBntUzKl7c3VTgcB8%3D&reserved=0
Re: Open source as employment?
Posted by Justin Mclean <ju...@classsoftware.com>.
Hi,
This may also be of use [1][2]. There's a section on employment. In short 70% are employed of those 65% contribute to OS, but less are allowed to contribute at work due to IP concerns.
Thanks,
Justin
1. https://github.com/github/open-source-survey
2. https://opensourcesurvey.org/2017/
Re: Open source as employment?
Posted by Kenneth Knowles <ke...@apache.org>.
In my experience, software engineers are not assigned to projects. They
choose to join a project that needs someone. If the project is rewarding,
they stay for a while. Paid contributors are still voluntary.
Google publishes https://diversity.google/annual-report/. The ASF committer
survey is not as representative, because of the bias and also because it
does not include non-committers. This is not really specific to Google. If
it were possible to compare "diversity of FooCorp employees" versus
"diversity of FooCorp employees regularly contributing to Apache projects"
that might be a useful data point to track over time.
Kenn
On Thu, Jul 18, 2019 at 7:27 PM Niclas Hedhman <ni...@hedhman.org> wrote:
> And I think it doesn't matter for my point... a lot of people has open
> source contribution as a part of their job (I think commit stats clearly
> shows that). So instead of debating whether it is 30% or 90% of the
> contributions coming from employed people, I hope to get comments on my
> point.
>
> Thanks
>
> On Fri, Jul 19, 2019 at 10:23 AM Sam Ruby <ru...@intertwingly.net> wrote:
>
> > On Thu, Jul 18, 2019 at 9:50 PM Justin Mclean <ju...@classsoftware.com>
> > wrote:
> > >
> > > Hi,
> > >
> > > > Ok, cool... so 50% of those that like to reply to, and have time for,
> > > > surveys…
> > >
> > > It probably a little more representative than you may think. While
> there
> > might be a couple of assumptions made, plugging in the numbers in say [1]
> > gives a 95% confidence of 3.3% margin of error or 99% confidence of a
> 4.4%
> > margin of error. But hopefully the new survey will explore things along
> > these lines and give us a better picture.
> >
> > Unfortunately, such calculations don't help if there is statistical
> > bias[2] involved. Note that bias in this case is a technical term and
> > does not necessarily have anything to do with intent.
> >
> > Source: was a math major.
> >
> > > Thanks,
> > > Justin
> > >
> > > 1. https://www.checkmarket.com/sample-size-calculator/
> >
> > 2. https://en.wikipedia.org/wiki/Bias_(statistics)
> >
>
>
> --
> Niclas Hedhman, Software Developer
> http://polygene.apache.org - New Energy for Java
>
Re: Open source as employment?
Posted by Niclas Hedhman <ni...@hedhman.org>.
And I think it doesn't matter for my point... a lot of people has open
source contribution as a part of their job (I think commit stats clearly
shows that). So instead of debating whether it is 30% or 90% of the
contributions coming from employed people, I hope to get comments on my
point.
Thanks
On Fri, Jul 19, 2019 at 10:23 AM Sam Ruby <ru...@intertwingly.net> wrote:
> On Thu, Jul 18, 2019 at 9:50 PM Justin Mclean <ju...@classsoftware.com>
> wrote:
> >
> > Hi,
> >
> > > Ok, cool... so 50% of those that like to reply to, and have time for,
> > > surveys…
> >
> > It probably a little more representative than you may think. While there
> might be a couple of assumptions made, plugging in the numbers in say [1]
> gives a 95% confidence of 3.3% margin of error or 99% confidence of a 4.4%
> margin of error. But hopefully the new survey will explore things along
> these lines and give us a better picture.
>
> Unfortunately, such calculations don't help if there is statistical
> bias[2] involved. Note that bias in this case is a technical term and
> does not necessarily have anything to do with intent.
>
> Source: was a math major.
>
> > Thanks,
> > Justin
> >
> > 1. https://www.checkmarket.com/sample-size-calculator/
>
> 2. https://en.wikipedia.org/wiki/Bias_(statistics)
>
--
Niclas Hedhman, Software Developer
http://polygene.apache.org - New Energy for Java
Re: Open source as employment?
Posted by Sam Ruby <ru...@intertwingly.net>.
On Thu, Jul 18, 2019 at 9:50 PM Justin Mclean <ju...@classsoftware.com> wrote:
>
> Hi,
>
> > Ok, cool... so 50% of those that like to reply to, and have time for,
> > surveys…
>
> It probably a little more representative than you may think. While there might be a couple of assumptions made, plugging in the numbers in say [1] gives a 95% confidence of 3.3% margin of error or 99% confidence of a 4.4% margin of error. But hopefully the new survey will explore things along these lines and give us a better picture.
Unfortunately, such calculations don't help if there is statistical
bias[2] involved. Note that bias in this case is a technical term and
does not necessarily have anything to do with intent.
Source: was a math major.
> Thanks,
> Justin
>
> 1. https://www.checkmarket.com/sample-size-calculator/
2. https://en.wikipedia.org/wiki/Bias_(statistics)
Re: Open source as employment?
Posted by Justin Mclean <ju...@classsoftware.com>.
Hi,
> Ok, cool... so 50% of those that like to reply to, and have time for,
> surveys…
It probably a little more representative than you may think. While there might be a couple of assumptions made, plugging in the numbers in say [1] gives a 95% confidence of 3.3% margin of error or 99% confidence of a 4.4% margin of error. But hopefully the new survey will explore things along these lines and give us a better picture.
Thanks,
Justin
1. https://www.checkmarket.com/sample-size-calculator/
Re: Open source as employment?
Posted by Niclas Hedhman <ni...@hedhman.org>.
Ok, cool... so 50% of those that like to reply to, and have time for,
surveys... The exact numbers aren't important, but I think it is agreeable
that the number is quite high.
On Fri, Jul 19, 2019 at 8:30 AM Justin Mclean <ju...@classsoftware.com>
wrote:
>
> Hi,
>
> > Something struck me yesterday, as I was reading a post (could have been
> an
> > old one) from someone who claimed that 50% of contributors work on ASF
> > projects as part of their $day job.
>
> See Question 3 from the committer survey [1]
>
> Thanks,
> Justin
>
> 1.
> https://cwiki.apache.org/confluence/display/COMDEV/ASF+Committer+Diversity+Survey+-+2016
>
>
>
--
Niclas Hedhman, Software Developer
http://polygene.apache.org - New Energy for Java
Re: Open source as employment?
Posted by Justin Mclean <ju...@classsoftware.com>.
Hi,
> Something struck me yesterday, as I was reading a post (could have been an
> old one) from someone who claimed that 50% of contributors work on ASF
> projects as part of their $day job.
See Question 3 from the committer survey [1]
Thanks,
Justin
1. https://cwiki.apache.org/confluence/display/COMDEV/ASF+Committer+Diversity+Survey+-+2016