Do you remember enterprise application integration? Those were the days. First you paid to bury your information in someone’s proprietary silo, then you paid to excavate it from there, then you paid again to bury it again in someone else’s silo. Everybody was happy. Except for the guys paying the bills.
I went to see the guys in Osmosoft yesterday, it’s always a pleasure visiting them. At BT Design, our approach to innovation has a significant community focus: Web21C, now integrated into Ribbit, was formed on that basis; both Osmosoft as well as Ribbit are excellent examples of what can be done with open multisided platforms.
While I was there, I spent some time with Jeremy Ruston who founded the firm and leads the team. Incidentally, it was good to see Blaine Cook there, I hadn’t seen him since he joined BT. Welcome to the team, Blaine.
When it comes to opensource, Jeremy’s one of the finest brains I know, we’re really privileged to have him. We got to talking, and somehow or the other, one of the topics that came up was the ways and means we have to figure out if someone’s any good, in the context of hiring. After all, there is no strategy in the world that can beat the one that begins “First hire good people”.
When you’re hiring people with experience, the best information used to come from people you knew who’d already worked with her or him. Nothing beats a good recommendation from a trusted domain. You can do all the interviews you want, run all the tests you can find, do all the background searching you feel like; over time, the trusted domain recommendation trumps the rest.
Now obviously this does not work when the person has not worked before, where there is no possibility of a trusted domain recommendation. Which is why people still use tests and interviews and background checks.
Which brings me to the point of this post. Jeremy brought up an issue that he’d spoken to me about quite some time ago, something I’m quite keen on: the use of subversion commit logs as a way of figuring out how good someone is.
And that got me thinking. Here we are, in a world where people are being told: Don’t be silly and record what you do in Facebook; don’t tell people everything you do via Twitter; don’t this; don’t that; after all, the bogeyman will come and get you, all these “facts” about your life will come back to haunt you.
As a counterpoint to this, we have the opensource community approach. Do tell everyone precisely what you are doing, record it in logs that everyone can see. Make sure that the logs are available in perpetuity. After all, how else will people find out how good you are?
Transparency can and should be a good thing. Abundant transparency can and should be a better thing, rather than scarce transparency. Right now we have a lot of scarce transparency; people can find out things about you, but only some people. Which would be fine, if you could choose who the people were. Do you have any idea who can access your credit rating? Your academic records? Do you have any idea who decided that?
Scarce information of this sort leads to secrets and lies and keeps whole industries occupied. Maybe we need to understand more about how the opensource community works. Which, incidentally, is one of the reasons why BT chose to champion Osmosoft.
An aside: David Cushman, whom I’d known electronically for a while, tweeted the likelihood of his being near the new Osmosoft offices around the time of my visit, so it made sense to connect up with him as well. It was good to meet him, and it reminded me of something I tweeted a few days ago. How things change. In the old days relationships began face to face and over time moved into remote and virtual and electronic. Nowadays that process has been reversed. Quite often, you’ve known someone electronically for a while, then you get to meet them. Intriguing.
Finally, my thanks to gapingvoid for the illustration, which I vaguely remembered as “Excavation 47”. It was a strange title so it stuck. Which reminds me, I have to start saving up to buy one of his lithographs, they’re must-haves.
Have you read David Brin’s The Transparent Society? In the same vein, though more about how many eyes will lessen the negative effects of government surveillance. If everyone has a camera, many eyes go to work.
The recent NY city policeman incident shows the tip of the iceberg:
http://news.cnet.com/8301-1023_3-10125084-93.html
One hurdle to overcome is the understanding and maturity of those reading the twitter/facebook etc.
I think for people to be able to share that level of information and link it to their work persona will require an understanding that not everything will be positive, that references to failure are not a bad thing and that it is acceptable to express negative emotions, albeit in a positive way.
These are the things I see the Osmosoft guys, and others, doing and it is wonderful to see leaders like Jeremy and you encouraging this.
When I think of my own industry (Management consulting) I can see we have a long way to go to get to the stage where this is possible.
Speaking of using svn commit logs to see what folks do in OSS – have you seen Ohloh.net? Its kinda great and kinda scary at the same time :)
e.g. here’s some of what I’ve been up to over the years…
https://www.ohloh.net/accounts/jstrachan?ref=Detailed
Its a pretty good service – lets you see who’s contributed to what and so forth. It doesn’t handle CVS/svn migration or svn relocation so often cuts off years of OSS contributions – but its about the best I’ve seen so far.
I guess we could do with some aggregation of all this with JIRA / forum / mailing list & wiki activity maybe – as there’s more to OSS than just svn commits
record it in logs that everyone can see. Make sure that the logs are available in perpetuity. After all, how else will people find out how good you are?
I am not sure why you pick out commit logs for special mention. Perhaps you are selecting for a specific trait when you talk about figuring out how good a person is by looking at the logs. Encouraging developers to acquire a good taste for commit logs is no different from encouraging a good commenting style, or even coding style, and a programmer’s worth is a function of how well she does all of those things (… and more)
Commit logs are not a way to show off how good you are. Sharing the right quantity and quality of information in the commit log increases productivity of the team – the motives behind revealing more information on social sites, on the other hand, are quite removed from that, I suspect.
To be able to truly judge the quality of a log one should be familiar with the code base (but not necessarily the commit in question, as pointed out), in which case you have much more to go by – the difficulty of the bug that was fixed, the elegance of the code fix, the style and quality of the comments – all of which are also recorded for posterity in the commits along with the logs. If you are unfamiliar with the code, you will tend to have bias for verbose logs, which will skew your analysis anyways.
Perhaps I am reading too much into your two sentence observation of the issue, but I would be delighted to hear Jeremy’s views as well as why you found them interesting.
On the issue of reference from trusted sources – LinkedIn, I think, presents a very interesting case. While I am yet to see (and I do not expect to see one, to be honest) a poor recommendation on someone’s page, I feel it would be extremely unlikely that a person would put substantially false claims on his public page – even without any explicit recommendations. A candidate is far more likely to distort facts in a resume submitted in private, rather than out in the open on a system like LinkedIn. I call this the Implicit Recommendation of the Cloud (IRoC).
I’ll stop writing “oops. silly me” into my commit comment from this point on.
Karra – Commit logs are just a part of the residue that an open source developer leaves behind them, and they need care and plenty of context if they are to be useful in making a judgement about the quality of someone’s work. The point for me is how the wealth of information about an open source developer contrasts with the past, where one might have had to make hiring decisions about someone on the basis of what they said in interview, and the reputation of their previous employer.
Jeremy: The point for me is how the wealth of information about an open source developer contrasts with the past
Absolutely. I agree.
Great to meet you too JP. Our chat inspired one of the 2009 predictions I’ve thrown into the mix today. I’m sure you’ll be inspiring much more over the coming years.