Around 2000 I started work at Thomson (now Thomson Reuters) in a small team attempting to transform legal SGML into websites. Thomson always felt massive. There were offices all over the world though I only properly visited the head office in Minneapolis. Which was also massive. So big it had a minibus to bring people in from the outskirts of the carpark. And a data centre run by robots built at the end of Minneapolis airport and, we were assured, completely plane proof. Though not building it at the end of a runway might have proved cheaper. And a train that brought in logs and shipped out legal documents. Which was cool.

Thomson was a traditional company with a traditional hierarchy of upper, middle and lower managers. Feedback fed up, orders cascaded down. At the top of the tree was Lord Thomson whose main job appeared to be deciding what to buy and what to sell. Which meant that most of our jobs became about integrating data from companies he'd bought and attempting to prise apart data from companies he'd decided to sell.

It was an ok place with a nice team and occasionally interesting work in a dreadful building (most of my working life has been spent in dreadful buildings). But we never really met other people either inside the organisation or out. There was an interesting trip to Maastricht that proved much lovelier than I'd expected (though next to Swiss Cottage and Minneapolis most places are lovely). But we didn't really get out and meet the people using the things we made and (at the time) there was no discernible sign of a community around legal data.

That said, there were informal networks for corporate gossip. Though in pre-Twitter days that mostly involved hanging out with PAs on cigarette breaks. Which is still probably the best way to find out what's going on. It's true - smokers know more.

I think I joined the BBC in 2005. It could have been 2006. Or somewhere around then.

It also felt massive. Bigger somehow because the workforce were more concentrated. The work was roughly as boring as Thomson. Someone from "the business" had decided that a thing should be done, talked to a project manager who'd made a chart and eventually the work ended up on your desk. Although, unlike Thomson, where the clear motive was profit, BBC work seemed a little more... random. My first job was to rebrand Jazz Fest 2005 (which had been red) to JazzFest 2006 (which was supposed to be green) which wasn't the best job to give to a colour blind person. Most of the other work arrived via a "task manager". It was largely making a new banner or tweaking the width of a spacer gif. You had no idea why you were doing it or who might be interested but avoiding work cost more effort than just doing it. And it seemed to make the managerial chain of command happy.

About 6 months into my BBC life I started to realise that the BBC didn't work how the BBC management or BBC HR thought it worked. Maybe most organisations don't. There were other lines of affinity running through the BBC that looked almost nothing like the formal org charts that were "cascaded" down every six months or so. Conversations happened outside of meetings and outside of formal reporting lines and that's how things got done. The more dense the network, the more chance you had of getting to work on the interesting stuff. And eventually all lines seemed to lead to Tony, Bill and Tom.

The other part of BBC life that was hard to escape was the perpetual cycle of re-orgs you were forced to live through. None of which seemed to have any material impact on the work you did. There was a fairly constant flux of centralisation and decentralisation that flowed in and out like the tides. One minute you'd be re-orged into a central IT function to better erm "leverage" "platforms" and stop 6 different teams from creating the same thing. Six months later they'd realise that the people working for a "division" had no clue what that division did and you'd be moved back to where the domain knowledge lived. And back round again. Maybe hierarchical org charts and re-orgs make it easier to divvy up the budgets. Maybe the managers just wanted a tidier view of what they think they thought everyone was up to. But conversations are orthogonal to org charts so no amount of deckchair tidying really makes a difference other than pushing up the budget for rebranded mouse mats and mugs. Though you do question what they learn at management school.

I've often wondered what the social network of an organisation would look like alongside a formal org chart. And whether the further they diverge and the more blurred the boundaries of inside and outside, the more productive and creative an organisation gets to be. People are social animals and work best in conversations. Sadly managers tend to not see the graphs for the trees. I guess there are people looking at this?

Around the time I arrived at the BBC, Backstage started. It was an attempt to get people to make things with BBC content and data (before anyone quite realised the BBC owned almost no content and had no data). It opened up some feeds and organised events but more importantly it had a mailing list. That anyone could join. I met more people outside the organisation through Backstage than any other route. And I met more people inside the BBC through Backstage than any other route. Probably the majority of my Twitter acquaintances came through Backstage.

Having the mailing list opened up the BBC (or at least the computer part of it) to licence fee payers (or at least the computer parts of them). But it also opened up new conversations in the BBC. The lines of who was inside and who was outside blurred and didn't really matter so long as people were talking and helping each other. Which mostly they were.

Some of the people who started off hacking BBC stuff on the "outside" ended up on the "inside". Which given the team already knew them was a better route to success than sitting through another round of, "can you tell us about a time you worked with a difficult person" with a bored looking bod you'd borrowed from HR. Some of the people on the "inside" headed off into the wider world. But none of it really mattered because the same people kept on talking and the knowledge of why you'd done a certain thing a certain way didn't dissipate when people left. I still try to get involved in conversations about my old BBC R&D projects. Or at least the interesting bits. Because you never really leave it all behind. And if anyone wants to know about programme URLs....

Maybe this is part of why Microsoft software still feels slightly wonky (to me at least). It's designed for the hierarchies of an industrial management age (that prominent 'send to manager' feature at the top of Outlook on a PC). Maybe because that's how Microsoft choose to work. Even as they move to the cloud, they never quite move to the web. Which might be fine if you only ever need to collaborate with people inside the organisation but that's not really how I'd choose to work.

Towards the end the Backstage mailing list tended to get bogged down in discussions about the BBC's adoption of DRM which all got a little tedious. But it's probably still a conversation a public service broadcaster should be having with itself.

In 2010 Backstage was shut down. The BBC seemed to turn in on itself and become a more claustrophobic place to work. Though it also coincided with the arrival of a bunch of managerial blokes from Microsoft who talked excitedly about "shipping excellent product". I still have no idea what any of that meant.

All of this rant was occasioned by conversations at my new place of work about how we better engage with and support the communit(y|ies) around open parliamentary data. I was struck by the number of interested people who turned up at the Newspeak House event we ran. All of them happy to get involved and help. And also struck by the collection of clerks and academics who turned up at the Study of Parliament Group. The two groups couldn't have looked less similar (tshirts and stickers vs M&S pullovers) but the shared interests were remarkable. And probably most of the people at one have never met, spoken to or known the existence of most of the people at the other.

There should be some way to brings these various communities together. The internal library people and statisticians, the external civic tech, academics and journalists etc. "Creating a community" definitely doesn't feel right because the communities already exist quite happily without our help. So it's less about creating a new graph and more about increasing the link density between existing graphs of people.

Obviously, open by default is the new fashionable, but I'm pretty sure it's no longer enough for organisations to be "open" because no matter how hard you try, eventually that just becomes PR and marketing fluff. Much better to aim to be porous. For people and conversations and ideas to flow in and out irrespective of departmental and organisational boundaries. Rip down the barriers to conversation and collaboration internally and externally. When you're trying to get a thing done or hunt out advice it doesn't really matter who's paying the person you're working with. You don't become a different person when you walk through a different office door and you don't stop talking to the people you usually talk to.

So maybe this is just a pitch to create a mailing list and run a few more events. That would be super.