Linux and open source technology should be a good news story for everyone. However, the way these topics are presented in the media often leaves enthusiasts unhappy. There is a widespread belief that open source alternatives are neglected in favour of commercial products; that coverage often distorts the facts and exaggerates conflict rather than offering insight; and that the right-wing bias of much Australian media dooms the open source community to being dismissed as a kook minority led by some cult figure from Scandinavia whose name no-one can pronounce. The reality is more complex, as reality usually is.
Note: This is the text of a presentation I gave at this year's Linux.conf.au event. Lead picture by The Giant Vermin.
Ever since I volunteered for this talk, I've had people say to me: "Why are you doing this?" Apparently there's something really odd about a journalist giving a presentation at Linux.conf.au, unless that journalist is Jonathan Corbet from LWN. And actually he did have something to do with this, but we'll get to that in a minute.
The first reason is that I've gotten heaps out of LCA in the eight years I've been attending. I have written an awful lot of stories about this event, and I have been to every one without fail since 2008. So it seemed like it might be time to give something back.
The second reason is that throughout my career as a technology journalist, I've had a steady trickle of people asking me why Linux in particular and open source in general doesn't get more coverage in the IT media, and why they attract the kind of coverage they do get. I don't think the community is totally unrealistic about its expectations.
For instance, I don't think anyone here is expecting to see a picture of Linus Torvalds on the front page of the Herald Sun unless he starts dating the wife of an AFL footballer. But there is often a vague sense that Linux should get covered more often, or that Apple gets covered too much, or that not enough people actually understand the issues surrounding information freedom and access. And all of those points are probably true, but I wanted to explore the reasons why, because I think they're often misunderstood or ignored.
Why admit when you can concede?
But the third reason, and the specific trigger for signing up, was a presentation that Jonathan Corbet gave at LCA 2011 in Brisbane. As many of you will know, Jonathan does an annual review of what's happened in the kernel at each LCA. He did that last year, but early in his presentation he also discussed how the entire kernel development process is often misrepresented in the tech media. One of the things he pointed out was that what got said about the kernel, and by extension about Linux generally, often oversimplified the topic for the sake of drama.
The specific example he gave was an article by Kathryn Edwards from IDG . Kathryn had written a news article summing up Corbet's 2010 presentation at LCA in Wellington about kernel issues. During that presentation, Corbet was asked by an audience member about whether it was too difficult for new contributors to get code into the main Linux kernel. And he said this:
Any development process that takes code from 2700 developers over the course of a year can’t be too exclusive, but it can be intimidating to come into. There’s been a lot of work done to make it easier for people to try and come into our community… Things have improved a lot but there’s a lot further we can go.
Nothing very controversial there. The issue wasn't that he said it or that it got reported, but how it was presented and interpreted in the leading paragraph of the article:
A key Linux kernel contributor has admitted the developer community can be intimidating and hard to break into.
That's a bit over the top, especially the word "admitted". It would make a lot more sense in the context of something that had been repeatedly denied. This is not one of those things. I don't think anyone who has ever looked at the kernel mailing list would argue that it wasn't often rambunctious and argumentative. In that sense, it's not even really news. But the word "admitted" makes you think of a murder trial, not what was actually a very civilised discussion about technology. If someone summed up what I'd said in public that way, I am sure I would be annoyed, just as Jonathan was.
So I was sympathetic, but as soon as I thought that, I had a reality check moment. I suspected that if I Googled my name and the word "admitted", I'd find examples of the exact same thing pretty easily. And unfortunately, I was right.
SearchSecurity: "Scandinavian bank Nordea has successfully deployed card readers to minimise the risk of an attack on its customers, but admits that this couldn't have been justified purely on the basis of cost reductions." (Angus Kidman, May 21 2009)
Oh dear. It turns out I've been even more guilty of playing the same trick with the word "conceded", which performs a similar function:
APC: "Typically, anything above 3% is not good," he conceded. (Angus Kidman, November 10 2008) iTnews: "SAP conceded that the original Duet 1.5 platform, which integrated SAP's back-end systems with Outlook via a series of 11 pre-defined scenarios, had not been a major hit with Australian companies." (Angus Kidman, July 20 2010)
It seems undeniable: us journos like to cast ourselves in the role of prosecuting counsel, wringing out information that would otherwise go concealed. Sometimes we actually have done that, but sometimes we've just sat in a presentation and taken notes and then made the whole thing sound more dramatic than perhaps it was. Picture by Extraordinary Chambers In The Courts Of Cambodia
Anyway, I point out my own guilt up front because it's silly to pretend that I'm just an observer of the media, reporting scientifically. I'm a participant. From that perspective, I hope I can offer some insights into why this kind of reporting that makes people so unhappy sometimes happens, and why stuff the open source community is very keen to see reported doesn't always make the cut.
I also need to emphasise is that this is just one person's view on how the technology media operates, and there's a very Australian bias. I can't claim to speak for the entire IT press, any more than Tridge can claim to speak for the entire Samba community. With that said, I've been working in the industry since 1994, and I've written for pretty much every major local publisher, so hopefully there'll be some vague sense of merit about it all. Anyway, let's get to the first key question I promised to address in my abstract.
Do we even notice?
So: is the IT media biased against open source? Does it get ignored? One way to start measuring that would be to see if technology sites in Australia are actually covering Linux.conf.au. With 500 attendees, it's one of the biggest technology gatherings in Australia, after all. Let's take a quick look at yesterday's front pages from Australia's three most visited technology sites, in terms of audience. Firstly, here's CNET. Nothing. And here's the Sydney Morning Herald's tech page. Nothing. And here's news.com.au's tech stuff. Nothing. You'll have noticed that LCA is not visible anywhere, and I can assure you that I scrolled all the way down to the bottom of each page and didn't find any coverage whatsoever referencing this event.
So in this case it seems there's not so much bias as downright disinterest. But you could point out that those are really big generalist sites with a mass-market consumer focus. Does it get any better with more specialised outlets? Perhaps, but only a little.
My main full-time gig as a journalist these days is as the editor of Lifehacker Australia. And sorry, I'll have a boast-mode moment here: Lifehacker is doing pretty well for itself. In the traffic rankings of Australian technology sites by Nielsen Online Ratings, we sit at number six:
1 CNET Australia 2 smh.com.au - Technology 3 news.com.au - Technology 4 The Age - Technology 5 Gizmodo Australia 6 Lifehacker Australia 7 ZDNet Australia 8 Herald Sun - Technology 9 PC World Australia 10 TechRadar Australia
I'd argue that in that entire top 10 list, we're easily the most open-source-sympathetic site. We regularly review Linux software options and outline Linux tweaks, we discuss open source issues and philosophy, and I come to conferences like this one and BarCamp Geelong and write about them. Lifehacker has a simple and broad mission -- telling people how to do everything better -- and Linux fits very neatly into that mission.
But I still need to ask myself as the editor: how much Linux coverage should we be running? One obvious thing to check -- and something that's much easier with online publishing -- is how many of our readers are Linux users. So I hit Google Analytics and grabbed the numbers. Here are the operating systems that were used by people who visited Lifehacker over the last month:
- Windows 67.46%
- Mac 21.19%
- iPad 3.85%
- Linux 3.23%
- iPhone 1.95%
- Android 1.47%
If you're curious, out of the Linux users, about half of them use Firefox and about a third of them use Chrome. But the overall percentage is really low -- actually a bit lower than I'd have guessed myself.
Obviously there's a chicken-and-egg aspect to the debate: maybe if we ran more Linux content. we'd get more Linux readers. And equally obviously people can run multiple operating systems without using them all for the purposes of reading web sites. But looking at it from a cold, hard publishing perspective, it would be really hard to justify running more a handful of Linux posts based on that readership profile.
As it happens, we've been running a lot more posts than that this week because of the conference, and they have attracted good audiences. But I suspect the same visitor analysis on my competitors, if I had access to that data, would produce even grimmer results in terms of actual Linux users. If a Sydney Morning Herald journalist asked to run a Linux story, they might well have trouble persuading their editor a readership was there on this most basic level. And so that's reason #1 why you don't see a lot of open source coverage: editors believe that there's not that much interest.
Yes, we have no staff
There's another factor which is worth considering when it comes to how Linux.conf.au itself gets covered, and which also plays into how open source gets written about more broadly: there are not enough journalists to go around. Case in point: when I started working as a journalist on PC Week back in 1994, we had 10 editorial staff, and we only had to put out an edition once a week. Picture by Tomi Knuutila
On Lifehacker, I'm essentially the only full-time staff member, and I write eight posts a day, at least one of which will be pretty lengthy. We have a truly awesome night editor who deals with refeeding stories from the US, but I have to share her with my sibling sites Gizmodo and Kotaku. And that pattern is reflected on every title I know of in the local market: there are fewer staff and they're expected to do much more and to keep web sites ticking over with fresh content all day.
In that context, it can be really hard to justify spending a week in glorious Ballarat, since you'll have to do all your regular work as well. So often it's easier not to get involved in the first place. My point is that what looks like conscious neglect is just as likely to be due to sheer exhaustion or lack of capacity. There are only so many stories you can write in a day. So that's reason #2: it's just not possible to cover everything that might be of interest.
Yes, we have no freelancers
One obvious way editors can overcome that issue is by using freelancers. Between 2000 and 2010, I worked as a freelance IT writer for pretty much anyone who would have me, and this is actually the first LCA I've attended as a salaried staffer rather than as a full-time freelancer. Assuming the editor has a freelance budget -- and that's actually a big assumption -- they might persuade someone like me or Stilgherrian to come along and write up any interesting stories. Picture by Maya83
Sometimes this works really well as a model. When I went to LCA 2010 in Wellington, I was the only freelancer there, and I set a personal record by writing stories about the event for six different publications. That meant I could find angles that worked for each of those titles, and they didn't lose any staff tracking them down.
But sometimes this doesn't work really well as a model. The main reason is a simple one: basic economics. No-one becomes a journalist to get rich, and writing about technology is a difficult area to make money in. And the biggest difficult is this: on virtually every Australian tech title, the freelance rates are either the same or lower than what they were 20 years ago. Apparently publishers haven't heard of inflation. As well, there are thousands of people offering excellent coverage on blogs purely because they love the subject area. There are thousands more rehashing other people's coverage for Google AdWords revenue. So you have to do more and more work to generate the same amount of money.
The other big issue with being a freelancer is that you have to convince an editor that the story is worthwhile for their particular audience, If you're lucky, you'll get asked to file a given number of stories from an event and you can pick the angles yourself. But more often than not, you'll have to come up with a particular angle and then convince the editor that it's worthwhile. And that involves at least two more challenges.
The first is that your editor really might not know much about the issues involved. That shouldn't be the case for anyone who has been around tech for a while, but you can never be sure. It can be difficult to get someone who isn't already versed with the issues to really understand the difference between "free as in beer" and "free as in speech". Media types like both options, but they don't necessarily want to discuss them in detail. But at least you can overcome that if you have time to explain the whole issue to the editor. Of course, again, you're not really getting paid enough to do that, but it's how the job goes.
The second issue is harder to get around, and plays back into the broader issue of how open source is covered by the IT media as a whole. There are fashions in what gets covered. If you're at a conference covering something that's deemed to be popular, it'll be easier to sell the story. But the problem with being in fashion is that eventually you go out of fashion.
Are we driven by fashion?
When I first covered LCA back in 2004 in Adelaide, stories about Linux were actually very much in demand. Indeed, many editors operated on a simple but fairly effective principle: if they ran enough stories about Linux, some of them would get picked up by Slashdot and traffic would go up massively. There was so much interest in the topic that I attended that conference with a commitment from two separate rival sites to take whatever stories I could come up with. I'm not so sure that would happen now. Picture by nycstreets
Let me give you another example of how Linux was once seen as a big money spinner. Back when I worked for Australian Consolidated Press around the year 2000, we had great success with the Pocketbooks series, which included a CD for Linux installation along with detailed instructions on how to install and use it. We sold around 100,000 copies of that, and at least two other rival publishing houses produced similar titles. But that title isn't in production any more. Things change.
The go-to technology topic which everyone wants to write about these days is Apple. Facebook comes a close second for the mainstream tech titles, but across the board, writing about Apple, whatever the angle, consistently brings in the readers. And again I'm guilty of doing this. Check how I wrote up Bruce Perens' opening LCA keynote earlier this week for Lifehacker. But in fact Perens made much the same point about open source's visibility in his keynote:
We are seeing some signs that Linux and open source have peaked. The locked-down platform is beating us in many ways today.
I don't think you can argue that change is solely due to the media covering open source with less traffic-chasing fervour. But equally I suspect that has played a part. So that's reason #3: it's just not a fashionable topic. The upside to that is that fashions cycle around.
Another minor issue when it comes to covering open source is this: because everything is developed collaboratively, it can be hard to get a definitive statement on anything. Journalists are always under pressure to find someone to comment on issues when they write stories. That can be challenging if the project has dozens or hundreds of developers, and the only way to get hold of them is via email. It's not that there's anything wrong with the process; it's just sometimes not as easy as going through the more typical corporate process. So perhaps that's reason #4 for some journalists: the whole process of writing about open source takes them too far out of their comfort zone. That, though, is something they should be able to get over.
Of course, if someone is perceived as a spokesperson or a leader , that can have its downsides too. A complaint that I often hear about how IT journalists cover open source is that when we do write about it, it can be on a very trivial level. The most obvious manifestation of this is in the cult of celebrity.
As far as most of my editorial colleagues are concerned, there is only one celebrity in the world of Linux: Linus himself. If I were to tell an editor that I'm coming to LCA, the first thing they'll ask is: "Will Linus be there? Can you get an interview?" And of course the answers are "He might be there, but I won't know until I get there" and "I'll certainly ask, but he's probably not going to say anything he hasn't already stated on the kernel mailing list". What I won't say is "I really don't feel like hassling the bloke when he's down here just to ask him what the future holds for Linux AGAIN and get him to make random comments about Apple". But being a journalist often means having to ask awkward questions.
Again, I've certainly been guilty of stupid trivial coverage. I once asked Linus one of the most ridiculous questions of my career. I used to write an advice column called 'Trivial Geek Questions' for Atomic, and one reader wrote in with a very odd question:
My 30th birthday is coming up soon, and I'd like to make it an event that none of my friends will forget in a hurry. How much would Linus Torvalds charge to come along and enliven the evening with his presence?
Remarkably, when I emailed this question to Linus, he actually answered it:
I don't think you can pay me enough. We have theme birthday parties too -- the best one we had was 'Reptile Lady', who had some cool snakes and lizards -- but a geek theme sounds a bit too sad for words.
You might view this as utterly trivial, and you'd probably be right. But it does have a humanising effect. And if Bruce Perens is right and the open source community needs to increase its visibility, a humanising effect might not be a bad thing.
Are we pressured by advertisers?
The final argument that I sometimes hear is that open source doesn't get covered because we don't want to offend the paying advertisers. Like most conspiracy theories, it lacks good evidence. It certainly has never actually happened on any of the technology magazines or sites I have worked on. For the most part, editorial staff have no idea whatsoever what's happening on the sales side, and that's a healthy situation. That doesn't mean the perception isn't there. I remember getting a very irate reader letter when I was at APC, saying that the magazine was clearly only writing so much about Windows because Microsoft was our biggest advertiser. It was an utterly uncompelling argument, largely because Microsoft never ever advertised in the magazine. Its brand awareness amongst geeks couldn't really get any higher; it was spending its money advertising on TV and in Woman's Day. I responded to the reader and told them that, but never heard back. People don't like the facts getting in the way of a good story, even if they do yell at journalists when that happens in print.
Does open source need the media?
Of course, the whole question of whether open source is covered sufficiently or fairly by the media presumes that getting that coverage is somehow important. And it wouldn't be too hard to argue that it really doesn't matter much. The fact that the Herald-Sun and the Daily Telegraph, the biggest-selling papers in Australia, will barely ever mention what's happening with the kernel or that LibreOffice might be a better choice than Microsoft Office, might be deemed irrelevant in the age of the Internet. If you want the information, it's out there. No-one needs to rely on just a handful of sources, everything is one search away, and anyone can publish their own views and news with minimal effort and cost.
It's also worth acknowledging that for a hefty chunk of the population, journalists in general are viewed as ill-informed, illiterate scum who have no sense of ethics or accuracy when it comes to doing their jobs. Does that apply to journalists who write about technology? It's hard to tell. One survey last year suggested that only 32 per cent of Australians trusted the media, a much lower figure than the global average of 49 per cent. On the other hand, the technology industry itself was trusted by 68 per cent. I'm hoping that means technology journalists fall somewhere in the middle on the hate scale. But you could also take it to mean that trying to get more journalists interested in open source issues and technologies is a pointless exercise.
Of course, the problem with taking that approach -- and this is something else Bruce Perens touched on in his keynote -- is that it effectively condemns the Linux user base to remaining somewhere around the 'three per cent of Lifehacker readers' mark. And I can't imagine that's the position that too many open-source enthusiasts want to take. You know that the open process can produce great software, and you'd love more people to use that software. Getting open source written about more often certainly isn't the only way that can happen, but it can't hurt.
So on that note, I want to offer a few really quick thoughts on how to go about approaching a journalist or a media outlet if you do want to get them to write about a project or an issue that you're involved with. None of this stuff is rocket science -- that's a different LCA presentation -- but getting the basics right is the most important thing if you're trying to get noticed by a busy journalist.
So #1: Identify the audience. The most crucial and obvious step, and the one that gets ignored an amazing percentage of the people who contact me, is to work out what issues the title writes about and what approach it takes. Lifehacker writes about how to do everything better, so we're interested in cool new software and ways to hack existing software and ways to make life more productive generally. We like free stuff, and actual stuff you can use right now. From an editorial point of view, we're not interested in how much money a project makes or who gets voted onto the LCA board or the fact that a given company has started using a new ERP platform or plans for a new site that hasn't actually launched. You can work all this out in about five minutes of looking at what we post. If you can't see an obvious reason why Lifehacker might want to write about your project and where it fits in, why on earth would you presume that I will?
#2: Be fashionable. It definitely pays to work out what topics are being covered regularly, as I've put it, and see if you can tie yourself into those. If you're championing a new open-source music player, explain exactly why it's better than iTunes. If you're promoting an open source conference, explain why the keynote speaker thinks Facebook is evil. If you're not sure what topics get covered a lot, you can do a lot worse than simply checking Google News.
#3: Be concise. The most likely way you'll contact a journalist is by email. I get several hundred emails a day, so the most important message is: keep it brief. Give a basic explanation of what you're promoting and a link to relevant materials.
#4: Finally, make sure you've included contact details: an email address for sure, and a phone number if at all feasible. If something needs to get checked in a hurry, phoning is how we're going to do it. Again, really obvious, but lots of people just forget.
If you follow those steps, your odds of getting coverage will jump right up. They're not guarantees, but merely getting them right will put you ahead of many of the professional PR people I know.
So, the big lesson? If there's a big anti-open-source conspiracy in the media, someone forgot to tell me about it. You don't have to take a cheesy tabloid approach to get open source noticed. But remember, if you don't get noticed, don't take it personally. Journalists, like coders, are entirely fallible human beings, not robots. We can always do our jobs better, and we can always use help in doing that. And if you lead us to a good story, we'll love you to bits.