ATO2014: Open source, marketing and using the press

Steven Vaughan-Nichols was up to talk to us about open source, marketing and using the press.

Before Steven was a journalist he was a techie. This makes him unusual as a journalist who actually gets technology. Steven is here to tell us that marketing is a big part of your job if you want a successful open source company. He has heard a lot of people saying that marketing isn’t necessary anymore. The reason it’s necessary is because writing great code is not enough – if no one else knows about it it doesn’t matter. You need to talk with people about the project to make it a success.

We like to talk about open source being a meritocracy – that’s not 100% true – the meritocracy is the ideal or a convenient fiction. The meritocracy is only part of the story – it’s not just about your programming it’s about getting the right words to the right people so that they know about your project. You need marketing for this reason.

Any successful project needs 2 things – 1 you already know – is that it solves a problem that needs a solution – the other part is that it must be able to convince a significant number of people that your project is the solution to their problem. One problem open source has is that they confuse open source with the community – they are not the same thing. Marketing is getting info about your project to the world. The community is used for defining what the project really is.

Peter Drucker, says “The aim of marketing is to know and understand the customer so well the product or service fits him and sells itself.” Knowing the customer better than they know themselves is not an easy job – but it’s necessary to market/sell your product/service. If your project doesn’t fit the needs of your audience then it won’t go anywhere.

David Packard: “Marketing is too important to be left to the marketing department” – and it really is. There is a tendency to see marketing as a separate thing. Marketing should not be a separate thing – it should be honest about what you do and it should be the process of getting that message to the world. Each person who works on the project (or for the company) is a representative of your product – we are always presenting out product to the world (you might not like it – but it’s true). If your name is attached to a project/company then people are going to be watching you. You need to avoid zinging competing products and portray a positive image about you and your product. Even if you’re not thinking about what you’re saying as marketing it is.

Branding is another thing that open source projects don’t always think this through enough – they think this is trivial. Branding actually does matter! What images and words and name you use to describe your product matter. These will become the shorthand that people see your project as. For example if you see the Apply logo you know what it’s about. In our world of open source there is the Red Hat shadow man – whenever you see that image you know that means Red Hat and all the associations you have with that. You can use that association in your marketing. People might not know what Firefox is (yes there are people who don’t know) but they do recognize the cute little logo.

You can no longer talk just on IRC or online, you have to get out there. You need to go to conferences and make speeches and get the word out to people. And always remember to invite people to participate because this is open source. You have to make an active network and get away from the keyboard and talk to people to get the word out there. At this point you need to start thinking about talking to people from the press.

One thing to say to people, to the press, is a statement that will catch on – a catch phrase that will reach the audience you want to reach. The press are the people to talk to the world at large. These are people who are talking to the broader world – talking to people at opensource.com and other tech sites is great – but if you want to make the next leap you need to get to these type of people. Don’t assume that the press you’re talking to don’t know what you’re talking about – but just because they happen to like open source or what you’re talking about – it does not mean that they will write only positive things. The press are critics – they’re not really on your side – even if they like you they won’t just talk your products up. You need to understand that going in.

Having said all that – you do need to talk to the press at some point. And when you do, you need to be aware of a few things. Never ever call the press – they are always on perpetual deadline – you can’t go wrong with email though. When you do send an email be sure to remember to cover a few important things: tell then what you’re doing, tell them what’s new (they don’t care that you have a new employee – they might care if a bigwig quits or is fired), get your message straight (if you don’t know what you’re doing then the press can’t figure it out), and hit it fast (tell them in the first line what you’re doing, who your audience is and why the world should care). Be sure to give the name of someone they can call and email for more info – this can’t be emphasized enough – so often Steven has gotten press releases without contact info on them. Put the info on your website – make sure that there is always a contact in your company for the press. Remember if your project is pretty to send screenshots – this will save the press a lot of time in installing and getting the right images. Steven says “You need to spoon feed us”.

You also want to be sure to know what the press person you’re contacting writes about – do your homework – don’t contact them with your press release if it’s not something they write about. Also be sure to speak in a language that the person you’re talking to will understand [I know I always shy away from OPAC and ILS when talking to the press]. Not everyone you’re talking to has experience in technology. Don’t talk down to the press, just be sure to talk to the person in words they understand. Very carefully craft your message – be sure to give people context and tell them why they should care – if you can’t tell them that there they can’t tell anyone else your story.

Final points – remember to be sweet and charming when talking to the press. When they say something that bothers you, don’t insult the press. If you alienate the press they will remember. In the end the press has more ink/pixels than you do – their words will have a longer reach than you do. If the press completely misrepresents you be sure to send a polite note to the person explaining what was wrong – without using the word ‘wrong’. Be firm, but be polite.

ATO2014: Women in Open Source

20141023_095243

DeLisa Alexander from Red Hat was up next to talk to us about women in open source.

How many of you knew that the first computer – the ENIAC was programmed by women mathematicians? DeLisa is here to share with us a passion for open source and transparency – and something similarly important – diversity.

Why does diversity matter? Throughout history we have been able to innovate our way out of all kinds of problems. In the future we’re going to have to do this faster than ever before. Diversity of thoughts, theories and views is critical to this process. It’s not just “good” to think about diversity, it’s important to innovation and for solving problems for quickly.

Why are we having so much trouble finding talent? 47% of the workforce is made up of women but only 12% are getting computer and information science degrees – and only 1-5% of open source contributors are women. How much faster could we solve the world’s big problems with the other 1/2 of the population were participating? We need to be part of this process.

When you meet a woman who is successful in technology – there is usually one person who mentored her (man or woman) to feel positive about her path – we could be that voice for a girl or woman that we know. Another thing that we can do is help our kids understand what is going on and what opportunities there are. Kids today don’t think about the fact that the games they’re playing were developed by a human – they just think that computers magically have software on them. They have no clue that someone had to design the hardware and program the software [I actually had someone ask me once what 'software' was - the hardest question I've ever had to answer!].

We can each think about the opportunities in open source. There is the GNOME for women program, Girl Develop It and the Women in Open Source award.

The challenge for us is to decide on one person that we’re going to try and influence to stay in the field, join the field, nominate for an award. If each of us do this one thing, next year this room could be filled with 50% women.

ATO2014: Open Source at Facebook

Faebook and Open Source

James Pearce from Facebook started off day 2 at All Things Open with his talk about open source at Facebook.

James started by playing a piece of music for us that was only ever heard in the Vatican until Mozart as a boy wrote down the music he heard and shared it with the world. This is what open source is like. Getting beautiful content out to the world. Being open trumps secrecy. At Facebook they have 211 open source projects – nearly all on Github with about 21 thousand forks and over 10 million lines of codes. In addition to software Facebook also open sources their hardware. Open source has always been part of the Facebook culture since day 1. The difference is that now that Facebook is so large they are much more capable of committing to share via open source.

Here’s the thing people forget about open source – open source is a chance to open the windows on what you’re doing – “Open source is like a breeze from an open window”. By using open source it means they have to think things through more and it means they’re doing a better job on their coding. Facebook however was not always so dedicated to open source – if you looked at their Github account a few years ago you were see a lot of unsupported projects or undocumented projects. “The problem if you throw something over the wall and don’t care about it it’s worth than not sharing it at all”. About a year ago Facebook decided to get their open source house in order.

The first thing they needed to do was find out what they owned and what was out there – which projects were doing well and which were doing badly. The good news was that they were able to use Github’s API to gather all this information and put it in to a database. They then make all this data available via the company intranet so that everyone can see what the status of things is. Once of the nice side effects of sharing this info and linking an employee to each project is that it gamifies things. The graphs can we used to make the teams play off each other. Using things like Github stars and forks they compete to see who is more popular. Why they’re not optimizing on the number of stars, but it does make things fun and keeps people paying attention to their projects.

Also using the data they were able to clean up their “social debt” – they had some pull requests that were over a year old with no response. This gets them thinking about the community health of these projects. They think about the depth of a project, how they’re going to be used and how they’re going to continue on. Sometimes the things they release are just a read only type thing. Other times they will have forked something and will have a stated goal to upstream it to the original project. Sometimes a project is no longer a Facebook specific project. Sometimes Facebook will deprecate a project – this happens with a project that is ‘done’ or is of no longer of use to anyone. Finally they have in the past rebooted a project when upstreaming was not an option.

After giving talks like this James finds that lots of people approach him to talk about their solutions and find that they’re all coming up with the same solutions and reinventing the wheel. So these groups have come together with the idea of pooling their resources and sharing. This was the way TODO started. This is not a Facebook initiative – they’re just one of 13 members who are keen to contribute and share what they learned. This group is thinking about a lot of challenges like why using open source in the first place, what are the policies for launching a new project, licenses, how to interact with communities, what are the metrics to measure the success of a project, etc etc. What they hope to do is start up conversations around these topics and publish these as blogposts.

ATO2014: Pax Data

Doug Cutting

Doug Cutting from Cloudera gave our closing keynote on day 1.

Hadoop started a revolution. It is an open source platform that really harnesses data.

Doug CuttingIn movies the people who harness the data are always the bad guys – so how do we save ourselves from becoming the bad guy? What good is coming out of good data?

Education! The better data we have the better our education system can be. Education will be much better if we can have a custom experience for each student – these kinds of observations are fed by data. If we’re going to make this happen we’re going to need to study data about these students. The more data you amass the better predictions you can make. On the flip side it’s scary to collect data about kids. inBloom was an effort to collect this data, but they ended up shutting down because of the fear. There is a lot of benefit to be had, and it would be sad if we didn’t enable this type of application.

Heathcare is another area this becomes handy. Medical research benefits greatly from data. The better data we collect the better we can care for people. Once again this is an area that people have fears about shared data.

Climate is the last example. Climate is changing and in order to understand how we can effect it data plays a huge role. Data about our energy consumption is part of this. Some people say that certain data is not useful to collect – but this isn’t a good approach. We want to collect all the data and then evaluate it. You don’t know in advance what value the data you collect will have.

How do we collect this data if we don’t have trust? How do we build that trust? There are some technology solutions like encrypting data and anonymizing data sets – these methods are imperfect though. In fact if you anonymize the data too much it muddies it and makes it less useful. This isn’t just a technical problem – instead we need to build trust.

The first way to build trust is to be transparent. If you’re collecting data you need to let people know you’re collecting it and what you’re going to use it for.

The next key element is establishing best practices around data. These are the technical elements like encryption and anonymization. This also includes language to agree/disagree to ways our data is shared.

Next we need to draw clear lines that people can’t step over – for example we can’t show someone’s home address without their express permission. Which gives us a basis for the last element.

Enforcement and oversight is needed. We need someone who is checking up on these organizations that are collecting data. Regulation can sound scary to people, but we have come to trust it in many markets already.

This is not just a local issue – it needs to be a global effort. As professionals in this industry we need to think about how to build this trust and get to the point where data can be stored and shared.

ATO2014: Saving the world: Open source and open science

Academese

Marcus Hanwell, another fellow opensource.com moderator, was the last session of the day with his talk about saving the world with open source and open science!

In science there was a strong ethic of ‘trust, but verify’ – and if you couldn’t reproduce the efforts of the scientist then the theory was dismissed. The ‘but verify’ part of that has kind of gone away in recent years. In science the primary measure of whether you were successful or not was to publish – citations to your work are key. Then when you do publish your content is locked down in costly journals instead of available in the public domain. So if you pay large amounts of money you can have access to the article – but not the data necessarily. Data is kept locked up more and more to keep the findings with the published person so that they get all the credit.

AcademeseJust like in the talk earlier today on what Academia can learn from open source Marcus showed us an article from the 17th century next to an article today – the method of publishing has not changed. Plus these articles are full of academese which is obtuse.

All of this makes it very important to show what’s in the black box. We need to show what’s going on in these experiments at all levels. This includes sharing your steps to run calculations – the source code used to get this info should be written in open source because now the tools used are basically notebooks with no version control system. We have to stop putting scientists on these pedestals and start to hold them accountable.

A great quote that Marcus shared from an Economist article was: “Scientific research has changed the world. Now it needs to change itself.” Another was “Publishing research without data is simply advertising, not science.” Scientists need to think more about licenses – they give their rights away to journals because they don’t pay enough attention to the licenses that are out there like the creative commons.

What is open? How do we change these behaviors? Open means that everyone has the same access. Certain basic rights are granted to all – the ability to share, modify and use the information. There is a fear out there that sharing our data means that we could prove that we’re wrong or stupid. We need to change this culture. We need more open data (shared in open formats) and using open source software, more open standards and open access.

We need to push boundaries – most of what is published in publicly funded so it should be open and available to all of us! We do need some software to share this data – that’s where we come in and where open source comes in. In the end the lesson is that we need to get scientists to show all their data and not reward academics solely for their citations because this model is rubbish. We need to find a new way to reward scientists though – a more open model.

ATO2014: Open Source in Healthcare

Data center

Luis Ibanez, my fellow opensource.com moderator, was up next to talk to us about Open Source in Healthcare. Luis’s story was so interesting – I hope I caught all the numbers he shared – but the moral of the story is that hospitals could save insane amounts of money if they switched to an open system.

There are 7 billion people on the planet making $72 trillion a year. In the US we have 320 million people and that’s 5% of the global population, but we make 22% of the economic production on the planet – what do we do with that money? 24% of that money is spent on healthcare ($3.8 trillion) – not just the government, this is the spending of the entire country. This is more than they’re spending in Germany and France. However we’re ranked 38th in healthcare quality in the world. France is #1 however and they spend only 12% of their money on healthcare. This is an example of how spending more money on the problem is not helping.

Is there something that geekdom can do to set this straight? Luis says ‘yes!’

So, why do we go to the doctor? To get information. We want the doctor to tell us if we have a problem they can fix and know how to fix it. Information connects directly to our geekdom.

Data centerToday if you go to a hospital our data will be stored in paper and will go in to a “data center” (a filing cabinet). In 2010 84% of hospitals were keeping paper records versus using software. The healthcare industry is the only industry that needs to be paid to get them to switch to using software to store this information – $20 billion spent between 2010 and 2013 to get us to 60% of hospitals storing information electronically. This is one of the reasons we’re spending so much on healthcare right now.

The problem here (and this is Luis’s rant) is that the hospitals have to pay for this software in the first place. And you’re not allowed to share anything about the system. You can’t take screenshots, you can’t talk about the features, you are completely locked down. This system will run your hospital (a combination of hotel, restaurant, and medical facility) – they have been called the most complex institution of the century. These systems for a 400 bed hospital cost $100 million – and they have to buy these systems with little or no knowledge of how they work because of the security measures around seeing/sharing information about the software. This is against the idea of a free market because of the NDA you have to sign to see the software and use the software.

An example that Luis gave us was Wake Forest hospital which ended up being in the red by $56 million. All because they bought software for $100 million – leading to them having to fire their people, stop making retirement payments and other cuts. [For me this sounds a lot like what libraries are doing - paying salaries for an ILS instead of putting money toward people and services instead and saving money on the ILS]

Another problem in the medical industry is that 41% (less than 1/2) have the capability to send secure messages to patients. This is not a technology problem – this is a cultural problem in the medical world. Other industries have solved this technology problem already.

So, why do we care about all of this? There are 5,723 hospitals in the US, 211 of them are federally run (typically military hospitals), 413 are psychiatric, 2,894 are non profits and the others are private or state run. That totals nearly 1 million beds and $830 billion a year is spent in hospitals. The software that these hospitals are buying costs about $250 billion.

The federal hospitals are running a system that was released in to the public domain called VistA. OSEHRA was founded to protect this software. This software those is written in MUMPS. This is the same language that the $100 million software is written in! Except there is a huge difference in price.

If hospitals switched they’d spend $0. To keep this software running/updated we’d need about 20 thousand developers – but if you divide that by the hospitals that’s 4 developers per hospital. These developers don’t need to be programmers though – they could be doctors, nurses pharmacists – because MUMPS is so easy to learn.

ATO2014: Open Source & the Internet of Things

Internet of Things

Erica Stanley was up next to talk to us about Open Source and the Internet of Things (IoT).

The Internet of Things (Connected Devices) is the connection of things and people over a network. Why the Internet of Things? Why now? Because technology has made it a possibility. Why open source Internet of Things? To ensure that innovation continues.

Some of the applications we have for connected devices are: Health/Fitness, Home/Environment and Identity. Having devices that are always connected to us allow us to do things like monitor our health so that we can see when something might be wrong before we feel symptoms. Some devices like this are vision (Google glass) related, smart watches, wearable cameras, wristbands (fitbit), smart home devices (some of which are on my wishlist), connected cars (cars that see that the car in front of you has stopped versus slowed down) and smart cities like Raleigh.

Internet of ThingsThere are many networking technologies these devices can use to stay connected, but bluetooth seems to be the default that is being used. There is a central device and a peripheral device – the central device wants the data that the peripheral device has. They use bluetooth to communicate with each other – the central device requesting info from the peripheral.

Cloud commuting, another important technology, has been one of the foundations for the Internet of Things – this is how we store all the info we’re passing back and forth. As we get more ability for our devices to learn we get more devices that can act on the data they’re gathering (there is a fitness app/device that will encourage you to get up and move once in a while for example).

Yet another technology that’s important is augmented reality showing us results of data in our day to day (Google glass showing you the directions to where you’re walking).

One challenge facing us is the fact that we have devices living in silos. So we have Google devices and Samsung devices – but they don’t talk to each other. We need to move towards a platform for connected devices. This will allow us to have a user controlled and created environment – where the devices I want to talk to each other can and the people I want to see the data can see the data. This allows us to personalize our environment but also secure our environment.

Speaking of security, there are some guidelines for developers that we can all follow to be sure to create secure devices. When building these devices we want to think about security from the very beginning. We need to understand our vulnerabilities, build security from the ground up. This starts with the OS so that we’re building an end-to-end solution. Obviously you want to be proactive in testing your apps and use updated APIs/frameworks/protocols.

Some tools you can use to get started as far as hardware: Arduino Compatible devices (Lilypad, Adafruit Flora and Gemma), Tessel, and Metawear. Software tools include: Spark Core, IoT Toolkit, Open.Sen.se, Cloud Foundry, Eclipse IoT Tools, and Huginn (which is kind of an open source IFTTT).

One thing to keep in mind when designing for IoT is that we no longer own the foreground – we might not have a screen or a full sized screen. We also have to think about integration with other devices and discoverablity of functionality if we don’t have a screen (gesture based device). Finally we have to keep in mind low energy and computing power. On the product side you want to think about the form factor – you don’t want a device that no one will want to wear. This also means creating personalizable devices

Remember that there is no ‘one size fits all’ – your device doesn’t have to be the same as others that are out there. Try to not get in the way of your user – build for people not technology! If we don’t try to take all of the user’s attention with the wearable then we’ll get more users.

ATO2014: How Raleigh Became an Open Source City

Open Raleigh

Next up was Jason Hibbets and Gail Roper who gave a talk about the open source initiative in Raleigh.

Gail started by saying ‘no one told us we had to be more open’. Instead there were signs that showed that this was a good way to go. In 2010 Forbes labeled Raleigh one of the most wired cities in the country, but what they really want is to be the most connected city in the country.

Raleigh has 3 initiatives open source, open data, and open access – the city wants to get gigabit internet connections to every household. So far they have a contract with AT&T and they are working with Google to see if Raleigh will become a Google fiber city.

The timeline leading up to this though required a lot of education of the community about what open meant. It didn’t mean that before this they were hiding things from the community. Instead they had to teach people about open source and open access. There were common stereotypes that the government had about open source – the image of a developer in his basement being among them.

Why did they do this? Why do they want to be an open city? Because of SMAC (Social, Mobile, Analytics, Cloud). Today’s citizens expect that anywhere on any device they should be able to connect to the web. Government organizations like Raleigh’s will have 100x the data to manage. So providing a government that is collaborative and connected to the community becomes a necessity not an option.

“Empowerment of individuals is a key part of what makes open source work, since in the end, innovations tend to come from small groups, not from large, structured efforts.” -Tim O’Reilly

Next up was Jason Hibbets who is the team lead on opensource.com by day and by night he supports the open Raleigh project. Jason shared with us how he helped make the open Raleigh vision a reality. He is not a coder, but he is a community manager. Government to him is about more than putting taxes in and getting out services – it’s about us – the members of the community.

Jason discovered CityCamp – a government unconference that brings together local citizens to build stronger communities where they live. These camps have allowed for people to come together to share their idea openly. Along the way the organizers of this local CityCamp became members of Code for America. Using many online tools they have made it easy to communicate with their local brigade and with others around the state. There is also a meetup group if you’re in the area. If you’re not local you can join a brigade in your area or start your own!

Jason has shared his story in his book The foundation for an open city.

ATO2014: What Academia Can Learn from Open Source

Git process

Arfon Smith from Github was up to talk to us about Academia and open source.

Arfon started with an example of a shared research proposal. So you create a document and then you edit the filename with each iteration because word processing applications is not good at tracking changes and allowing collaboration. Git though is meant for this very thing. So he showed us a book example on Github where the collaborators worked together on a document.

In open source there is this ubiquitous culture of reuse. Academia doesn’t do this – but why not? The problem is the publishing requirement in academia. The first problem is that ‘Novel’ results are preferred. You’re incentivized to publish new things to move ahead. The second problem is that the value of your citation is more powerful than the number of people you’ve worked with. And thirdly, and more generally, the format sucks. Even if it’s an electronic document it’s still hard to collaborate on it (see the document example above). This is state of the art technology … for the late 17th century. (Reinventing Discovery).

So, what do open source collaborations do well? There is a difference sometimes between open source and open source collaborations, this is an important distinction. Open source is the right to modify – it’s not the right to contribute back. An open source collaborations are highly collaborative development processes that allow anyone to contribute if they show an interest. This brings us back to the ubiquitous culture of reuse. These collaborations also expose the process by which they work together – unlike the current black box of research in academia.

Git processHow do we get 4000 people to work together then? Using git and Github specifically you can fork the code from an existing project and work on it without breaking other people’s work and then when you want to contribute it back you submit a pull request to the project. The beauty of this is ‘code first, permission later’ and every time this process happens the community learns.

The goal of a contribution of Github is to get it merged in to the product. Not all open source projects are receptive to these pull requests though, so those are not the collaborative types of projects.

Fernando Perez: “open source is .. reproducible by necessity.” If you don’t collaborate then these projects wouldn’t move forward – so they need to be collaborative. The difference in academia is that you have to work alone to and in a closed fashion to move ahead and get recognition.

Open can mean within your team or institution – it doesn’t have to be worldwide like in open source. But making your content electronic and available (which does not me a word doc or email) makes working together easier. Academia can learn from open source – more importantly academia must learn from open source to move forward.

All the above seems kind of negative, but Arfon did show us a lot of examples where people are sharing in academia – we just need to get this to be more widespread. Where might more significant change happen? The most obvious place to look is where communities form – like around a shared challenge – or around shared data. Science and big data are where we’re going to see this more hopefully.

There are challenges still though – so how do we make sharing the norm? The main problem is that academic reward ‘credit’ – so articles written by you solely. Tools like Astropy is hugely successful on github, but the authors had to write a paper about it to get credit. The other issue is trust – academics are reluctant to use other people’s stuff because we don’t know if their work is of value. In open source we have solved this problem already – if the package was downloading thousands of times it’s probably reliable. There are also tools like codeclimate that give your code a grade.

In short the barriers are cultural not technical!