|A Season for Open Source|
Coming to a lab near you…open source technology. Resources, software and instruments in the SLAS community model the end results of successful open source platforms. The time is right for user collaboration, but is the profession ready to embrace it?
“I estimate that 60 percent of all laboratory equipment and software already contain open source architecture or open source code,” states SLAS member and Journal of Laboratory Automation (JALA) Editorial Board member Neil Benn, M.Sc., co-founder and CEO of Ziath Ltd. “A lot of our products make use of open source technology. It is a building block to what we do. It cuts down the amount of time you need before you can create something,” he comments.
Benn offers an example within Ziath. When they decode two-dimensional barcodes, the team has its own barcode-detecting algorithm. “However, if we decode a one-dimension barcode, then we can take advantage of an open source library that’s produced by Google,” he explains. In open source spirit, Ziath communicates any bugs it detects to Google so others don’t have the problem in the future.
“What is needed is for people to comprehend what open source means,” Benn explains. “That’s difficult because there are a lot of areas where open source has really taken over with computer programmers and engineers driving it.”
SLAS member, SLAS Informatics Special Interest Group (SIG) chair and LabAutopedia editor James Gill, Ph.D. agrees with Benn. “It’s tough for people to get over the assumption that open source is a ‘wild west show’ with constant change and lack of reliability,” he comments, pointing to the fact that of the top four servers on the Internet, two are open source (Apache and NGINX), which together host more or less 74 percent of active web servers according to a November 2011 web server survey.
“Open source simply means that the end user has access to the source code and if needed can make changes to fit their needs. Most users of open source never need to make a change, but when a change is critical open source makes that change possible,” Gill continues.
“One hurdle for open source software is the laboratory community’s tendency to buy rather than build, leaving the decision to use open source to the vendors. However the vendor community is rapidly leveraging open source whether or not those companies provide access to their core product code to their own clients.” He offers as examples SLAS2012 exhibitor IDBS BioBook and the many systems that leverage The R Project for Statistical Computing for data analysis. “So whether we know it or not we are using open source every day," Gill continues.
Scott Weiss, IDBS director of product strategy, mentions that the E-WorkBook Suite (of which BioBook is part) has been developed in Java using the Eclipse framework and several other popular open source technologies. As a Java development shop, IDBS supports and contributes code to multiple open source initiatives.
“E-WorkBook has a rich API that allows our customers and partners to add new components (or plug-ins) that extend functionality using the Eclipse toolset,” Weiss comments. “So, yes, we do embrace using open source technologies, industry standard data formats and encourage third-party development. Currently, E-WorkBook Suite itself is not shared as an open source project, and IDBS does not, as a rule, curate or redistribute customer/third-party components as part of the core product offering. However, in many cases, great ideas first developed in the field or in partnership with a customer have found their way into the product.”
While open source software has made inroads, Gill states that open source hardware is harder to find in the laboratory. “I don’t know that we are seeing that same amount of open source hardware or instrument control,” he states. “Part of that is obvious. You are not going to open source a plate reader because the vendor needs a lot of time and effort to focus on developing the hardware. It’s a higher bar to get into the lab through hardware because of the capital involved in development.”
An open source project that has caught fire in the past few years is the Arduino platform used by many university labs that are developing robotics and other electronics projects, says SLAS member Roger McIntosh, vice president of engineering at IntegenX. “It’s a complete open source platform that includes electronic board designs, complete schematics and layout files so that the user can send them to a shop to build, or buy them pre-built,” explains McIntosh. “Along with the electronics, they have a nice software development tool to go along with it.”
The Arduino community is extremely open, according to McIntosh. “They have a committee that tests submissions and places approved material into their official library. People can submit work as an official set or post it on any number of open libraries on the web and place links in the forums,” he explains.
Gill gives some credit for the wider acceptance of open source platforms to the Maker Movement, which is creating a more mainstream profile for collaborative ventures. He describes this movement as “a group of people who embrace the openness of open source,” by posting and adapting plans for everything from clocks to complex robotry.
What does this mean within the laboratory science and technology community? The higher the profile for open source projects, the more likely all industries will begin to employ it. “We will see people realizing that they don’t have to pay hundreds of dollars to an instrument vendor for a small sensor. They can buy an Arduino and attach a sensor,” Gill says, noting that the appropriate Arduino may retail for $30; the sensor, around $10. “Because of all the effort put in by the Maker Movement to offer tools that are accessible and easy to use, I can bring up my sensor and connect it to my system in a couple of hours for $30-$40.”
This is particularly true, Gill notes, in academic labs. “If this trend continues to grow, it wouldn’t surprise me to see a vendor develop small products mostly composed of open source hardware parts,” he continues, mentioning that one such company is Ziath.
The primary problem is for open source to gain acceptance as a useful tool and not be viewed as an open season on intellectual property. “I think the concept makes people nervous,” says Benn. He shares a recent interaction he had with a representative of a large pharmaceutical company who asked how many people were using an open source framework developed by Benn before founding Ziath. “I told him I didn’t know because people download it, use it and tweak it. He asked, ‘what if it doesn’t work?’ I explained that you ask in the community for support with it,” Benn continues. “It’s difficult to get the concept of what open source technology actually is into the laboratory. Some people may not actually know that they are using it.”
A Collaborative Concept
Open source technology initiatives normally start small – perhaps you purchase something that doesn’t do quite what you need, for example. “You are reduced to begging the supplier to implement what you need it to do for you,” Benn explains. “If someone comes up with a good idea in open source format, they can implement it themselves. In an open source community, everyone works with the code, and it grows as people help each other to adapt it to a variety of purposes. When the application continues to be used and modified, it grows.”
At some point during this growth, companies emerge to offer support. “Then you have two choices,” according to Benn. “Pay for professional support from this newly formed company, or continue to get the free product support from the community, understanding you are improving the product for the good of other people. The product will always be free. The user does not have to buy the paid support because community support is available,” Benn says. “However, if people do not have the in-house resources, then they would take the free software and purchase support from a company.”
He mentions projects-in-progress that have been greeted with interest, such as LabVia, a resource for open source laboratory instrument interface software, and the open source initiative in high-content screening called OMERO. OMERO interest continues to escalate.
“When I started working with OMERO, it was good,” Benn comments. “I no longer work in that field, so I checked back to see how it was doing, and it has moved forward in leaps and bounds in two years. It is being driven by an unmet need.”
“I think the way it starts is by beginning to look at the open source community and bringing it into the lab,” says Gill. “I think the biologists have a head start in that there are quite a few open source initiatives. The Open Biological and Biomedical Ontologies (OBO) and other options open up biology information and make it available. Bioinformatics and sequencing were the cradle of open source, really.”
The first step is to find a resource for organizing the initiative behind someone who knows what needs to be developed and how to develop it. Benn notes that some of the large framework projects in academia have six or seven different sites to coordinate. “It’s no easy job,” he observes. “Linux only exists because the person on top of it is a good project manager.” This strong, organized figurehead, according to Benn, also should be someone charismatic who can get industry and academia to collaborate.
“In the scientific world, we need to have somebody who understands both the informatics side of it and can also adapt that for use within both biology and chemistry,” he advises, adding that this person likely would come from one of the more industry-minded academic institutions in the world. “That’s the issue. It has to be supported by the community itself. People need to work together to keep the applications working smoothly. It’s everybody’s job to support each other,” Benn continues.
Other important elements in establishing a successful open source platform, according to McIntosh, include setting up an infrastructure and making it conducive for people and organizations to use. “Establishing this infrastructure first involves setting up a web-based repository for the files, setting up the structure for accessing it and communicating the repository’s presence to the intended user group,” he describes. “Then you need to solicit a few initial high-quality projects to assess the system.”
Developing a critical mass of participation is another hurdle to vault, according to McIntosh. “I’ve seen a couple of open source attempts that have withered on the vine. You have to have participation and enough high-quality materials to make it of interest almost immediately. It requires a significant amount of work up front to get that together and keep it current,” he says.
“It is one of the classic problems you see in some of these efforts. Interest is developed, and then the key coordinators get busy and subject matter goes dark for a while. Then interest begins to decline,” McIntosh continues. “Some of these groups and efforts that are now successful have realized their success because they filled a void early on and were able to reach that critical mass because there weren’t so many options out there.”
As for attaining the proper size of the group, McIntosh advises that while an online community offers a much broader scope, some groups span too many different companies. “You could get bogged down and have decision-making paralysis with a larger group,” he warns. “There is a balance between too few and too many. You need a few to steer the project and then a broad base of people to support it.”
The SLAS Connection
Benn believes that SLAS is the perfect organization to motivate people around an open source initiative. “SLAS can be the grit around which the champagne bubble forms,” Benn says. “Essentially open source success needs that: One central organization that has no vested interest to put a drive into its development.”
According to Gill, a successful open source project within the SLAS community was launched in 2004 when the Society for Biomolecular Screening (SBS) published its microplate standards. “It was developed in a very open source way by SBS, and remains open source today. Some would say that it’s not, but I think it really is,” Gill states. “It points to the open source nature of science itself. Of all communities, this community should embrace that. There has always been a steady drumbeat of standards, consortia and people working together. An example would be the National Institute of Standards and Technology (NIST) work with the Analytical Information Markup Language (anIML) standards, which models a form of open source in the community. “Standards lend themselves to be open source activities,” Gill comments.
To further assist SLAS membership in appreciating the open source nature of standards development and ongoing activities, SLAS member and NIST Research Chemist Gary Kramer presented a new SLAS Standards Initiatives SIG at SLAS2012. The SIG showcased a number of current standards efforts in varying stages. Each presentation focused on the project's definition, approach and scope, players and backers, current status and future plans.
Benn believes that a successful open source campaign should be evangelized more. “If you look at what SBS did with the plate standards and how that revolutionized the industry – it’s the same evangelical nature that needs to happen with open source technology,” he asserts.
Gill agrees. “I think those groups have a real opportunity,” he notes. “I think our SIGs need to be trumpeted as the opportunity to build momentum around interest areas with unique needs and cultures. They need to be allowed to grow up into membership sections. The SLAS Informatics SIG could be a place in which open source finds its home.”
It comes down to volunteers, Gill asserts. “How much do you want to put back into the community? SLAS has the staff and infrastructure to help the communities do whatever it is they want to achieve, but the community has to provide the vision and take the lead.”
Benn cautions, too, that in the process it is important to maintain relationships on the commercial side, because vendors make the equipment and provide the services to the industry. “If they feel excluded by it, then they are not going to join in,” he observes.
Gill states that open source will be the way of the future. “A lot of kids coming up are being taught that they don’t need to rely on an industry to make a magic black box that does something. They can figure it out and make it themselves. Potentially cheaper, potentially better,” he concludes. “Imagine the kid who came up through the first robotics program, how they will look at automation. It will be completely different than how my generation looks at automation. For us it was a challenge, it was a big industrial capital request and you had to convince management of the need for it. For these kids, who have already built five robots before they show up at the door, it is something different. It’s so cool. The change that those kids are going to bring into our environment is awesome. The kids who are coming up have incredible capabilities and knowledge, and they are coming in expecting an open and sharing environment.”