Cyberinfrastructure, Innovation, and University Policy
Keck Center for the National Academies
February 21, 2008
Chuck Vest introduction
Chuck’s “six things speakers do”–he does one of each.
1. A profound-sounding question: Are we actually trying to design something or simply trying to understand what in the world is happening to us?
In the early days of the web we did understand what was going on–we were just using new technologies to do the same old things. Cyberinfrastructure (CI) is different because we don’t really understand what is happening to us. Including social scientists and manager is good–we need those people to help us understand it. When we look at new phenomenon like Second Life and YouTube, we really see youth using technologies in new ways.
2. A mysterious statement: We must both compete and cooperate
CI aids cooperation and collaboration in exciting ways. But we also need elbow room for healthy competition that drives excellence and stimulates innovation. When chuck was president of MIT, he got faxed letter from Bill Gates for meeting at Indiana University. No other university presidents were signing up, but Chuck ended up getting 30 min with Bill in a car to the airport. Chuck asked Microsoft’s help in funding $25 million to MIT to work on new uses of information technology in education. The take away: if you really have the right ideas, you’ll find the resources to do it.
3. A time-worn quotation: Universities never change, universities never innovate
Chuck sees a great deal of innovation within universities, but agrees that most of it ends up being closely held and not propagated to a larger community. Oftentimes different universities reinvent things. We don’t always continuously improve ourselves and boost others. But CI will give us the opportunity to do this in education, like it’s doing in research and scholarship.
4. A topical question: Why is it that we argue endlessly about rising health care cost and never do the things that will reduce them?
We need to utilize CI to directly address important public policy issues like a sensible deployment of health informatics. While technology is not always the solution to our problems, but this one is obvious. Involving the social sciences is key and can provide a useful venue to get this into the public debate. Careful, smart use of IT can go a long way–we need to get it on the public’s radar.
5. A macro-observation: In the 20th century, we had information stovepipes…scientists discovered, engineers designed, and medical doctors healed people. In the 21st century, science, engineering, and medicine are interdependent and blending together in many ways.
Chuck claims that around 80% of what he does at the National Academies work involves 2 or 3 of the big groups working together–no longer everyone doing their own thing.
6. Ask someone else’s question: Which institution is going to be the global university in the 21st century?
Chuck sees this as the wrong question. There are reasons for having physical educational presence in other countries, but this is not in the long run what is most important. The real question is how CI will allow for education to be restructured in the 21st century. He sees the evolution of a “meta-university.” The meta-university will form a substrate that allows institutions to move knowledge and pedagogy around the world, share the cost of development, won’t have to reinvent the educational wheel, will be robust, adaptable, and open. Chuck looks to John Seely Brown’s idea of “tinkering” as a framework for the possibilities of CI. We’ve used a distributed, collaborative process for the gathering of knowledge in astronomy for a long time. How do we move the open education and open content movements forward?
Conclusion:
The management of knowledge is a key issue of our time. We need to balance openness and free access to information with the formation of business models that will maintain the overall system. We need to maintain flexibility, form self-adapting systems that we cannot predict as we start now.
Cyber-Enabled Knowledge
Peter Freeman
Simon Porter
Guru Parulkar
Peter:
We need to focus on action by bringing together activists with provosts and university presidents. He agrees with Chuck in that we need to do more than talk about the high cost of health care? He questions why is there not widespread, ubiquitous technology implementation within the U.S. health care system? For example, webMD–where’d these go?
Pop quiz reveals that many of us do not know what “CI” means, what “innovation” means, what “university policy” is, and that our individual definitions often do not match with others’ definitions.
Eli Noam in a 1995 Science article predicted the demise of universities within 10 years. Noam said we overestimate near term future and underestimate the long-term future. Noam said that the essence of university is (1) production of knowledge, (2) storage of knowledge, and (3) transmission of knowledge. Peter wants to investigate how ICTs can work to affect and benefit this framework.
He says we understand that IT is going to change the world, but we don’t have a clue about how it’s going to change the university. Gives an example of an article in the NYtimes from that morning about the discrepancies in use of the Internet by youth, in that girls utilize Internet more than boys.
The key take-away: how can we utilize ICTs in production, storage and transmission of knowledge to enhance innovation?
Guru’s 4 points:
1. As we plan the 21st century CI, we need to make sure we build it on the right foundation. There are various layers of infrastructure including (1) physical, (2) internet, (3) web, and (4) knowledge. Normally, when most people talk about CI, we focus on the upper layer and assume the lower layers are fine. This might be reasonable to assume, but may be incorrect that the lower layers are solid. We need to ensure a strong base, a solid foundation of the lower CI layers. Guru is concerned about the Internet infrastructure layer, because he says all the ideas there are 40 years old. He says we need to champion the cause of reinventing the lower levels and not assume it as a given.
2. We all agree that Internet has been a great platform for innovation–gives example of Creative Commons, end-to-end principle and support for net neutrality. He think that letting the edges of the network be programmable are what has spurred the best innovation from organizations like Google. Therefore, he suggests we need to preserve the ideas of empowering the end users. Furthermore, we need to allow people to program not only the edges, but also program the infrastructure itself.
3. point 3 skipped due to time constraints
4. Universities need to lead these changes
If we are to reinvent the internet (or at least re-envision the structural makeup of the lower levels) universities need to be at the center of the movement. We need to employ CI within and between university campuses, and not leave it up to the vendors. We need to reinvigorate computer science academics to study diverse areas within their respective programs. We need to re-integrate network research and traditional CS research again.
Simon:
Simon suggests universities explore a position of “cyberadministrator” for research.
1. CI closes the gap between doing research and finding out about research. He says that provenance information is crucial within the virtual organization. He questions that if CI is prevalent, what does the world look like when we can find out about all the research projects within the university? He suggests that granting bodies need to buy into virtual CI itself.
Simple things CI can work to do:
-Begin to do is manage and aggregate university research portfolios
-Make sure research projects have a research documentation plan
-Find out what the research data storage needs are
-Present research data to the public can in interesting and transparent ways
-We can do away with disparate proxy web pages and instead opt for a university or inter-university “research portal”
-How can research learn from innovative businesses like Amazon or Ebay? We need to challenge the way research universities propagate their identities
-We need to connect those who have data and those who have the structure
-If we understand the structure of knowledge, we can build value on top of it through annotations
Q&A summarized:
In order for CI to operate effectively, we’ll need to see a change in people’s work habits. Openness and transparency will be key. Some universities will get if faster than others. Stanford “gets” it because they’ve had people working at places like Yahoo and Google. But in general, many universities have a disconnect between computer researchers and computer operators.
Longstanding question about what we should be focusing on, research first or standards first? Cargill says standards first, he’s the one who’s footing the bill and says we need an educational program in standards at universities. Guru says universities need research before standards because that is where the real innovation is born. Many things were created from a pre-standards phase, like Google, Second Life–things bubble up, move fast, standards follow afterwards. Simon takes the middle road, because standards help people access and actually access and understand data and research. Weinberger says that the nature of knowledge reflects the infrastructure. New CI allows more people to participate and contribute. Wilbanks says that when we move from network standards to knowledge standards, it proves a very difficult for semantic standards. Part of the notion of standards themselves need to change somewhat to better reflect meaning, semantics.
Group sees universities are stuck when provosts expect them to secure research funding from outside source before the university will throw in some money.
Sometimes industry says to academics: “you’re training folks to be just like yourself!” How can we create new industry/academic partnerships?
What is the incentive for professors to make their research open? Today, rewards are not geared towards publishing on the Internet? What can the incentive system to in order to take that leap? One idea might be to provide web space at a reputable place like NSF for researchers, which would house one’s homepage. This provides a good validation effect (one of the key functions that journals perform) from user standpoint because it’s difficult to sort out the quality of various works. Simon suggests researchers might have a persistent identifier.
The failure with OSI in the 80s came because they implemented standards before the research was done.
Doing research with no notion of “productization” is the key to success to universities - we don’t want universities to have “training” for industry as a main focus.
Jackson urges us to adopt more expansive notion of CI–so we don’t always pull back to the technology and leave everything else as the “soft stuff.” The layer model makes us do this, but we often see that the hardest problems are soft.
The Empowered University in the Global Economy
Jeff Lehman
Pradeep Khosla
Lesa Mitchell
Gerald Barnett
Susan Tuttle
Jeff:
Something he’s been doing–how can we use the world’s CI to have a more successful tool to enrich experience of students in overseas law school? Anecdote: an American lawyer talking with Chinese partners and American universities using existing products to produce better quality learning about American study of law. Back to Chuck’s question, “why globalize?”
There’s been an explosion in increasing outbound research and globalization, but what does this mean for CI? What does it say about university policy? What does it mean to administration about the mission of the university?
Important public policy question that should be in our minds–should the government funding of research be structured to promote or inhibit the widest possible dissemination of new knowledge as free public goods?
Pradeep:
We’ve learned lessons from building campuses, but what do we mean by an empowered global university? Global companies are different than global universities. Just because a university has put together a school in China, the quality is not the same as the main location. But, in industry, product quality is generally static across the board. Education has not changed that much in 300 years, and we need do more to enact useful change. Companies are investing globally in research. But while universities are the source of innovation, they are also the last to adopt that innovation. We need to globalize research, not undergraduate education. We need education that is customized for that location, not just a U.S. export of a particular institution.
Tuition increases are a really a big issue. How can we use CI to reduce or manage tuition costs, by possibly outsourcing some aspects of university like business processes?
We also need to push for better assessment metrics so we can understand where things are going right and where things need work.
Lesa
This is “not your father’s university.” The modern university is entrenched in innovation, commercialization and partnerships. University outputs of research have great commercial potential and are very interesting. Four years ago, all we talked about was technology transfer and the IP to licensing models. Since then, we understand the other pathways.
Can we build an infrastructure for people to be able to “see” what universities are doing? We’re attempting to do this with the iBridge program. There are 37 universities are on the iBridge network.
At Science Commons, John taught us the importance of looking at the bigger picture than just IP–things like materials, cell lines….how do you get access to them? The pathways to push the knowledge aren’t there yet, but SC and iBridge can help.
With the “better world project,” we aim to make the projects that don’t make millions through the tech transfer office visible to the larger community. We need to look at collaborative models like CITRUS at UC-Berkeley. We need to develop principles on IP, open collaborative models, and a better strategy.
Gerald
He says, “computers are serving the function of the bonfire in the forest.” Our debate over CI revolves around where the fire should be and who should be around it. He claims that it should be a very social thing, and that CI will affect and will be affected by our behaviors.
IP is not fundamentally about law; it’s about behaviors (and often our confusion of the “new”). The behaviors around IP haven’t been able to keep up with how we want to use it. IP is looking to manage relationships between groups. Industry has done much better with IP because they have markets to govern. Universities haven’t done as well because they only rely on policy. Universities don’t necessarily want an innovation policy–IP policies are really screwed up on campuses.
We shouldn’t have to break policy to enable innovation. Research is supposed to produce inventions, inventions are supposed to spur industry to make products–this is a 100-year-old idea.
We need to learn from other countries as they put together their environments. We can look to what groups like Creative Commons has been doing in trying to educate, normalize, and create things that people can use. CI is obviously about the social and not just the technical.
Susan
Susan questions why everyone talks about innovation. She says that the key to success (at least for IBM) is competition in the global marketplace by creating a value and differentiating from competitors.
A vision of CI must be (1) open, (2) collaborative, (3) mulit-disciplinary, (4) global.
1. Open: Innovation takes place in how people use technologies in new ways, not the technical challenges directly resulting from those technologies. If you free things up, a lot of innovation can take place on top of it. This is IBM’s new approach.
2. Collaborative: In the great days of IBM, the company told the customers what they want. This doesn’t happen anymore–solutions to real world problems are tied to user-centered design. She says that those companies and organizations that are struggling are usually the ones creating the best things.
3. Multi-disciplinary: Skills that students are developing do not help them understand how businesses work. Students don’t have the skills to put them to work right away. IBM wants students who are both broad and deep, to adapt to changing needs of customers.
4. Global: IBM does business in 170 countries, employs 350,000+ employees, and does 40% of its business remotely. We need to leverage CI to help support these trends.
Q&A summarized:
Students, like faculty, need incentives to participate. Students have limited time and limited funding. An alternative is to give students credit. Another idea is to hold a competition or give a reward to best team from the university working with CI.
We realize that the engrained culture is difficult to change. Many U.S. universities operate in ways that value the lone wolf student or researcher, but this doesn’t necessarily map onto the real work world. The Chinese critique is opposite, there they complain of no ability to break out of the team/pack mentality.
Gerald mentions that the student is the leader now–the propagator of true original research and intimate knowledge of changing technologies, social practices. Faculty needs to recognize this. CI is more than just content delivery. In the old generation (those in the room), we look at the Internet as a tool. For the new generation (youth), they look at the Internet as a way of life.
Kahin points out that the peer review system is fundamentally conservative and prone to stove piping. He claims that universities have a basic institutional structure than prevents researchers from going too far. Susan says that NSF grants and other and funding for universities needs to be sorted out. There are new models for development and we see new areas of academia (like business schools) doing important research (and getting funding) not traditionally mapped to those programs.
Breaking the “space barrier” is an important area for further study in CI. We haven’t done the proper assessment of the quality and impact of the virtual experience. We need to see a similarity or even equality between these two experiences. CI is efficient, interoperable, and a huge harmonizing influence, but we should also strive to seek a diversity of tone. We don’t have this tonal component in the infrastructure yet. We need to be able to talk, hug, debate, and keep tone as a part of the policy goals with CI.
Someone brought up the importance of the cell phone as ICTs in a global context. With widespread cell service disbursement, huge swaths of traditionally sidelined populations are now thrust into the world economy. How can CI support cell technologies that create a knowledge base and linkages in the 21st century?
Weinberger ends with the comment that as we make decisions and build different infrastructures for different values, how do we build infrastructure to make it a tool of globalization and not a tool of colonization?
Tony Hey lunch talk
Talked about (1) innovation and open source, (2) scholarly communication and open access, (3) scientific data deluge and (4) cloud computing.
1. Innovation and open source: The Bayh-Dole Act generates $30 billion of economic activity each year, and 250,000 jobs come from it. Stresses the importance of tech transfer offices and faculty communicating and understanding institutional and research goals. He says that not all academics and tech transfer offices understand the profound effect that different types of open source licenses can have on promoting and sharing work.
2. Scholarly communication and open access: In terms of scholarly publishing agreements, he questions how universities can have a vibrant system if they can’t afford academic journals? He suggests that a new model must emerge. Academics need access to quality journals, but don’t library budgets in mind. The University of Michigan cancelled 2500 journals this year because the costs are skyrocketing. There have been other movements in open access publishing such as the NIH mandate to release publicly funded research online within 12 months, Wellcome Trust, the recent Harvard Open Access policy.
Open access research repositories are becoming more important and have helped put universities like Southhampton high on the Google Scholar list. Open access and visibility is what people see and what gets articles cited. Virginia Tech maintains an open access electronic database of theses and dissertations, and requests grew from 220,000 in 1997 to 20 million in 2006. Open access research repositories can be full text and may contain “grey” literature such as workshop papers, presentations, technical reports and theses. In the future they’ll contain multimedia and software too.
3. Scientific data deluge on the horizon: Tony asks what the library’s role is in this space. We need librarians to be able to understand and implement semantic markups for CI– microformats from the bottom up, ontologies from the top down. We need librarians to nurture the ecosystem of services and tools to support the semantically rich publication, consumption, aggregation, integration, and processing of information. Librarians will be guardians of intellectual output of the university.
4. Cloud computing: The future, in Microsoft’s view. How do we build a software plus services model for science? We can leverage computing and data storage in the cloud; use distributed virtual computing so institutions don’t have to buy their own supercomputing systems, outsourcing.
Q&A summarized:
Do we see too much homogenization of research with such a centralized model?
With CI, we can’t have wishful thinking. It’s so ridiculously hard to get things right within our institutions. Even introducing new ideas (like transitioning to Gmail within universities) butts up against a lot of red tape.
Group agrees that the subscription model is fundamentally broken. Universities need to take action themselves about scholarly publishing–possibly mirror Harvard’s initiative.
What should universities be doing to support a bottom up semantic network? Tony suggests that we need to develop intelligent repositories, and then get top researchers’ papers in the archive. Hopefully we’ll see a trickle-down and eventual network effect of other faculty and researchers wanting their work in the repository too. This has happened with MIT’s Open CourseWare.
Designing for Integration and Collaboration
Linda Katehi
John Wilbanks
Mackenzie Smith
Chris Mackie
Sara Kiesler
Linda
Knowledge infrastructures have been present at universities for many years, informal or formal. But structures today are the same as what we’ve seen for 200 years. When we try to cross-interdisciplinary boundaries, we have to create new things–this is not sustainable. Many universities are out of space and money. How do we design for university environments and then sustain it?
John
What is the end goal of having knowledge infrastructure? We use content effortlessly today. We put our digital photos Flickr, others tag those photos, and the value-added content is then used by 3rd parties (like his photo that was used for the Washington DC travel guide).
Social investment of users is high for content, and this doesn’t necessarily translate to the knowledge environment. The way we make choices for intermediate sets of conditions affects the end result. When we look at initiatives like ARPANET, we see that content was not the end goal. Translating data from information to knowledge requires a high transaction cost. What surprised Tim Berners-Lee about things like Google was not the end goal, but the design choices. In fact, we may even go so far as to say that the end goal of the web, and of CI, is to create a system we cannot imagine.
Our intermediate goal in building CI should be to sever the longstanding idea that academic publishing is all-important, to remove the ideas of database and information silo-ing (one database per child), and to rid ourselves of ideas for atomization of information and intellectual property.
We also realize that not all knowledge is digital–look at Ebay and Amazon. Google is great about content in web pages, but not good for papers that aren’t indexed. We need to make it easy to build knowledge queries. Furthermore, we need to be successful in building human capacity.
Mackenzie
What is the knowledge infrastructure we’re talking about? We see both a functional view and an IT view.
The functional view involves 7 layers:
1. repositories
2. data management
3. linking: we need meaningful interoperability, useful semantics for data, encoding standards, ORE (object reuse exchange). At the library, we need to be not just interdisciplinary, but all disciplinary. We need to be able to combine data not just across two life sciences, but also between a vast array of disciplines.
4. discovery
5. delivery
6. social layer: collaboration tools, virtual research spaces
7. business layer
The IT view maps to these 7 layers:
1. repositories: s3, Palimpcest, ePrints
2. management: dSpace
3. RDF, sfx
4. Google, Worldcat
5. tools for visualizing data
6. Sakai, aTutor
7. ?: we need to explore this space more
Chris
How do we know what success looks like? Can we really “design for collaboration”? We excrete collaboration. Is some CI just “steaming piles of integration” like the physical “steaming piles of cash”? He questions how far we can push researchers, how much can we expect? If you’re funding researchers to be researchers, is it unfair to make researchers do CI too?
Chris thinks that we need both bottom up and top down CI, but how? Is all this “simultaneous” top down and bottom up just hand waving to get foundations to fund projects? Each of these separately will fall short. When we talk about collaboration are we just talking past each other? Do these just turn into pooling of names to try to get projects funded but then not really working together?
Pragmatically, you can’t reduce some disciplinary boundaries by reapplying others under CI. Sometimes we get a false sense that a great deal has been accomplished. Chris offers the odd suggestion that foundations might support infrastructure projects by the arbitrary funding of arbitrary researchers in arbitrary places. He says you can’t sustain CI for elite research universities if it proves to be unsustainable or if it flattens or doesn’t carefully incorporate liberal arts and other broad considerations. We are at a real risk of producing a permanent underclass of institutions.
Sara
The problem of collaboration is not due to differences in disciplines. There are oftentimes individual intra-institutional problems. There are different bureaucratic procedures, different cultures within universities–this is what slows down CI, collaboration and innovation. Universities now have a lot of power and self-interest that is often at odds with the goals of collaboration. For examples, universities will cooperate as long as the internal budget stays put. She also sees the problem that small institutions like community colleges don’t even have the small pieces of infrastructure in order to participate with the big players.
Q&A summarized:
Some big collaborations (like those at NSF) work because the organization was able to look at and support each piece of the collaboration. It is also noted “a lot don’t work.” We need to understand what’s going on in our university that needs reforming. Also important is the productive climate of the research project. Change within the university needs to be done with social and cultural implications in mind. We also need to be able sustain relationships even if someone leaves from the school or organization.
We need to recognize that projects shouldn’t need to begin and end with a funding line. We need to build for sustainability and extendibility in virtual organizations after the initial grant money is dries up. We can do this by making sure the collaboration on campus is inexpensive as possible. Need to also make the business case when applying for funding. 75% of Mellon-funded projects now out of funding are still going.
Australia is about to undergo a federal innovation review, important to keep IP issues in mind and push for decentralization in approach.
Someone suggests there may also be an institutional layer to add to Mackenzie’s 7-layer model.
How do we align the incentives and interests of the individual researchers, who may have slightly different interests or ideas about what to pursue? We need to implement interdependencies between the different parties. We need deliberate design of the governance structure–this is something we see in successful projects. People need to be able to argue, but still be able to work with each other.
On the Edge
Kaye Husbands-Fealing
Arti Rai
Elliot Maxwell
Brian Kahin
Kaye
She’s studied much of the literature in economics of higher education, science of science and innovation policy, cyberinfrastructure literature. Something to take away from this study is the idea that we need to continue the conversation between technical and stewardship in working for better interdisciplinary understanding.
Arti
Arti is interested in how university knowledge diffuses into the marketplace. There’s been lots of data generated in life science patenting, but less quantitative data in respect to IT. How does the university contribute to the global ecosystem?
The percentage of software patents has been increasing due to legal decisions making it easier to patent. There’s a lot of tacit knowledge wrapped up in life sciences research and some in IT patents. Who is pushing patenting at the university–researchers or tech transfer offices? We see that some researchers are weary of TTO and might try to commercialize it themselves or not commercialize it at all. Innovative initiatives at places like UC-Berkeley have combined licensing and sponsored research offices.
Nonexclusive copyright licensing may also be useful–we see this through organizations like Creative Commons and projects like iBridge.
Eliott
We see two dimensions that affect levels of openness in terms of information or processes: accessibility and responsiveness. We can think of this as a continuum of “closed” to “open.” Closed information and processes are like “ideas under the mattress”–where there’s no sharing at all. Open information and processes are like the web–the maximum levels of sharing.
He notes that open source is not really open in terms of the pure sense because there’s still judgment by experts or keepers of the kernel. Things can still be taken down from Wikipedia. This doesn’t make these initiatives less useful, but we should understand that openness is limited to the purpose. For example, we want our health records open to those who need it, but not to everyone. Projects like OCWs demonstrate the idea that you share what’s going on in your class as an analogy about how you view the university as a whole.
Concerning openness and CI, we want to be able to have the indexed information on all researchers and topics, even if the full piece is not there. This helps with transparency and allows others to find it. The notion that we should know about research can be a life and death situation, especially in places like medical science or the environment.
What happens when every student has a digital repository in 5 years? What if we make collaboration a goal of the university? Universities should study how collaboration can work (or doesn’t work). The university needs to rethink tenure structures and determine how to build in collaboration and sharing within this framework.
We realize that funders make things happen. We need to push for open access provisions like those adopted by the NIH to be adopted by the other government and private funding agencies.
Can the process of openness be “forced”–just as it has to be forced in the privacy realm? What are the benefits/drawbacks to this model?
Brian
How do we present this expanding CI ecology in a way that connects people? The University of Michigan is driven by a particular vision of knowledge. Do universities hold core research principles that align with their missions? Are university missions aligned with their own researchers’ interests and goals?
We need to think about collaboration science, but it’s difficult because CI is such an interdisciplinary science.
Standards making is inherently collaborative, but one of the gaps of CI is that we talk about standards without specific action being taken.
We also need to examine the strategic significance of physical presence, of place.
Innovation policy is critical to economic growth.
Q&A summarized:
We need to keep in mind the end user–in the university this is the faculty member or student. Normally the end user is the consumer. There’s an important distinction to be made today that now consumers are producers too (and consumers can often make things better than producers–von Hippel).
There was a question about data sustainability and when does the data no longer need to be stored (and who decides this)? We need to take a domain-specific or community-specific approach to this. Sometimes we might not want everything to be sustainable or live forever.
A big question is how do we reflect the potentials of sharing across communities for the remixing of resources. How do we foster remix and reuse and also communicate with each other?
We’ve come to realize that we can’t define CI, but we need to work toward where we want to get and what we want to accomplish: more openness and knowledge sharing–this is what to strive for. It seems odd, but part of what CI is is what it’s supposed to do.
It’s also extremely difficult to define “infrastructure.” As we go along, we need to be aware of it and to remain cognizant of the process.
We shouldn’t discount the value of competition and rivalry between universities. Competition helps produce interesting things.
What about this set of questions in relation to teaching? We’ve been talking about research all day long.
For others to understand what CI is doing, we need to push for useful language describing it. We need a CI documentation project.
