Computing Research Policy Blog

Examiner Editorial on Math and Science Incentive Act

The DC Examiner ran an editorial today using the Math and Science Incentive Act of 2005 (CRA blog entry here) to focus on the lack of emphasis that primary, secondary, and university education place on teaching science and math. The editorial praises the Act, introduced by Frank Wolf (R-VA) in the House and John Warner (R-VA) in the Senate, which would forgive up to $10,000 in student loan interest for post-college work or teaching in mathematics, physical sciences, and engineering.

The piece notes, however, that this bill alone is insufficient:

Last week, the NSF's congressionally-mandated Committee on Equal Opportunities in Science and Engineering reported measurable but uneven gains in underrepresented groups. However, as Committee Chairman Robert Lichter put it, "bold, innovative and long-term initiatives are still needed, especially at the institutional level." Interest-free student loans are not quite in that league, but at least they're a start.

Updates on the status of the bill will appear in the blog if and when it gains traction in committee (Education and the Workforce in the House and Health, Education, Labor, and Pensions in the Senate).

Update posted June 29: The provisions of this bill have been rolled into the College Access and Opportunity Act, which was part of the higher education authorization. ACM has followed this issue in their blog.

Industry Group Calls for Increased Cyber Security R&D; Congress Hears Message from Former PITAC Members

In a report released this week, the Cyber Security Industry Alliance -- a group consisting of information security software, hardware and service vendors -- called on Congress and the Administration to ramp up support for fundamental research in cyber security R&D and increase the prominence of cyber security at key federal agencies. CSIA's report, Federal Funding for Cyber Security R&D (pdf) reiterates the findings of the most recent Presidential IT Advisory Committee (PITAC) report (pdf) on the state of federal cyber security research, concluding that the overall investment in cyber security research is inadequate and too focused on the short-term. The CSIA report agrees with the PITAC report's recommendation to increase funding for long-term research in cyber security, noting a number of key security technologies -- firewalls, intrustion detection systems, fault tolerant networks, operating systems, cryptography and advanced authentication -- that bear the stamp of federally-sponsored, long-term research.

The report differs from the PITAC report slightly in that it calls for the creation of a "designated entity" within DHS to coordinate the federal government's cyber security R&D effort; whereas, PITAC recommended that function remain within the interagency working group activity of the Networking and IT R&D program. CSIA rightly points out that the IWG of NITRD has very little actual influence on priority-setting at the agencies. Instead, they recommend that the new Assistant Secretary for Cyber Security at DHS serve as "the logical choice to drive the prioritization of requirements for research and development." My only concern with that recommendation is that DHS hasn't yet bought into the idea that long-term research efforts should be a priority. DHS's own budget for cyber security R&D remains a paltry $18 million for FY 05, out of an overall science and technology budget of just over a billion dollars. And of that $18 million, barely $2 million could realistically be described as "long-term" research efforts. (DHS's lack of priority for cyber security R&D has been a frequent topic here).

Otherwise, the CSIA report marches in lockstep with the PITAC report on cyber security R&D (pdf) issued back in March. We strongly endorsed that report and I'm pretty thrilled with the industry report issued this week.

Coincidentally, two former PITAC members (former because PITAC has been "disbanded" since June 1, 2005...) were on the Hill yesterday to participate in a briefing on cyber security R&D hosted by the Congressional Research and Development Caucus and put together by IEEE and IEEE-CS. Former PITAC Subcommittee on Cyber Security R&D Chair Tom Leighton (Chief Scientist and Co-Founder of Akamai) and former PITAC member Gene Spafford "Spaf" (Professor and Director of CERIAS at Purdue University) told the assembled congressional staffers, science community folks and assorted press about the problems we face in the cyber security arena and what PITAC recommended.

The briefing was the latest in a series of briefings on the PITAC report and follows a number of hearings on the scope of the cyber security challenge. In April, for example, Spaf and Leighton, along with former PITAC co-Chair Ed Lazowska, participated in a number of focused briefings for Hill staff on the PITAC report. The House Science Committee, as well as the House Homeland Security committee have both held numerous hearings on the subject over the last several years. Yet the extent of the problems we face -- the risk posed by cyber attacks on critical infrastructure, the exposure internet users have to fraud and abuse because of security vulnerabilities, the cost to industry due to cyber extortion and malicious acts -- still appears to shock to congressional staff. I'm not sure they really believe that companies have paid "protection" money to criminals who threatened to take down their web presence with massive distributed denial of service attacks. I'm not sure they really believe that "phishing" and "pharming" attacks are real threats to individual internet users. I'm not sure they understand that IT systems are in the control loop of just about every piece of critical infrastructure in the nation and are vulnerable. I think many believe that the impact of a concerted cyber attack would be limited to something like Amazon being unavailable for the day.

So despite the reports and briefings and hearings, we in the community haven't done a great job breaking through the noise around homeland security and conveying the importance of cyber security, or by extension cyber security R&D. In part, I think this is because the homeland security debate is really dominated by the specter of a nuclear, biological or chemical (NBC) attack (perhaps rightly so). The idea that a cyber attack could exist on the same scale as any one of the big three isn't so easily embraced by staff. Yet in terms of cost to industry and cost to government, the daily onslaught of cyber attacks must add up to dollar losses that exceed even some of the more dramatic NBC scenarios. But the investment in research to mitigate those losses, or prevent them entirely, pales in comparison to the investments in NBC research at DHS.

In any case, the continued efforts of folks like Spaf and Leighton, and industry partners like the members of CSIA and ITAA, are helping to educate members of Congress and their staff to the challenges in the area. And, for better or worse, the growing frequency of breeches of customer data held by credit card companies, banks, universities and others is forcing Congress to climb the learning curve....

Cerf in WSJ: America Gasps for Breath in the R&D Marathon

Turing Award winner Vint Cerf and ITAA head Harris Miller have a fantastic op-ed in today's Wall Street Journal raising concerns about US competitiveness in light of a declining federal R&D budget. The article is behind the WSJ pay wall, but can be viewed online for the next seven days here. Some snippets:

America will soon find its grip on the levers of international commerce slipping as we turn our backs on a proud tradition of technology innovation. The stewards of our national destiny are busily tightening the tap on the federal R&D budget, the most important source of funding for programs that seek to answer fundamental questions of science and technology.

...

In the 1960s and '70s, a collection of academics and private-sector technologists, including a co-author of this piece, used findings funded by the Pentagon's Advanced Research Projects Agency (now DARPA), to participate in implementation of the first wide-area packet switched network (the ARPANET) and the subsequent integrated collection of packet-switched networks (the Internet).

Now DARPA officials have revealed a shift in focus away from its history of open-ended long-range research, which typically has been performed in universities and nonprofit institutions. According to recent news reports, DARPA funding for university researchers in computer science has fallen from $214 million to $123 million from 2001 to 2004. Moreover, the focus of DARPA R&D is more near-term and more immediately defense-oriented than before. While this is defensible in some ways, the largest impacts of long-term research funded in the past by DARPA have been in areas that have wider or dual application to defense and the civilian sector.

The U.S. is already lagging behind in R&D funding. Our total national spending on R&D is 2.7% of our GDP, and now ranks sixth in the world, in relative terms, behind Israel (4.4%), Sweden (3.8%), Finland (3.4%), Japan (3.0%) and Iceland (2.9%). The federal government's share of total national R&D spending has fallen from 66% in 1964 to 25%.

Some of the outright cuts in the president's proposed R&D budget include the following:

The Department of Energy's Office of Science would see its R&D funding fall 4.5% to $3.2 billion.
 

The Department of Agriculture would see its R&D funding decline 14.6% to $2.1 billion.

Funding for all three multi-agency R&D initiatives would decline in FY 2006, a category that includes programs such as the National Nanotechnology Initiative and the Networking and Information Technology R&D initiative.
 
The proposed cuts come at a time when other nations have fixed their sights firmly on overtaking our technological lead, especially in information technology. For those of us in industry and academia, this shift in policy represents a major detour in the marathon race for global economic leadership.

The piece goes on to quote a number of indicators -- many of the same ones cited in the Task Force on the Future of American Innovation's influential Benchmarks of our Innovation Future report -- that show that while the U.S. remains in the leadership position in innovation and R&D investments, all of the trendlines are slanting the wrong way.

The facile solution is to turn to private industry and academia to make up the difference. But R&D funding from private industry is currently growing above inflation. It is susceptible to general economic cycles, and by its nature it is focused on the here and now. Meanwhile, many academic institutions are battling lagging enrollment and turning to unconventional fund-raising means merely to stay afloat. The difficulty in obtaining visas for foreign scientists has also restricted an important source of talent in the research community.

In a very real sense, today's R&D agenda determines where America will find itself in the future. The benefits of vigorous, federally funded academic R&D programs reaped by American society at large have been enormous. Our domestic and global economies thrive on the results of such work. Private sector programs alone cannot produce comparable results, in part owing to an ethical obligation to deliver bottom-line business results for their stockholders. The U.S. government needs a long-term strategy for continued economic growth. A strong and thriving academic R&D program is critical to that strategy. To choose otherwise is a recipe leading to irrelevance and decline.

I'm thrilled to see this piece in the WSJ today....

I'll have a bit more comment on this later when I have a few minutes, but I wanted to get the pointer to the article up asap. Read the whole thing, while it's still available!

Update: The article is finding it's way around Congress. Rep. Anna Eshoo (D-CA) circulated the piece in a "Dear Colleague" letter along with this text:

Once again, high technology leaders are warning that declining federal investments in research and development are allowing the rest of the world to catch up. This isn't a problem that can be blamed on Europe or developing economies in Asia. It's a problem that we're creating. If we're to maintain our economic leadership for future generations, we need to increase the federal commitment to R&D instead of cutting it.

Help Requested in Support of Defense Authorization Amendment

Update July 22, 2005: Jason Van Wey of the Coalition for National Security Research (CNSR) has more on the effort to see the amendment passed, including the news that the amendment has picked up a number of important cosponsors. As of this morning, joining Collins and Kennedy on the amendment are Sens. Hillary Rodham Clinton (D-NY), Elizabeth Dole (R-NC), Joe Lieberman (D-CT), Barbara Mikulski (D-MD), Pat Roberts (R-KS) and Rick Santorum (R-PA). Though this bipartisan list of co-sponsors bodes well for passage, your calls are still needed as the Senate works through the amendments to the Defense Authorization today and Monday!

Originally posted July 21, 2005: Word comes from AAU that Sen. Susan Collins (R-ME) and Sen. Ted Kennedy (D-MA) will introduce an amendment this afternoon to the FY 06 Defense Authorization bill under consideration today that would increase funding for a number of programs of interest to the computing research community, including a $10 million plus-up to "fundamental computer science and cybersecurity research at DARPA." Senators need to be made aware of the amendment and urged to support it. Here are the details from AAU:

During Senate consideration today of the FY06 Defense Authorization Act (S. 1042), Senators Edward Kennedy (D-MA) and Susan Collins (R-ME) will offer an amendment to authorize additional funding for the Department of Defense SMART National Defense Education Program, a new scholarship and fellowship program aimed at attracting more students into science and engineering fields.  The amendment also provides additional authorization for basic research programs at U.S. universities.
 
CFR members are urged to contact their Senators to ask that they support the Kennedy/Collins amendment when it is considered on the Senate floor.
 
A copy of an AAU statement (pdf) supporting the amendment is attached, along with talking points prepared by the Senators’ offices and the text of the amendment.
 
THE AMENDMENT
The amendment would authorize an additional $50 million for university research and education programs at the Department of Defense. 
 
Specifically, the Kennedy/Collins amendment:
 

Increases the SMART National Defense Education Program by $10 million;

Increases the Army University Research Initiatives (URI) account by $10 million;

Increases the Navy University Research Initiatives (URI) account by $10 million;

Increases the Air Force University Research Initiatives (URI) account by $10 million; and

Increases the DARPA account by $10 million and specifies that money should be spent on fundamental research in computer science and cybersecurity.
 
The amendment also includes a Sense of the Senate that the Department of Defense set a goal to invest 15 percent of its science and technology budget in basic research programs.  The current percentage varies between 11 percent and 12 percent. 

The amendment would "pay for" the increases -- every funding increase in an amendment to the bill has to be offset by a reduction somewhere else -- by reducing a planned $2 billion increase to the "defense-wide operations and maintenance fund for IT" by an equal amount.

The university community here in DC (along with CRA) is mobilizing to contact senators about the amendment. More calls would surely help. Urge your senator (by phone) to support the Collins-Kennedy amendment to the FY 06 Defense Authorization Bill. The bill is on the floor today, so the time is now! We'll have updates as developments warrant....

Here's a copy of the amendment as well as some talking points. Here's AAU's statement.

AAAS Report on Women and Minorities in the IT Workforce

MSNBC has some interesting coverage of an important but oft-overlooked part of our IT workforce: students seeking vocational rather than research-oriented IT training. The article covers the recent AAAS report entitled Preparing Women and Minorities for the IT Workforce: The Role of Nontraditional Educational Pathways. The article begins:

Pop quiz: Which schools produced the most degrees in computer science in 2001? MIT? Carnegie Mellon? Georgia Tech?

If you guessed any of these, you’re wrong: try Strayer University and DeVry Institute of Technology.

[...]

If you guessed [the typical student is] a young geeky guy with a pocket saver, guess again: try a 35-year-old African American or Hispanic woman who already has a full-time job at a company where information technology (IT) skills are a key to advancement.

She’s the one taking the night courses at one of the for-profit institutions like Strayer or DeVry that have a wide variety of locations, and offer courses in the early morning and evening, as well as on-line courses.

The study found that women, minorities, and non-traditional students were especially likely to take advantage of CS/CE educational opportunities from for-profit institutions. It is a helpful reminder that the future of computer science and engineering in the United States is dependent not just on researchers but on a non-research oriented IT workforce that can deploy the advances of CS/CE research and development throughout all areas of society.

CRA's Taulbee Survey maintains information about women earning CS/CE degrees from PhD-granting institutions. Results from recent years:

Gates on CS/CE Enrollment and Funding

InternetNews.com has coverage of the opening of Microsoft Research's sixth annual Faculty Summit, a "a unique opportunity for faculty members and Microsoft researchers, architects, and executives to collectively discuss a vision for the future of computing." Microsoft Chairman and CEO Bill Gates had some interesting comments to open the event (along with ACM past-President Maria Klawe). Here's a sample:

But today, Gates and Klawe focused on the present; specifically, how to encourage more students to enroll in computer-science programs so that the industry will have enough qualified engineers to work on those future innovations.

Klawe presented some grim figures: The popularity of computer science as a major has fallen more than 60 percent between 2000 and 2004, she said, even though the software engineering and several related jobs will be among the fastest growing through 2012.

Some of that slack might be taken up by girls if they didn't have such a seeming aversion to the field. Klawe said participation of women in computing has gone down over the past 25 years, with only around 15 percent of computer-science Ph.D.s going to women.

When Klawe asked Gates what could be done, he seemed to flounder. When he responded, "There's no magic answer. Maybe get women in the field to be more visible?" Klawe hooted him down.

"No, that's not the answer," she said. "We all do it, but we're not getting anywhere with it."

"You lose them at about five stages," Gates agreed. "And, if there aren't enough women in field, it makes it less attractive, even if everything else is good. There's a critical-mass element to this."

The decline in federal funding for academic research and graduate education doesn't help, the two agreed. Money from the Defense Advanced Research Projects Agency (DARPA) dropped by half last year.

"The biggest payoff for federal funding or research is in computer science," Gates said, pointing to the economic and technology boom of the 1990s. "Department of Defense money was one of the elements that allowed us to turn this into one of the greatest success periods the U.S. has ever had."

Computer science could fuel another such boom in the next 10 years, according to Gates.

"Computer science is becoming the toolkit for all the sciences," he said. As all disciplines become more data-driven, they're turning to computer science to make sense of the huge amounts of data. "Computer science helps model the world," he added.

Newsday also has coverage of the event, focusing on the declining enrollment in CS/CE question:

Speaking to hundreds of university professors, Microsoft Chairman Bill Gates said Monday that he's baffled more students don't go into computer science.

Gates said that even if young people don't know that salaries and job openings in computer science are on the rise, they're hooked on so much technology _ cell phones, digital music players, instant messaging, Internet browsing _ that it's puzzling why more don't want to grow up to be programmers.

"It's such a paradox," Gates said. "If you say to a kid, 'Yeah, what are the 10 coolest products you use that your parents are clueless about, that you're good at using,' I don't think they're going to say, 'Oh, you know, it's this new breakfast cereal. And I want to go work in agriculture and invent new cereals or something.' ... I think 10 out of 10 would be things that are software-driven."

...

Gates said computer scientists need to do a better job of dispelling that myth and conveying that it's an exciting field.

"How many fields can you get right out of college and define substantial aspects of a product that's going to go out and over 100 million people are going to use it?" Gates said. "We promise people when they come here to do programming ... they're going to have that opportunity, and yet we can't hire as many people as we'd like."

Both pieces are chock full of interesting quotes and worth reading. We'll have more on how the computing research community is organizing to take on these issues soon, so watch this space....

Update: Here's the transcript from Gates and Klawe's opening remarks. And here's a video.

Science Funding's Unintended Consequences

There's an interesting article by Sallie Baliunas at Tech Central Station today on research funding. The piece notes a recent Nature article that suggests scientific misbehavior might be linked to "perceptions of inequities in the [science] resource distribution process" and connects that with tendency among federal funding agencies to shift emphasis from basic to applied research.

Since 1970, total federal non-medical research spending as a fraction of Gross Domestic Product has declined by about one-third. No formal history has tracked research misbehavior, leaving it impossible to say if ongoing stresses on budget allocation systems would partly explain current misbehavior.

Continual budget pressures, though, are transforming U.S. research and development. Funding agencies now weigh more heavily a proposal's aim toward practical applications, especially those with near-term payoff.

The rest of the article focuses on this trend, citing as an example PITAC's 1999 report "Investing in our Future" that noted that federal funding in computing research was "excessively focused on near-term problems" (a problem that persists) and providing examples of the sort of serendipitous discovery that doesn't occur in that environment.

Though I'm not sure what to make of the linkage between this change in focus and scientific misbehavior, the article's point on the real cost of the push towards applied research is well-taken. "Questions of how funding is distributed are as critical as how much funding."

Here's the whole thing.

High End Computing Remains a "Priority" in Administration's FY 07 Plans

[Back from vacation. Blogging resumes...]

The Administration has released its annual guidance (pdf) to Federal agencies instructing them on the areas of research and development they should make priorities in their forthcoming FY 2007 budget requests to the White House. The memo, a joint production of the White House Office of Science and Technology and budget gatekeepers, the Office of Management and Budget, "provides general guidance for setting priorities among R&D programs, interagency R&D efforts that should receive special focus in agency budget requests, and reiteration of the R&D Investment Criteria that agencies should use to improve investment decisions for and management of their R&D programs."

As it was last year, High End Computing and Networking R&D remains a priority for the Adminstration, even at the expense of other items within the Networking and Information Technology R&D portfolio. HEC joins Homeland Security R&D, the National Nanotechnology Initiative, Priorities in the Physical Sciences, Understanding Complex Biological Systems, and Energy and Environment as focal points in the Administration's R&D portfolio. Here's the relevant language from the computing section:

While the importance of each of the Networking and Information Technology R&D (NITRD) program areas continues, investments in high-end computing and cyber infrastructure R&D should be given higher relative priority due to their potential for broad impact. Agency plans in high-end computing must be consistent with the 2004 Federal Plan for High-End Computing and should aggressively focus on supercomputing capability, capacity and accessibility issues by emphasizing coordination, leveraging the efforts of all agencies and, where appropriate, use of coordinated multi-agency investments. Advanced networking research (including test-beds) on hardware and software for secure, reliable, distributed computing environments and tools that provide the communication, analysis and sharing of very large amounts of information will accelerate discovery and enable new technological advances. Agency requests should reflect these program priorities by reallocating funds from lower priority efforts. Agencies supporting R&D in these and all NITRD areas are expected to participate in interagency planning through the NSTC to guide future investments. Reflecting the importance of cyber security, agencies should continue to work through the NSTC to generate a detailed gap analysis of R&D funding in this area.

Even though the FY 2006 budget process is still unsettled, this memo gives a good peek at the Administration's thinking for FY 2007. Not surprisingly, the memo implies that next year's budget will likely be as flat as this year, noting that

Agencies may propose new, high-priority activities, but these requests should identify potential offsets by elimination or reductions in less effective or lower priority programs or programs where Federal involvement is no longer needed or appropriate.

So, it will again be critically important that the computing community work with agencies to make sure that the right priorities are struck in this zero-sum game....

CRA Bulletin Now a Blog

For several years now CRA has sent anyone who was interested an electronic bulletin containing links to items of interest to the computing research community. While the content was always useful, the desire to aggregate links and not bombard subscribers with e-mail after e-mail meant that we'd let the bulletin ripen until we'd accumulated enough entries to make it worthwhile to send out. This had the disadvantage of making things a little less-than-timely. So it was time to evolve the format.

Behold, the new CRA Bulletin, now a blog complete with RSS feed for easy subscribing. CRA's Jay Vegso is the curator of the blog. Here's his description for the blog's function:

The focus of the bulletin will be student/faculty demographic, workforce, and R&D information. My intention is to create a source for reliable information, like footnotes, rather than 'breaking news' or editorials. Rather than deal with large reports in a single entry, individual graphs or issues will be given their own entries. For example, Science & Engineering Indicators might have 10 entries, viewable by clicking on the S&E Indicators 'category' in the right-side menu.

Expect frequent cross-links from here to there as Jay comes across more juicy morsels to post. There's already plenty of good content there, like Growth Among Computer/Math Sciences Workforce in the late 1990s, NSF Reports on Academic R&D Expenditures for FY 2002, Close to 40% of Those Employed in Computer and Math Science Occupations Do Not Have a B.S. Degree, and a whole lot more.

Media noticing lack of CS/CE majors

Two interesting stories came through the Triangle (North Carolina) Business Journal over the weekend focusing on the lack of undergraduates majoring in CS and CE. The first one, entitled "Fewer students majoring in industry could lead to labor shortage," notes that CS enrollments in at North Carolina State and the UNC campuses have dropped from 1,988 in 2000-01 to 1,333 in 2004-05. The story was picked up and nationally syndicated by MSNBC. A second story focuses on the lack of minorities entering CS-related fields.

Both stories quote Andrew Bernat and cite the CRA as a key source. Could this be a sign that at least the business media are showing an increased interest in computing research and its effects on the American economy?

Additional news stories mentioning CRA can be found at http://www.cra.org/reports/news/index.html.