Gregory wilson: we know little about how--or how well--scientists use computers. Almost 2,000 people took the time to tell us what they were doing, he says. 71 percent of respondents had a Ph.D. Or equivalent, he writes. Over half of our 1,972 respondents chose more than one category.
Gregory wilson: we know little about how--or how well--scientists use computers. Almost 2,000 people took the time to tell us what they were doing, he says. 71 percent of respondents had a Ph.D. Or equivalent, he writes. Over half of our 1,972 respondents chose more than one category.
Direitos autorais:
Attribution Non-Commercial (BY-NC)
Formatos disponíveis
Baixe no formato PDF, TXT ou leia online no Scribd
Gregory wilson: we know little about how--or how well--scientists use computers. Almost 2,000 people took the time to tell us what they were doing, he says. 71 percent of respondents had a Ph.D. Or equivalent, he writes. Over half of our 1,972 respondents chose more than one category.
Direitos autorais:
Attribution Non-Commercial (BY-NC)
Formatos disponíveis
Baixe no formato PDF, TXT ou leia online no Scribd
tools in every branch of sci- ence, but we know remarkably little A Web-based Respondents’ descriptions of their disciplines were much more diverse. Roughly 150 identified themselves about how—or how well—scientists use them. Do most scientists use off- survey offers as physicists, but no other discipline made up more than 5 percent of the the-shelf software or write their own? Do they really need state-of-the-art clues sample. These figures are necessarily imprecise, since we had to make a lot supercomputers to solve their prob- of judgment calls when coding them. lems, or can they do most of what they For example, should astrophysics be need to on desktop machines? And classified as a separate discipline from how much time do grad students re- over 50 or, in the case of 15 respond- astronomy and physics? If so, what ally spend patching their supervisors’ ents, didn’t answer. These figures are about plasma physics? And how ex- crusty old Fortran programs? consistent with reports about degrees: actly do we count “theological engi- To answer these questions, my col- Seventy-one percent had a Ph.D. or neering”? (In the end, we discarded leagues and I ran a Web-based survey equivalent, with 18 percent reporting that response entirely.) during the last two months of 2008. at least an M.Sc. We were surprised and gratified that When asked to identify their roles, Getting the Answers almost 2,000 people took the time to over half of our 1,972 respondents So what did these people tell us? First, tell us what they were doing. We were chose more than one category (be- respondents work an average of 48 equally surprised by what they told us. low)—which is probably an accurate hours a week, of which 30 percent reflection of how many jobs working is spent developing software and 40 Who Responded scientists actually do. percent is spent using it. They also re- First, a few facts about who answered. Thirty-one percent told us they were from the United States, 20 percent from Canada, and 8 percent from the United Kingdom. Germany and Nor- way came next with 7 percent and 6 percent respectively, while the rest of the world made up the remaining 28 percent. The high representation from Canada and Norway reflects the fact that my colleagues and I are based there, while the low response rate from areas such as Russia and East Asia is undoubtedly due to the fact that we only advertised the survey in English- language channels. Thirty-three percent of respondents were 18 to 30 years old; 35 percent were 30 to 40, and 17 percent were 40 to 50. The remaining 15 percent were
Greg Wilson is an adjunct professor of computer
science at the University of Toronto. His course material is available at http://www.third-bit.com/ swc. Address: Room 3230, Bahen Centre for In- formation Technology, University of Toronto, To- ronto, Ontario, M5S 2E4. Internet: gvwilson@ cs.utoronto.ca
American Scientist, Volume 97
port that these proportions are going er people. When we asked where that ter of the total, while larger programs up—45 percent of respondents say that software comes from, though, they re- account for the remaining 12 to 15 scientists spend more or much more of ported “commercial off-the-shelf soft- percent. To look at it another way, two their time developing scientific soft- ware,” “open source” and “we build it thirds of the programs used by these ware than they did 5 years ago, and ourselves” in almost equal numbers. scientists are less than 5,000 lines long. 70 percent say that they spend more It’s interesting to compare the latter The hardware scientists use is just as or much more time using it. These an- answers with those given for another interesting. Eighty-one percent prima- swers are much higher than we expect- question. Fifty-eight percent of scien- rily use desktop machines; only 13 per- ed, and probably signal that our (self- tists reported that they do development cent use intermediate-sized machines selected) respondents use computers on their own; 17 percent work with one such as departmental Linux clusters, more than the “average” scientist (if in other person, and 18 percent in teams of and a mere 6 percent use supercom- fact there is such a thing). 3 to 5 people, while only 9 percent work puters. This is consistent with their Second, most scientists generate and in larger groups. These numbers are the reports about how they use comput- archive a few gigabytes of data each reverse of what would be expected for ers: Most said that interactive use was year. This answer was more popular professional software developers, who most common, followed by prepar- than all the others together, which were usually work in teams. They also ex- ing and reformatting data, preparing “a few megabytes,” “a few terabytes” plain the relatively low uptake among things for batch processing, and finally and “more than a few terabytes.” One scientists of collaborative tools like ver- systems administration. thing we didn’t ask (but should have) sion control, which most professional As for what occupied the most of was how that data is archived: Is it software developers consider essential: our respondents’ time, coding and de- stored in a Web-accessible database If you expect to work alone, why invest bugging took first place. Planning and with searchable metadata, or on a DVD in tools for working with others? quality assurance tied for second place, stuck in the bottom drawer of some- The prevalence of solo and small- reading/reviewing code came third, one’s desk? Personal experience tells us team work is consistent with another documenting fourth, and packaging the latter is far more likely…. finding. Roughly 38 percent of the pro- software came last. It is ironic to com- Third, most of the software that sci- grams scientists write are between 500 pare this complaint with answers to entists work with is widely used: Only and 5,000 lines long; smaller programs, another question: What “pain points” 10 percent reported that the programs and programs between 5,000 and 50,000 hurt you most? Lack of documentation they rely on are used by three or few- lines long, each make up about a quar- was the number-one answer for more
www.americanscientist.org 2009 September–October
than 40 percent of respondents, and in The three areas in which respond- tists how to use computers effectively as the top three for 80 percent. ents felt they didn’t know as much as research tools. One reason for this fail- Where do scientists learn how to they should were, in order of increasing ure is that commercial software devel- develop software and use computers gap, software construction, verification opment tools and practices often don’t in their research? Almost all said that and testing. Again, this isn’t surprising, fit the needs of people doing explora- informal self-study had been most im- since the whole point of science is to tory research in domains where years portant. Peer mentoring came second, be able to prove that your answers are of training are required to understand with formal instruction at school or on valid--and that requires confidence in the problems being solved. At the same the job trailing well behind. the methods and tools used to get them. time, university science and engineer- To close off, we wanted to find out The necessity of keeping test tubes clean ing departments feel their curricula are how good scientists are at developing and calibrating equipment is drilled already overfull. As a physicist said to and using software. However, self- into students from high school onward, me some years ago, “What should we assessment is notoriously unreliable, and but most are uncomfortably aware that take out to make room for more pro- administering a proficiency test over the we know a lot less about how to ensure gramming—thermodynamics or quan- web would have been impractical. We that software is correct. The fact that tum mechanics?” Figuring out how to therefore asked our respondents to rate there always seems to be one more bug square these circles is, in my opinion, how well they felt they understood vari- to fix only reinforces the feeling. the only grand challenge in scientific ous aspects of software development, computing that really matters. and how important those aspects are. Helping Those Who Need It The results were consistent with Our results can be interpreted in many Acknowledgments answers given to other questions. In ways, but I think two things are clear. This work was made possible by a grant most areas—requirements, design, The first is that if funding agencies, ven- from The MathWorks, Inc. I’d like to thank maintenance, product management dors and computer science researchers my co-investigators, as well as Jon Pipi- and project management—scientists really want to help working scientists tone and Dr. Laurel Duquette, who helped reported that they knew as much as do more science, they should invest with data coding and analysis. they felt they needed to know. This isn’t more in conventional small-scale com- surprising: Scientists are usually their puting. Big-budget supercomputing Bibliography own customers, and as our findings projects and e-science grids are more Hannay, Jo Erskine, Hans Petter Langtangen, about team and program size suggest, likely to capture magazine covers, but Carolyn MacLeod, Dietmar Pfahl, Janice those who develop software are creat- improvements to mundane desktop Singer and Greg Wilson. 2009. “How Do ing small programs for their own use. applications, and to the ways scientists Scientists Develop and Use Scientific Soft- ware?” Proceedings of the Second Interna- Skills relevant to large projects done for use them, will have more real impact. tional Workshop on Software Engineering other people are therefore unlikely to My second conclusion is that we’re for Computational Science and Engineer- loom large in their minds. not doing nearly enough to teach scien- ing. New York: IEEE Press.
Summary: 12 Months to $1 Million: How to Pick a Winning Product, Build a Real Business, and Become a Seven-Figure Entrepreneur by Ryan Daniel Moran: Key Takeaways, Summary & Analysis
The Star Interview: The Ultimate Guide to a Successful Interview, Learn The Best Practices On How to Ace An Interview As Well As Crucial Mistakes You Need to Avoid In Order To Land the Job