I’ve often wondered what Nobel laureates “do for an encore.” Are they content to do mundane work? Or do they step up their work as public intellectuals? Maybe just enjoy their reign as deans of their fields?
Thus I read with interest yesterday’s Wall Street Journal op-ed, “Government Can Do More to Support Science and Innovation” by Paul Romer, who earlier this year shared the Nobel Prize in Economics, for work done related to the title of his piece. I was shocked, indeed appalled, at this passage:
Take Python, the open-source programming language that has become the language of choice for software developers in the fields of artificial intelligence and data science. Programmers have used Python to power innovation in everything from the detection of gravity waves…to reducing the cost of developing new drugs.
First, many of us in AI/data science greatly prefer the R language to Python; ironically I just wrote on this yesterday in my statistics/R blog. But the much more salient point is that choice of programming language is a matter of taste. For Romer to claim — and I fear, believe — that without Python we’d still be searching for gravity waves and drugs would take longer to develop is absolutely absurd.
Romer is also confusing programming languages with software written in those languages. Any major language is free, but commercial software is written in all of them, including Python. And contrary to another implication of Romer’s piece, federal grants are used to develop open-source software all the time.
Actually, I myself am very big on open-source software. In fact, all the software I use is open-source. But Romer’s ignorance here is alarming.
Upon Closer inspection 
“In fact, all the software I use is open-source.”
I guess you won’t be having problems with retirement. I read somewhere that many retired professors lament not having access to the software they used while being employed by the university.
LikeLike
I can’t read the article behind the pay wall, and from other stuff Romer seems to be just another rah-rah kind of guy, but the fragment you quoted does not appall me, and in fact I have no idea what it is about it that appalls you. I mean, he might have said the same thing about pencils as about Python, but so what?
LikeLike
Romer should keep his day job
LikeLike
A Nobel Prize winner in Economics doesn’t mean that he is an expert in computer languages or software.
At Sac State in the early 80s, a professor told my classmate that Algol was the language of the future. The curriculum was then heavy in Pascal and Fortran. C, which I learned on my own at my 2nd job, didn’t take off at Sac State till I had been gone for 5 years.
From my 35 years experience, I would advise IT professionals to know some or all of the following:
Languages: C#, Java, Python, C++
Databases: Oracle, SQL Server, MySQL
Web: HTML, CSS, Javascript, JQuery
OS: Windows 10, Unix, Linux, Solaris,
These will of course change over time.
Not familiar with R except through your blog. However I recognize the importance of languages that are tailored for specific areas as R is for statistics. From what you have said, I’d clearly recommend it for those doing scientific programming.
LikeLike
> Take Python, the open-source programming language that has become the language of choice for software developers in the fields of artificial intelligence and data science.
I have to wonder if Romer didn’t just read or hear this from someplace such as the article at https://www.netguru.co/blog/how-did-python-become-the-language-of-choice-for-data-science titled “How Did Python Become the Language of Choice for Data Science?”. Other articles, such as one at https://www.zdnet.com/article/the-best-programming-language-for-data-science-and-machine-learning/ give what I believe is the correct view that there is no clear winner. It’s title is “The best programming language for data science and machine learning” with the subtitle “Hint: There is no easy answer, and no consensus either.”
> Romer is also confusing programming languages with software written in those languages.
Agreed. I became interested in Python because of an interest in security and financial programs and Python seemed to have more libraries or, at least, more libraries referenced by online courses. Also, I just couldn’t seem to find enough contracts working in R. However, R seems to have plenty of libraries/packages for artificial intelligence and especially for data science. And I am yet to find the equivalent of R Shiny in Python. As explained at https://shiny.rstudio.com/ , Shiny is an R package that makes it easy to build interactive web apps straight from R. Of course, I’m not an expert in Python and the software and packages written in Python and R are constantly changing. It seems evident that Romer similarly does not have expertise in all of the uses and capabilities of Python and R and would have done well to admit as much and not limit his focus to Python.
LikeLike
Some of the nastiest security bugs are buried in the libraries of Python. Look up the ones on how URLs are parsed. They give you access to usernames and passwords, debug away.
LikeLike
I follow the financial markets and I never heard of Paul Romer. Please consider that I was both a CS and Economics major. Economics is not a science. In the past it was considered a branch of political philosophy, which in no way is a science. If you dissect the economics curriculum, you may be surprised by the following, an economist only needs one semester of college level math, usually Calc I and/or Statistics. This in no way qualifies an economist to evaluate software or complex systems of any kind.
Don’t bother disagreeing with Romer or the NYT. Neither is qualified to address the topic.
LikeLike
I don’t mind any write up or talk about Python no matter whether distorted or not. After I got laid off about 2 1/2 years ago, I began to see that there existed a cottage industry of HR consultants, career consultants, interview experts, corporate flunkies, etc, whose main function was to justify that you (the older person) got laid off because you are like generations behind in the latest technologies and resume building skills–that foreign born engineers know more than you when they leave their equivalent of high school. One expert insisted American engineers were like 8 generations behind the latest and greatest technologies.
And like wow, a language created in 1991 is the coolest thing around. How many generations am I or others behind now? Well, I don’t know Python, but Verilog/VHDL the same thing.
LikeLike
Paul Romer is also confusing creative thinking with sitting at a computer and typing code. While programming languages are good tools, they can be a distraction from creative thinking or even worse – a form of intellectual “crack”.
Newton, Gauss, Euler and Leibniz did not get their great ideas chucking out Python code.Their great mathematical innovations came about by quietly pondering the scientific problems of their day and thinking up new ways to approach them. Einstein had his “thought experiments”.
Romer had fallen for what I call the “dot-com-ization” or “Microsoft-ization” of scientific work. The erroneous and dangerous perception that sitting at a computer typing code leads to technological advancement.
Innovation only comes about by working with atoms and molecules NOT bits and bytes.
LikeLike
Just shows that most people in our govt have no idea what they are doing. Kind of like when Greenspan said it would be a good idea to flood the US with H-1Bs in 1999 in order to effect “wage leveling” and “end the concentration of income”.
That didn’t exactly work out too well.
LikeLike
Mr. Greenspan repeated this opinion in 2009 (before the Subcommittee on Immigration, Border Security and Citizenship) and in 2015 (at a Council of Foreign Relations conference).
This is the man who could/would not see the seriousness of the economic bubble that ballooned under his watch.
LikeLike