SAS programming language is #5 (for Jobs) in IEEE Spectrum 2024 ranking

[ChrisHemedinger]

Hi SAS Enthusiasts,

If you follow tech like I do, you've come across a variety of "top programming languages" lists. These annual or biannual lists always capture my attention, and this year I was pleased to see that SAS earned a higher placement on the IEEE Spectrum 2024 list of Top Programming Languages. SAS specifically shows well in the Jobs ranking (#5), reflecting the demand in the market for SAS skills.

I want to know what you think about this list and the SAS placement: do you feel it reflects the importance and usefulness of SAS in your professional life? I would love to hear about how SAS skills opened career opportunities for you. 

While you're thinking about your reply, I'll share my take. I have studied these lists for years (from IEEE, from TIOBE,from StackOverflow and others). Most of the criteria that is used to compute rankings, such as Internet searches or GitHub repositories, will always undercount SAS usage. Most SAS customers work in industries where they cannot share their code on GitHub -- their SAS work is tied up too much in the secret sauce that makes their company productive. And as far as the Internet goes, most SAS information sharing happens here in these communities or in conference proceedings that are hosted on our websites. But when you look at job opportunities, there is no denying that SAS skills are critical in many large organizations -- and the job boards reflect that. (See more about IEEE Spectrum's methodology here. They used R for their analysis...oh well.)

The high ranking of SAS, with Python and SQL, tells a bigger story, I think, about the importance of data and data science as a discipline that is still very code-intense, despite the "low code" movement to bring these capabilities to non-programmers.

Tell us -- what do you think of SAS' placement here, and how do you see its unique value in your programming toolbox?

[AndersS]

Hi! When comparing programming languages there are several aspects to consider:

I - programming environment:  Editor,  Documentation, Documentation aids, Debug tools, OS Platforms, data storage, how to move programs and data to another OS platform.

II - How good are they. Perhaps best judged by finding say 5-10 problems, that shall be solved.
      Then measure - coding time, cpu-time, memory usage, How easy is it to implement a small change in the
      specifications.

III - Facilities in the language. Like Formats, functions etc.

(I do not know very many languages - Fortran, SAS + Some languages where I found and corrected the errors made by users)

/Br Anders Sköllermo

[ChrisNZ]

SAS also shines by its backward compatibility. Take a program written for mainframe 40 years ago and it will likely run as-is today under Linux.

[AndersS]

Hi! Almost, almost correct. An option must be taken away. Something like
TSOPTION.
When trying to run the program you will get an error message. VERY EASY to
fix when you know about it.
/Br Anders Sköllermo

[Quentin]

As background: I used to say I was a programmer who only knew one language (SAS), until it was repeatedly pointed out to me how many languages there are in SAS. I've always been vaguely interested in these lists, as in enough to read the occasional LinkedIn or blog post, but not enough to worry about it too much.

I don't find the general programing language lists very useful.  As somebody who uses SAS for data management and analytics, it never worried me that there were many more .NET or javascript programmers out there than SAS programmers.  There are more doctors, car mechanics, and restaurant workers also, and the comparison feels almost as irrelevant.

I do find the comparison between SAS, R, and Python interesting, and keep 1/4 an eye on those trends.

Not surprisingly, Python is well in the lead in terms of numbers, but probably 90% (?) of Python is not for analytics or data management, it's a general programming language.

But still, I think the relative growth/trends of R, SAS, and Python are interesting.  And I think focusing on the job market is probably the best metric available.

I was actually surprised how much lower R scored than SAS.  But as I think about it, R feels like a more limited language than SAS. I feel like R (because it came from S/S+) is basically still a statistical language.  And while it has AI/ML packages, python is way ahead in AI/ML.  And I'm not sure I'd want to serious do data management  / data engineering in R.

Within the field of analytics, I think one of the reasons SAS shines is that it has packages/solutions that do everything you need:

• Data management/Data engineering
• Traditional stats/biostats
• AI/ML

Whereas R and Python probably only do one or two of those well.

That said, I'm a big fan of SAS's strategy to embrace open source, both calling R / Python / etc. from SAS, and allowing R/Python/etc to call SAS analytics.

When I join occasional R community calls, in the past it felt like R wanted to be a "SAS-killer".  And probably some of that community still does.  But now on those calls I feel like the R community is sometimes super-defensive about the rise of Python.  I see a lot of python-bashing  in the chat, as if they're worried that R is no longer the revolutionary open-source language, and may itself be over-thrown.  It was super-interesting that the company R-studio changed their name to Posit and embraced Python.  We'll see how that goes. 

Meanwhile at SAS user group conferences, it's been great to see the growth of R and Python.  I feel like not just SAS Inc, but also the SAS community, has benefitted from truly embracing open-source languages.

[AndersS]

Hi! Very well written. /Br AndersS