https://futureofcoding.org/ logo
#linking-together
Title
# linking-together
w

Walker Griggs

06/27/2023, 11:00 PM
Everyone excited for Fortran 2023!? Only mostly joking, but I just noticed today that fortran is scheduled for a new standard release next month. I got a big smile at some of the proposed imporvements • Trig functions that work in degrees • Using integer arrays to specify subscripts and section subscripts • Enum types
2.4 US 22. Conditional expressions and arguments Conditional expressions, expressions whose value is one of several alternatives, are added. A simple example is value = ( a>0.0 ? a : 0.0)
Fortran now supports ternary conditions?
k

Konrad Hinsen

06/28/2023, 6:35 AM
Nobody ever gets excited about Fortran, and that's why it exists. It's boring technology that you can rely on. People get excited about what they are doing with their Fortran programs, happy for not having to deal with tech churn. A useful lens for understanding Fortran is Stewart Brand's pace layers . Excitement is about the fastest layer, which Brand labels "fashion/art". Fortran is in the infrastructure layer. It's as exciting as highways and sewage systems.
k

Kevin Greer

06/28/2023, 3:31 PM
It would have been a good investment in your time if you had learned FORTRAN in 1957 (provided you wanted/needed to code in FORTRAN for 66 years).
I was talking to a COBOL programmer and I assumed that the language had probably evolved a lot over the years, but they assured me that it hadn't and had remained relatively stable and unchanging, at least over their career.
w

Walker Griggs

06/28/2023, 4:16 PM
An irc contact turned close friend did their entire phd in fortran99 because that's what their fluid dynamics library was written it. I read through some of the systems and (aside from the math I didn't grasp) it was seriously readable
I love seeing it supported and invested in (at whatever speed). The world may well crumble without it
k

Konrad Hinsen

06/29/2023, 5:58 AM
@Kevin Greer It's indeed interesting to compare COBOL, Fortran, and Lisp, the three oldest languages that are still widely used. COBOL has hardly changed, Fortran has evolved significantly, but with a strong emphasis on backward compatibility, and Lisp has always continued splitting into dialects that have taken lots of different directions. I see this mostly as Conway's law in action, the languages reflecting their communities: for COBOL, banks and insurance companies, for who stability is top priority, scientific computing based on mature theories for Fortran, and builders of experimental software for Lisp.
k

Kevin Greer

06/29/2023, 1:42 PM
Makes sense and I think this points to what is both Lisp's greatest strength and weakness: its versatility and mutability at the cost of being able to develop a community around an actually common standard.
Maybe the curse of powerful languages.
w

Walker Griggs

06/29/2023, 4:33 PM
I have little context on the modern iterations of these languages and am trying to attribute some reason as to why they split in popularity. The few thoughts that came to mind: • Lisp is a standard and the dialects can be more flexible / meet modern use cases. • COBOL found a home on mainframes which are rarely used outside of banking • Fortran found an early footing in academic systems and has only been a stable, reliable base for scientific computing. Also performance Do any of those points hold water?
k

Kevin Greer

06/29/2023, 6:14 PM
COBOL has very little competition, in that it is a language actually designed for business applications. Java was designed for set-top boxes, javscript for light scripting of webpages, Python for education, C for OS system programming, Smalltalk also for education. Given that the majority of the world's code is for business applications, this is a remarkably underserved area. COBOL's rich record support makes working with structured data much easier than something like C's anemic structures. FORTRAN also benefited from rich standard libraries like https://www.netlib.org/linpack/
k

Konrad Hinsen

06/29/2023, 6:53 PM
@Walker Griggs All three have official standards, which have multiple highly conforming implementations. It's really the application domains and the associated user communities that make the difference.
8 Views