programming – Looking Forward

In last week’s post on the Coding Craze, I referred to the continuing
reduction in the need for low level coding – even as what is defined as
low level continues to rise and be more abstract, more separated from
the machines that the software controls. I even noted the work in
artificial intelligence to create programs that can program.

All
of this is a reflection of the fact that pure coding itself is only a
small part of what makes software successful – something that many
coding courses don’t discuss.

Many years ago in the programming
world, there was a relatively popular methodology named after its two
creators – Yourdon and DeMarco. While it has been mostly been
remembered for its use of data flow diagrams, there was something else that it taught which too many coders don’t realize.

There
is a difference between what is logically or conceptually going on in a
business and the particular way it is implemented. Yourdon asks
software designers to first figure out what is essential or as he put it:

“The
essential system model is a model of what the system must do in order
to satisfy the user’s requirements, with as little as possible (and
ideally nothing) said about how the system will be implemented. … this
means that our system model assumes that we have perfect technology
available and that it can be readily obtained at zero cost. [Note: this is a lot closer to reality today than it was when he wrote about zero cost.]

“Specifically,
this means that when the systems analyst talks with the user about the
requirements of the system, the analyst should avoid describing specific
implementations of processes … he or she should not show the system
functions being carried out by humans or an existing computer system. …
these are arbitrary choices of how the system might be implemented; but
this is a decision that should be delayed until the systems design
activity has begun.”

Thinking this way about the world operates
and how you want it to operate is the start of software design.
Software design really has two, related, meanings – like C++
overloading.

First, there is the design of the architecture of
the software and overall solution. Diving into coding without doing
this design is what leads to persistent and embarrassing bugs in
software. The internal design is also necessary to avoid spaghetti code
that is hard to fix and to improve performance, even in these days of a
supposed abundance of compute resources.

Second, there is the design of the interface that the user sees –
with all the things to worry about that we associate with the “design
thinking” movement.

(One way of planning software is to imagine
the software designer is a playwright, who is responsible for all the
parts of the play aside from one, the user’s part. I guess this is more
like improv than a play, but you get the idea 🙂 )

So maybe in
addition to the coding class, the wanna-be software developer should go
to improv or drama school. That’s a more likely path to knowing how to
generate the WOW! reaction from users that makes for software success.

[http://njacknis.tumblr.com/post/141545808566/beyond-the-craze-about-coding
]

A computer coding craze has taken over the country. Everywhere you turn, public officials from President Obama on down in the US and around the world seem to be talking about the need to train folks in computer coding.

Governors
and Mayors are asking their school systems to teach students how to
code instead of learning other subjects. Many people who had little
previous interest in computers or software – except as blissfully
ignorant users – have signed up for, often expensive, courses on
programming.

It’s not just in California or Seattle or Austin
(pictured), but back in traditionally less high-tech places on the East
Coast as well. Recently there were stories about coding classes in the
Borough of the Bronx in New York City and as far south as Miami.

There
may be good reasons to take these courses. A bit like the courses that
schools used to teach about how the car combustion engine worked,
learning to program may help people better understand how computers
sometimes operate.

As with any creative activity, at the start,
you can get a sense of accomplishment when witnessing your software
creation come to life — once most of the bugs are eliminated 🙂 You can
even reprise this feeling under special circumstances later on in your
career. But much of the work of coders can, in the long-run, become mind
numbing.

The opportunity to design and create a great new app is
like being invited to paint the Sistine Chapel. But the more frequent
opportunities are like being invited to paint someone’s apartment.

Don’t
get me wrong. As a long time software developer myself, I can say there
are many satisfactions for developers who have both the knack and
passion for software. But people who don’t have those attributes and
just do it like any other job will be frustrated too easily.

And, honestly, even the positive side of life as a developer is not what is primarily driving this coding craze.

Much
of the interest by public officials (like the governor of Arkansas in
the picture) as well as the people enrolled in coding classes is based
on their belief that these courses make possible employment
opportunities that will endure for decades in a world in which
traditional jobs have been automated or shipped overseas. Will it?

Surely, some people are going into coding just for an immediate bump in short-term income. Studies
on the relatively new phenomenon of coding bootcamps seem to support
this notion – that is for the 65% of students who graduated and are
working in a programming job. Even in those cases, the best results were
for students who graduated from the more expensive and selective
programs.

Yet, on balance, count me as a skeptic. I think this
craze is, well, crazy. In the long run, I don’t think coding courses for
the millions will lead to the affluent future and lifelong careers that
many proponents envision.

First, as I’ve alluded to, these are
not jobs that everyone who is learning to code will find satisfying. We
may be too early into this craze to know how many people go into the
field and last for more than a short while, but I’d expect the dropout
rate to be high.

Second, there is the low level nature of what is being taught – how to write instructions in a currently fashionable language. While most of the coding courses focus on currently popular languages, like Ruby and Javascript, many of their students do not understand how popularity in languages can come and go quickly.

Some
languages last longer than others do, of course. Through sheer inertia
and unwillingness to invest, there are still some existing programs
written in old computer languages, like FORTRAN and COBOL. But there
aren’t that many job openings for people coding those old languages.

Wikipedia lists a variety of languages that have been created over the last three decades, approximately one a year:

1980 C++
1983 Ada
1984 Common Lisp
1984 MATLAB
1985 Eiffel
1986 Erlang
1987 Perl
1988 Tcl
1988 Mathematica
1990 Haskell
1991 Python
1991 Visual Basic
1993 Ruby
1993 Lua
1994 CLOS (part of Common Lisp)
1995 Ada 95
1995 Java
1995 Delphi (Object Pascal)
1995 JavaScript
1995 PHP
1996 WebDNA
1997 Rebol
1999 D
2000 ActionScript
2001 C#
2001 Visual Basic .NET
2003 Groovy
2003 Scala
2005 F#
2009 Go
2011 Dart
2014 Swift
2015 Rust
2016 ???

So if all they learn today is the syntax of one language and lack a
deeper education, they may find that one skill to fall out of favor.

Indeed,
many of those students aren’t even being taught about the different
kinds of programming languages – even classes of languages vary in
popularity over time.

Instead, they are usually learning imperative languages, especially with a focus on low level procedures.

It
is also not clear that the popular languages are the best ones to even
teach basic coding, never mind understand software more generally.

Even
the idea that any language is good enough to educate students about how
computers work is misleading. Different classes of languages lead to
different ways of thinking how we can represent the world and instruct
computers.

And finally, the trend in software, in fits and starts,
has been to reduce the need for low-level programming. Originally, it
was a move away from “machine instructions” to higher level languages.
Then there were various tools for rapid application development. Today,
there is the Low-Code or even No-Code movement, especially for Apps.

You’ve
heard of the App Economy, another part of the promised job future?
Putting aside the debate as to whether the app phenomenon has already
peaked, with these low-code tools, fewer coders will be needed to churn
out the same number of apps as in the past.

And then over the horizon, computer scientists have been busy “Pushing the Limits of Self-Programming Artificial Intelligence” as one article states in its title.

Finally,
with this background, pure coding itself, even in past years, was only a
small part of what made software successful. And a successful long term
career in software requires an understanding of what goes beyond
coding.

But this is enough in one post to get many people irked, so I’ll save that for a future post.

[http://njacknis.tumblr.com/post/141088761737/the-coding-craze]

Tag: programming

Beyond The Craze About Coding

The Coding Craze