What is prompting this post is some recent articles on my cellphone Google News Feed talking about how much faster “low-code/no-code” development is.
I coded my first computer program back in the early ’70s. I have been an interested (mostly) bystander since then.
There have been many attempts to make software development quicker and more efficient. Some have worked. Some haven’t.
Some of the earliest improvements to speed up software development were caused by doing more/most of the software development in “higher-level” languages rather than Assembler. Someone reported that if you had a good optimizing Fortran IV compiler you could hire 4 Fortran programmers for the cost of a single Assembler programmer and get very nearly the same performance out of the code that was created while getting nearly 4 times more software production.
Then along came the idea of “4th Generation Languages” that were designed to allow you to tell your computer “what” you wanted to be done and then “it” spent its time during the “compile” phase figuring out “how” to do it. The results were spectacular in the speed of development but the resulting code wasn’t always very efficient. The original version(s) of FOCUS was one of many examples of “how-not-to” replace a very fast data-entry system in Command Level CICS/Cobol with a much slower rapidly developed equivalent in FOCUS. I heard a report that they had to quadruple the staff using that version of the Application to get enough productivity.
SQL is still the most commonly used and efficient language of this type.
Another idea that has come into and gone out of fashion is the “low-code” or “no-code” approach to software development. This can speed up the development process significantly. The catch is the “tool/interface” you are using has to be able to fully support the solution of the problem you are trying to solve.
When programmers started having to “customize” the code after it had been built by the no-code environment this caused problems until something like “round-tripping” was invented. Up to that point every time you generated a change with the no-code environment you made all your customization(s) disappear so you have to “re-apply them”. This was very inefficient. So the no-code people tried using techniques like “round-tripping” to preserve the customizations. Apparently, there was never a really good way to preserve customizations because no-code systems to create customized development appear to have fallen out of favor some time ago. Now they are apparently back.
One other thing that came into favor and is still a favorite was “COTS” packages that provided limited parameter-driven flexibility for well-understood business processes. The major brand names on this are products like SAP.
So now people are promoting “low-code/no-code” development again. But wait, there’s more. Apparently, they are trying to use Machine Learning and other Artificial Intelligence techniques to leverage up the power of computer hardware and reduce the human element in coding a solution.
I am a great fan of being able to tell my “computer” what to do and having “it” figure out how to solve the problem(s). And while things are getting better I am not sure how soon that this new trend is going to succeed in un-employing countless journeyman software developers.
Tom M
ps. One of the no-code developer systems is called “Divi” and it allows you to create a custom-built WordPress-based website without coding. It is on my TTD list to implement and start using.




