Tyler Jensen’s Post

One of the first no-code products that crossed my path. I even bought a copy long long ago. Perhaps in 1997, I think.

Note to commenters: Your comments are welcome, pro or con, but self serving spam with links to promote your own product will be deleted and I will block you.

Simon Schaefer posted a share and asked some questions. I brushed him off at the time because of a personal distraction. With that out of the way, I've tried to answer his questions properly there and want to share them here with my answers in the main post for those who are interested. See my replies to this introductory comment.

I have to be honest. The attention this post has garnered has surpassed any other post of mine by a long shot. I'm quite surprised by this. I appreciate everyone's comments and interest.

This is kind of already the case though, even if you write the code yourself. The operating system you run the code on is “not your code”. The software libraries you download from the internet and use in your software code are “not your code”. The tools you use to deploy your software and set up CI pipelines (Jenkins, Nomad, Kubernetes, Docker) are “not your code”. Heck, even your development tools like formatters, linters, coding editors, even the language compilers and interpreters themselves are not your own code. I think the big things you sacrifice with no code solutions are customizability, and, assuming you have a halfway decent engineering team, quality. But these days, code or no code, we’re borrowing and trusting each other’s code left and right. I think as AI matured in the next 5-10 years, we’ll see some truly amazing advances that result in no-code solutions that actually do well enough for the average enterprise. As a software engineer, I’m still building my software development skills, but I’m also developing other skills anticipating that I might one day in the near-ish future find my career replaced by robots. It’s always good to look ahead and have a plan B.

I doubt there’s a person on this extended thread who doesn’t have a tool in their stack that’s low-code (given the ubiquity of e.g. SSIS). My opinion is that this isn’t about “low code” but about appropriate levels of abstraction. Modern programming languages are “low code” compared to writing in assembly. Relying on someone else’s library in a full-code solution has very similar failure surface. I think the current crop of low-code tools have their place- and if they deliver a few years of business value, for stuff that is almost inevitably going to change anyway, then they make some sense. Especially on the side of things like answering analytical questions (the Alteryx’s of the world), where change and variety make coding things from scratch daily every time a non-starter.

I agree with you on one point: the two vendors made a mistake trying to sell their solution to you, a person that actually writes code for a living. This is the reason why I perceive your argument somehow biased. I do not think there are people who really believe that a no-code platform could completely replace developers' work and if those vendors tell you that, well, maybe they should devote their time to other things. In my opinion the point of no-code platforms is to encourage digitalisation at any level and in any size company, even those that couldn't afford some pretty expensive tailored made solutions. In the end digitalisation is all about increasing productivity, streamlining business processes and speeding up operations. Now, as long I can get that while keeping protected my business data, who cares about what code I am using? I am not interested in knowing who wrote it, if the code is custom or there are other individuals/companies that are using it, as long as it does what promised, works smoothly and meets my business needs, I am happy with that. At this point it turns out to be just a matter of money, and no-code solutions definitely result more sustainable under this aspect.

In more generic terms: every platform you use is code from others. Software stacks may be application-logic at the top, but underneath it is platform-logic all the way down. Each platform is geared towards supporting certain logic to be 'coded' and deployed in it. Business do not write their own operating systems, databases, they use a lot of packaged code from others (cough — log4j — cough 😉). However you program (in) your platform. Programming with point and click is still programming, even 'configuring' is programming, point-and-click it is just severely limited in its expression and as such is only a silver bullet from the perspective of those that do not understand IT. But that does not mean it is by definition useless, but I agree with the judgement that NC is yet another over-simplistic hype. Comparable: At some point a company decided to replace 5 self-coded application by a new shiny off-the-shelf system. They thought they were rationalising their landscape. They ended up with a platform in which they had to write 5 applications, but now with proprietary languages. That was not all bad, using the platform had advantages, but rationalisation it was not. I allowed, I'll post the link to that story (no product involved).

I'm not big on the name, but isn't it just another step in the evolution? Is it really that different from the advancement to an IDE from command line?? Those controls that you drag and drop in the IDE are written in someone else's code? In fact, aren't the commands for strongly typed languages written with someone else's lower level code? I suppose it all comes down to your desired control for development and the tools manufacturers you trust. When MS came out with windows and then they're suite of development tools for it, I sure as hell didn't trust MS as I was reporting bugs to them within a week or 2 of using they're new development tools during the early 90s, but it was all we had to work with.

Sure, though I hope they did not actually express anything about trying to remove pesky code monkeys? As in, I like your straw-man sir, but does it come in hay too? I think a valid point is that some no-code products make claims for a future they do not know much about and indeed existing attempts have come and gone, with little of those promises becoming reality. I know you are likely accurately mirroring the pitches you were given in your text, but I think it is nuanced. What about a person who leaves their job with enough savings to create a small runway for themselves to work on their dream project. Not enough money to hire a coder, no investments yet. But, the MVP can just about be built with a no-code platform that shall remain nameless, that has a simplified database integration. This person all of a sudden gets started and eventually replaces the product once they have some success, with actual coders. I mean, in this (true story by the way) case, no-code served its purpose. Plus it is also a matter of perspective, I do not see why no-coders would claim it is their code, because, you know, NO code, nothing to claim. But can I show you a project and say: Here look at my code, when I use an arbitrary amount of libraries?