Simple not simple

For readers in a hurry

  • Low-code and no-code platforms make it easier for software developers to implement applications. However, they require a sound understanding of databases, business logic, interfaces, user interfaces, security, operation, architecture and error analysis.
  • Every low-code/no-code platform generates classic source code for professionals in the background, even if the composition of the application is largely graphical. If you want to create sophisticated applications or are faced with errors, professional know-how is essential to solve the problem.
  • Every software development tool or IDE uses or will use low-code elements in the future, unless it relies on AI-generated source code to increase developer productivity. Graphical development tools are always more intuitive when it comes to user interfaces and databases.
  • The simpler the platforms make development, the smaller their current range of functions and the greater the lock-in effect. Private equity-financed companies in particular will sooner or later have to fleece their customers in order to realize the profit expectations placed in them. Interested parties should take this into account when choosing a platform.

Tip to try out

If you want to develop appealing application interfaces and rely on a modern and feature-rich platform, you should take a look at FlutterFlow should take a closer look. It allows software developers to develop appealing user interfaces for all end devices and combine them with backend functionality. In contrast to Flutter, FlutterFlow offers numerous low-code-based accelerations and pre-built elements, but requires the solution to be operated in the FlutterFlow Cloud. Nevertheless, it is a great option for rapid development in the context of SaaS applications.

Oracle APEX for a quick win

The low-code market is facing major upheavals. Leading providers such as OutSystems and Mendix have lost the focus of IT companies due to the developments surrounding AI and OpenAI. What's more, their pricing model is not convincing. Too expensive to get started, too opaque for scaling applications. Who wants to estimate user screens, interfaces, database tables or function blocks for several years as part of licensing before they have even implemented an application? This takes the fun out of license negotiations, even if the platforms are consistently good and usable.

Oracle APEX takes a different approach here. Here, computing power, i.e. CPU, RAM and storage space, is licensed. If an application is little used, you pay less; if the application is heavily used or very large, you pay more. This is understandable. Oracle APEX is also significantly cheaper in absolute terms - especially when getting started with low-code development - than OutSystems or Mendix.

The current disadvantage is that Oracle APEX, which integrates seamlessly into Oracle's Autonomous Database, requires a relatively large amount of PL/SQL code and JavaScript in order to implement more complex backend and frontend logic. OutSystems and Mendix are much more graphical in this respect, even if you have to think like a software developer on both platforms (e.g. iterations, stringifies, set theory for tables, return values, data types and much more).

Nevertheless, powerful and visually appealing applications can be realized in a short time with Oracle APEX without having to get lost in the depths of JQuery, Angular, Vue or React. This is exactly what we have done, asyou can read here .

Babylonian linguistic confusion

Let's make things a little more complicated now: In an international environment, many applications have to be adapted for a variety of languages - including the one above. In the meantime, UTF-8 has established itself as the character set encoding so that letters with an accent such as an ö in German or an é in French are displayed correctly. That's the theory!

The relevant question is: Can you rely on it? The short answer: unfortunately not.

It can happen that a German text suddenly becomes an ISO-8859-1 encoded text that the target application cannot read. Instead, umlauts are represented by special characters to the annoyance of the user. Now it is time to find the cause. And this is exactly where in-depth know-how is required. Low code or no code. Hexadecimal byte sequences with different character set coding tables must be translated manually where the errors occur in order to identify the actual encoding used based on the special characters. The exact location of the error must then be checked along the system steps. Finally, its cause must be found and rectified.

A lot of work that requires concentration and, above all, expertise.

Why is this important?

Low-code sounds nice as long as the use case is simple and there are no problems. As soon as the application is used in complex environments with many peripheral systems and numerous participants, a low-code platform can only survive if it is used by professionals. Then it also retains its main advantage: the fast and relatively secure development of visually and functionally appealing applications.

This is a welcome simplification and relief for the developer, especially for applications that consist of numerous user interfaces, business logic, database and interfaces.

So it's all sunshine after all? Unfortunately not.

Reliability

Anyone wishing to purchase a firearm in Germany is subject to a reliability check. This is intended - with varying degrees of success - to prevent violent criminals from gaining access to these objects.

The reliability of the provider must also be checked when choosing a software development platform. Otherwise you could be in for some expensive surprises.

While the technical evaluation is quite successful, the economic and organizational reliability of low-code providers is not so easy. IT has been discovered by banks, speculators and other moderately productive economic forces as an ideal tool for making money. Hardly any customers understand what is going on under the "server hood". This is an ideal prerequisite for exorbitant prices.

Almost every US startup is financed by private equity and has thus become a means of hopefully cashing in on the numerous PE firms and their fund investors. The money - which for a long time was easy to borrow from central banks at zero percent interest - is and was also needed to make major investments in product development. But what does this look like in times of rising interest rates?

This is where the wheat is separated from the chaff.

The golden rule

He who has the gold makes the rule. In our case, that is the customer. We don't want to dwell on the short-term view of the financial industry. Rather, the question is, how do you counter this when you have found a good low-code platform?

The answer is very simple: choose an alternative if the economic cornerstones are not conclusive. These would be

  • Is the pricing model transparent and comprehensible, i.e. are the price drivers measurable, predictable and constant?
  • Does the sales department or the provider in general have a high fluctuation rate, so that there is no continuity among contact persons and service providers?
  • Is the licensing model changed regularly, or are "new products" regularly added that have to be licensed separately in order to realize disguised upsells - especially in the case of innovations?
  • Does the management constantly talk about the planned IPO or M&A, for which the "bride" has to be regularly financially spruced up and rigid payment modalities are to be anchored in contracts like a sacred cow?
  • Is the product roadmap unclear or intransparent, so that, for example, core functionalities are neglected in favor of new fashions (keyword AI) and suddenly function unreliably?
  • Is customer service "on the ball" or sluggish in its responses? Is it hiding behind chatbots or uninformed outsourcing companies?

The list could go on and on. Why go into so much detail? Because in case of doubt, other mothers also have beautiful sons. If your gut feeling during the fact check tells you that they might not be reliable, then stay away - or at least be careful and persistent.

Where there's smoke, there's fire.

Time is working against hasbardeurs

Anyone who sits on a high horse will fall painfully low at some point. This can also be observed with low-code platforms. Arrogant management usually overlooks the technical substitutes. In this case: artificial intelligence, especially large language models. Because software development is always language-based, even if the language is only moderately attractive to humans: JavaScript, Python, Java, C#, C, SQL and many more. Low-code platforms also convert their graphically created programs into these high-level languages so that they can then be compiled or interpreted in machine-readable form as normal.

It's not rocket science.

Low-code has not yet established itself on a broad front. It dominates traditional software development. Not even Microsoft has made a recognizable push towards low-code with Visual Studio. The PowerPlatform has been rather neglected.

We are not surprised, as the low-code platforms are still far too expensive today and their license models are not predictable enough, while subsequent export of created software to a native environment is not possible or only possible with great effort.

Not good starting conditions.

At the same time, an elephant entered the room completely unnoticed: OpenAI and GPT. With its codex model geared towards software development and additional extensions such as Copilot for GitHub or the competitor AWS CodeWhisperer, AI suddenly accelerated traditional software development enormously. Unit tests are created automatically, error searches are carried out, source code is documented automatically and entire function blocks are written as if by magic. This immediately raises the question: why do I still need a low-code platform if I need competent developers for demanding applications anyway and the speed advantage is significantly reduced, as is the dependency on the platform?

This is the question we are facing today.

The next 2 to 3 years will show us the way.

Late regrets, everything will be fine?

Will the low-code providers reflect on their position in the cosmos of software development and adjust their prices so that they are affordable as a development platform across the board and become a reliable partner for their customers? We are curious! Does this mean that low-code platforms should be avoided today?

No, not at all! If you are aware of their situation and behavior, you can take this into account in your negotiations and stoically talk them down. If you can't come to a satisfactory solution, you take another one. There is plenty of choice, and most decisions are not made for eternity.

However, it would be wrong to rely on the promises and assertions of low-code sellers that are not contractually enforceable and enforceable. This is because, in the case of unreliable companies, these vendors are usually gone again when a problem becomes virulent.

The "rats" leave the boat.

For this reason, we use several low-code platforms in parallel and check before each project which platform is not only technically but also economically suitable. If something can be implemented with little additional effort using architecturally simple pro-code (i.e. JavaScript, Python, Java, etc.), this alternative is presented to the customer for a decision on an equal footing. Clever risk and portfolio management should not only be taken to heart by companies in the financial sector, but also by every value-adding entrepreneur.

Never put all your eggs in one basket.

About Business Automatica GmbH:

Business Automatica reduces process costs by automating manual activities, increases the quality of data exchange in complex system architectures and connects on-premise systems with modern cloud and SaaS architectures. Applied artificial intelligence in the company is an integral part of this. Business Automatica also offers automation solutions from the cloud that are geared towards cyber security.