No-code and low-code, VPL and declarative programming models – what they offer, and what they offer not?
Is it possible, at this day an age, for everyone to create software for their own needs? Are the works involved, regarding the multiplicity of available universal IDEs, frameworks, virtualisation and containerisation tools, being more and becoming more simple, or does it result with a skyrocketing entry level, similar to the learning curve?
The answer to the question may be supported by turning attention to the trends that – at least in the principles – are supposed to change how we think of software development. What do the no-code and low-code development methodologies offer at the end of the second decade of the 21st century, including visual programming languages and declarative programming?
Software made in the creator
The dreams of developing software and expanding software to be made available via a transparent, efficient wizard and a creator for dummies, have accompanied us for years. Most often, the vision is reduced to setting a necessary programme by means of already prepared components, obviously in the GUI mode, with a schematic graphic representation. In this variant, programming would be reduced to actions resembling the assembly of blocks – as if the prefabricates in a particular configuration perceived as such, could respond to every need.
Not much experience is required to realise, that this vision is far from being fulfilled. Interestingly, the forecasts regarding parting with the monolithic software architecture have proven true to a degree. Currently, based on micro-services, we actually build even the specialised tools based on a division into smaller processes. However, it has little to do with the automation of the process of combining the aforementioned services – their integration may be performed with success due to a team of specialists, that have better knowledge in the DevOps methodology, rather than browser wizards.
No-code and low-code development
It would be erroneous to discredit all no-code and low-code development initiatives in the beginning, i.e. those that entirely exclude working in a traditionally perceived code or reduce the necessity of its knowledge to its bare minimum. Tools like those adopted by Google AppSheet deserve attention, due to which small enterprise may successfully automate works related e.g. to filling forms.
One cannot say, that the no-code/low-code development tools have no educational value. Such approach is rather universally applied in gaming engines in combination with visual programming languages. Particular components based on the use of VPL can be found i.a. in Unity, Unreal Engine, CryEngine or opensource Gogot. The mechanisms based on using prepared visual components have a wide variety of applications in declarative tools for designing user interfaces, with the best example being Apple SwiftUI.
Clash with reality
However, one should not cherish illusions – if you need a well-established native, pseudo-native software or even software that works in a browser environment, that is adjusted to the specified needs of a particular organisation, the NCDL/LCDL tool in the most majority of scenarios will not be sufficient. Without a doubt, the elements of the declarative model of programming will be used much more willingly, e.g. due to the comfort that they provide in terms of designing UI/UX, but it would be difficult to expect in the nearest future, that the high-efficient code in the back-end will be constructed of blocks, just as the participants of bootcamps addressed to children, create their simple games in Scratch.
Flexibility, complementarity, and the efficiency of the operation of a professionally managed, and experienced DevOps team, will, without a doubt, remain a value that is difficult to overestimate, offering something inaccessible to the no-code/low-code approach – a high quality of a highly specialised solution strictly adjusted to the individual needs of the end user.
