The thing is: that takes a lot of design and planning. You even agree with that. While the current design philosophy seems to be: quick and dirty throwaway apps. Which either run in a browser or on your phone, or both. Yes, that's something else than business software, right? Or is it?
With buisness software I meant the typical buisness to buisness software, where small and medium sized developers are contracted by a few companies to build special purpose software. This is still a large part of software development.
And there the big part is functionality, the reason the customer buys this software is because there is no software that fully satisfies their needs out there, and they want something thats either custom build, or an existing software that is customized for them. Usually characterized by a small pool of customers who pay a lot of money to get the software they need.
On the other hand you have what I call "end user applications", which are readily packaged off the shelf software anyone can buy and/or download by anyone, often with lots of competition. These are your tax software, photo or video editing software, music players, etc. There you have a very little time window to convince the customer. And looking shiny is a good way to do this, and web tech is really good to easiely create very fancy graphics and animations (after all it's what the web is built for, presenting things in a pretty way).
But this does not say much about the internal project and development structure. Such projects can be staticall planned through (e.g. with complete design documents and requirements analysis and so on from the beginning), or could be more agile, with regular checkups, both with developers and customers, regular prototypes and user tests, etc. Depends heavily on the sector. For example when the customer base was the automotive and avionics industry (I have worked on one software which where they were customers, even though it was not the actual system control software, but I had some contact with people from that area), everything needed to be extremely rigorous planned beforehand, with a clear decision chain and responsibilities.
While in other areas, the customer might themselves not really know what they want and you need regular meetings and user tests to find out what is the best way to approach the solution. This is inherently not as plannable as the former.
Both kinds of software can fall in any of these categories (or even somewhere in between).
How many companies still produce their own in-house software? Less and less. Software should be a service and belongs in the Cloud, right? The big software providers, like Microsoft, wholeheartedly agree. They cannot produce decent software anymore, anyway, and subscriptions bring in money every month, no need to produce something better every year.
If I remember correctly, here in Germany 98% of software developers work for small and medium sized buisnesses. And it's an ever growing industry. Sure "cloud native" is one of the buzz words thats currently popular, and espeically when you go into the industry 4.0 stuff you will see alot of that, but I don't see it as that dominating. Infact there are a lot of companies around who do not want to have all their customer data in the cloud, and if so they usually use "private clouds" (i.e. on non shared servers).
The subscription model is not so popular with buisness software but very popular with end user software but also mostly from big companies like Microsoft or so, the reason is that this buisness model while being more predictable with respect to income is more risky for new services (e.g. Spotify has in it's 15 years it exists not made a single cent in profits, 2022 was the best year so far when they "just" lost a bit more than 220 million USD)
So you often see these subscription models with software that already has a large user base (adobe suite, office, etc.). Because what do you want to do, stop using Office? Good luck with that. Note that Office is still sold as a pay once software, because while the subscription model works well for end users, buisnesses don't keep up with that and want to "own" their copy.
But thats also a completely different story, many of these subscription softwares are Desktop software, Office and the Adobe Suite are still written in C++ (even though there is now are web versions, but they often are just a lite version and may even be just a webassembly port of the existing native code).
So you can have cloud native subscription based software, thats still a native desktop application, that just has some connections to the cloud (e.g. as storage, to check licenses, etc.).
So I don't see this as a contradiction
What runs on the server is mostly a collection of databases and micro-services. Written in the popular languages, as promoted by Big Tech. Mostly as a collection of anonymous functions that directly handle the URL routes and map to data in the databases. Here and there you find an obscure library with some business logic, often inherited, but it's mostly routing the data from the databases to the application that runs in the browser, on demand.
Together with the display logic in the browser, that's basically the stuff Lazarus does automatically for you when you make a desktop app. Rapid Application Development, as coined.
But only if you need databases and cloud and so on. There is still alot of software that needs to run offline (e.g. because it is in an encapsulated network and security parameter), or where there is simply no need to run it in the cloud. All this cloud stuff costs money for the provider, while when everything runs locally on the customer side, once the software is deplyed no running costs of operation.
You often see cloud for distributed systems, e.g. when you have control software for other machines, where then simply both can communicate over the cloud. The reason for this is quite simple, most of those cloud services (Azure, AWS) provide core modules like MQTT communication, database access, authorization management, etc. already making it really easy to use their cloud instead of writing your own networking architecture.
With end user software it's a bit different, because there you often have free software that makes money either by collecting data of the users, or by showing advertisements, or by locking functionality behind a paywall. While those are also sometimes mostly offline (e.g. StrongBox password manager for iOS is a freemium app with a subscription model, that fully works offline), it can be easier to enforce those things using a cloud (e.g. collecting user data is only possible when the user sends you the data).
But again, I don't see this as the majority of software, quite the contrary, because browsers get more powerful (JiT Javascript compilation, webassembly, optimizations, etc.), you now see a shift back to the end device. Sure it's still web based, but many software that previously required a backend, now runs fully frontend sided in the browser, simply because it is much cheaper to run the workload at the customer side, than to have to host it yourself.
And you can still combine this with the more predatory buisness models (like subscription, ad payed, data collection, etc.).
I see no reason why you shouldn't be able to have a cloud connected, user data selling, subscription and ad driven application fully written in Lazarus. Webtech is just often used because it is easy cross platform allowing for very nice graphics and animations (e.g. look at the wizard I posted above, it also draws it's own menu items, but canvas just looks really bad because it doesn't have sub pixel anti aliasing, I could probably build the same UI in HTML + JS in the same time and it would look much better)
Recent
