Delphi’s selling point (from my perspective, at least) is it’s ability to do visual UI design. Whether that be web (I wish), mobile or desktop. Without that it is irrelevant.
So to disregard mobile would be at it’s own peril, because to solely rely on desktop would be backing a shrinking market.
Some interesting views here.
I’m in the weird position of actually being an Embarcadero employee and someone who creates mobile apps (outside of my work for Embarcadero).
For the mobile apps it’s roughly 50% SaaS gateway/adjunct to a desktop or web portal and 50% providing services which are, in essence, mobile first. For example a secure ‘wound camera’ for one app and an encrypted Enterprise contacts/messaging app. Yes, of course, all in Delphi.
Speaking as a developer rather than representing Embarcadero, the choice of Delphi for the mobile apps was simply because it was the best solution available. Low code, component-based development, code re-use of existing business logic, familiarity with the platform, bells and whistles like the themes, access to full source of the created solution, and so on.
We did look at other options - for example there was a lot of pressure from one direction to use Xamarin at the time, and lots of representation from people to use React and similar solutions. I’m glad we went with Delphi and the clients are still using the app, with zero problems (except for when Apple does something weird, especially with certificates or generally by being Apple) over the course of several years and several iterations of RAD Studio.
Has it all been plain sailing? No, development is an absolute PITA particularly when it comes to mobile because it’s still an ecosystem where the key players are almost total monopolies who do whatever the heck they feel like to keep their grip on the customers and are very laissez faire when it comes to the end developer (remember the guy who wrote a very popular flashlight app which was instantly killed overnight by Apple?)
But yeah, I do write real apps for desktop, web, and mobile - and, for me, RAD Studio is the tool to use. It’s not perfect (and Lord knows I am working hard for us to get better) but it is a viable, practical, solution for me.
I agree.
Alex
If we (Embarcadero) were to invest in SaaS/web - what would that look like to you in order to be credible? What could we do/create that would help with that?
Ignore the “but they never would” or going over the many historical missteps that may or may not have happened - I am asking you, as someone who can genuinely influence the decision-making process, for you to tell me something practical you think could be done. Blue sky is fine. Be constructive. Let me know. Maybe it can be done, maybe it can’t, but as the old adage goes: if you keep doing, what you’ve always done, you’ll always get what you’ve always got - except, in this case, we might not even get that.
Delphi really doesn’t have a great ip/web architecture. If you look at where asp.net, ruby on rails and other web platforms are now, you can see how much work would need to be done to catch up. That would take a massive investment in the right people with the right skills and the will to work on it (who may not exist). They would effectively need to start from scratch - that’s a mammoth task.
Also, the web and SAAS in general is not running on windows - don’t see a Linux IDE appearing any time soon.
Not sure that distils it down enough for me to understand what it would look like as a proposal.
Are you saying:
Take the ideas of ASP.Net. Blazor, Ruby?
A Linux IDE with that kind of committed new architecture would be essential?
It would not have to be free. 
I wasn’t really proposing anything saas related, just pointing out some the difficulties involved - but I’m game.
I’m not sure this is really achievable in a realistic timeframe. Half baked attempts would be pointless - and the cost of doing this properly would be astronimcal. It would take quite some time to recruit a large enough team with skills needed.
If you look at the evolution of dotnet core, they had many misteps and it’s taken 8 major versions to get to the point where it is now (pretty good but not perfect). That includes significant contributions from the open source community.
I’m not sure something like Blazor (not a fan of it) could be created easily in delphi, very much relies on the managed nature of dotnet and source generators - straying into a whole new compiler architecture territory there.
Then there is the business side of it - the development costs would in the millions of dollars - and RIO is extremely uncertain - how would it be priced (royalty based would be a non starter imho) - what would companies be prepared to pay?
Even more uncertain is - would customers be willing to go down that route - closed source - proprietary dev environment with a limited pool of developers with the skills to use it properly. Even dotnet has only a small market share in the SAAS world - most are using a mix of different open source technologies.
So the reality is this is all pie in the sky stuff - and instead embarcadero should focus on consolidating what they already have, which is still the best windows ui developement tool on the market - it just needs a lot of TLC to make it viable into the future.
I wasn’t really proposing anything saas related, just pointing out some the difficulties involved - but I’m game.
- Go all in on Linux - lets face it, the only company doing saas on windows is microsoft.
- A Linux ARM compiler - Cloud providers are moving towards arm in a big way - due to reduced power usage, bigger bang for buck.
- Linux devs want to develop on Linux, so yes you need a Linux IDE (not a Kylix resurrection!). Remote debugging doesn’t cut it.
- An RTL base which is highly performant and optimised for low memory usage.
Amen
- Base socket library on which high performance libraries can be created.
- A high performance web server (like kestral).
- A web application framework which is performant and extensible.
The open source community is better positioned to fill these things in.
In fact… it has. Lazarus is all that you asked for, and starting to be reliable now
Grahame
I don’t think any work should commence on this.
You have to start on a project like this by having someone in the company with a vision for something better. I don’t see vision inside the current Embarcadero.
At the very fundamentals of where we are now, you can’t say ‘here, we made something that is the
same as already exists, for $1400+/dev/year’ and expect any real movement of developers.
The starting point now is basically everything NodeJS, React, Next.js etc as well as what Microsoft is doing - for essentially free - just to start.
When Visual Studio Code already exists on all platforms, why would you invest in a new IDE today? Its really hard to see what benefit you could bring - if there is some they should be listed and people should be able to decide if that really matters to them. Its hard to even see where RAD Studio exists in a VS Code world. Whilst web deployment is on Linux, the number of dev’s using Linux for their work operating system is very low. MacOS would have a higher take up.
The Delphi language has to be competitive with modern language features. These are hard things that no one else can really do except Embarcadero. Let’s say we had a modern language that was competitive with Swift UI, C#, Javascript/TypeScript including some form of native JSON support in the language, maybe then we could be back in the game.
Relationships with the best and brightest in the pascal world need to be fixed. The likes of mORMot don’t even get free licenses to Delphi so some of the most important code being built in pascal doesn’t even work in Delphi.
Delphi has to support high performance modern web standards to build modern web servers, microservices and associated socket applications. These have to be built at commercial standard. All the plumbing developers need also needs to exist like professional pooling, polling etc.
TLS1.3, WebSockets, Secure SMTP, High Performance Web Servers. Linux first optimisations.
Finally, you need to have a proper conversation about how web UI should really be built. The drag/drop into a pixel surface approach to web really doesn’t work well.
It would be good to keep people with large Delphi code bases on the platform, but to survive new code has to be made in Delphi as well. This requires a reason to choose it.
Essentially, to build something right you need to open up your ideas to the community and let it be discussed, not spend another 5 years building something that we could already have told you won’t work, just as mobile has failed because coding got in the way of assessing the business problems that needed to be resolved first.
This is not the first time that has been mentioned - and I specifically said if you work on the project or know someone that does - get in touch. I’m here to make these kinds of things happen. I can make those kinds of things happen (and have, even today, for a different project).
I have already reached out to some Lazarus and Free Pascal people and when I am in Amsterdam in June I’ll be talking to a lot more in person too.
WRG to the web - I totally agree with where you’re coming from on that. Pixels and the web…no that’s not the right way, and in some ways the Delphi visual design paradigm being applied to the web is likely a specious path to take. I know of some interesting as-yet-unreleased third-party projects which seem to have a very good chance at doing some remarkably interesting stuff when it comes to the web.
Separately I personally think Linux is a significant thing which needs to be embraced properly, fully, and urgently, for a number of reasons, and that definitely means Linux ARM as a part of that.
All I can do is keep doing what I am doing - inviting people to have their say, to engage with me (and by proxy Embarcadero as a whole) and do exactly that; invite people to tell us their beliefs, ideas, and feelings. I understand why people might be skeptical or cynical. As I’ve said previously the only way we will establish credibility is by walking the walk, keep striving to increase quality, be as authentic and genuine as we (and I) can, listen to people, try to do the right things.
Our benefit is 29 years of history - but that’s a very checkered history (for all sorts of reasons) and, as Detective Monk would say, that longevity - it’s a blessing…and a curse.