Its on valentines day, and on her birthday, that we can now look back on the 30 years of life of the Delphi product line and reflect on the wonderful product she was and the sad death by a thousand cuts that finally saw her depart in 2024.
Thank you Delphi for all the great memories.
3 Likes
What absolute drivel.
3 Likes
I understand his perspective (Idera) Ian.
Be nice 
Itās wholly inaccurate, especially with reference to 2024 and to the current period.
2 Likes
I nearly died when I read āā¦the sad death by a thousand cuts that finally saw her depart in 2024ā
The old girl is going strong here.
She is pushing industrial automation and machine vision around on thousands of our client sites. We would be lost without her.
2 Likes
A product canāt survive without
- A work force
- A jobs market i.e. new Delphi jobs
- Up-to-date libraries
- Commercial trust in the future of the product and the above factors
- New projects coming onboard each year
I did not suggest legacy projects donāt exist, that they donāt work great or that their developers arenāt great - they do and they are.
But the sad reality is, these projects live only because of their developers who are ageing and their projects will die with them they retire or sadly pass away. This cliff has now been reached. How many Delphi developers really plan to be still working in 5 years time in Australia?
No major new work has been started in Delphi, in the western world over the last 10 years at any commercial scale to support a work force, support the third party vendors or support the future of Delphi.
Feel free to document a major app or project that started out new in the last 10 years in the western world in delphi and reached commercial success.
A product canāt be considered serious if its only being used for maintenance and hobbyist. This is the definition of dying.
Ian and his friends nailed the last nail in the coffin when they refused to release a yearly roadmap.
Theyāve also never shown any factual data to counteract all the factual evidence out there about Delphiās commercial death. No subscriber numbers etc to prove we are wrong.
YouTube views of Delphi VS Microsoft releases, forum usage, professional group membership etc all provide a picture so clear you really have to be true to the faith to believe its not the case.
Iām sitting here editing over a terabyte of video, having worked all week with a bunch of motivated people all pulling the same direction. The product is still around and the company going - something Borland, Inprise et al didnāt manage, having increased market share in both real terms and on questionable indexes like TiOBE (currently number 9, up from the low point of number 23 at one stage). Weāve managed in the past 12 months to get Delphi taught again in several universities and created a specific team to help push that further (I think education is key, I donāt think Iām the only one).
Yes, lots of missteps on the way. Many taking years to bounce back and weāve still got a long way to go. I get it, lots more to do and some people righteously annoyed. Many not. I think Iāve been open and engaged with people on it, listen to criticisms, get as many as I can acted on when itās practically and financially possible.
I donāt think I personally could do any more. Iām galvanized to reinvigorate things. I make absolutely certain everyone and anyone who wants to air their views, positive, negative or in between get to have their voice heard and heard at the highest levels of the company.
And then I come here, a Delphi userās forum, and I see a post which is absolutely, soul-destroying negative in a way that is, to me, quite unacceptable.
Iām not made of wood.
6 Likes
Well looking at the solutions out there, nothing quite matches Delphiās ability to connect to anything and work on anything. Currently looking at a system, that uses one language for iOS apps, another for Android App, Another for web services, Another for Office App and Data transfer, Another for Point of Sale.
Only a large software corporation with many different language experts could maintain it. The whole system is a house of cards. In my view had they had used Delphi for everything it would have made the system easy to maintain throughout the years.
āDelphi taught again in several universitiesā
I hate to pile on here, but I want to mention my (negative) experience with Delphi at university.
I was an Engineering lecturer at Griffith University, and ran the specialised engineering computer laboratory (To note these computers already had the full Microsoft suite (including Visual Studio)).
My Delphi interest, and āCā programming courses, meant I wanted to install Rad Studio (Delphi and C Builder) on those computers.
After assuring the IT support people that āDelphi / C Builder is NOT deadā, I got the okay to have this as another programming environment option.
So I then contacted Embarcaderoās (or more specifically the Australia Delphi distributor) to get a free installation of RAD studio installed. Yet the conditions and major hassles that were imposed on getting RAD studio on the university computers, definitely sealed the ānail in coffinā for RAD studio to ever be installed.
(Especially with Microsoft offering incentives to install Visual Studio, it was a surprise Embarcadero just put conditions on the install.)
Although it has been over 5 years since I retired from university teaching, so things may have changed. Also (at the time), there was no free RAD studio Community Edition.
With MS Visual Studio offering [at the time] a community edition, it was definitely seen that there was no interest in Embarcadero encouraging new users to the platform.
The inability for Borland/Embarcadero/insert-name-here to allow Delphi in educational institutions was, IMHO, the worst decision made in terms of impact to the uptake of Delphi. Bean-counters make poor strategic decisions. I would have been jumping over buildings to get Delphi into Universities. These decisions go back a couple of decades
Hi all,
Realistically the game was lost when Anders Hejlsberg was poached from Borland to Microsoft and Microsoft then allocated huge resources to C#.
For Delphi to survive as a commercial mainstream language it would instead have had to be sold to Microsoft, perhaps with Microsoft acquiring Borland. I can understand why Borland did not want to do this existentially fatal move , but the consequence was loss of market share for Delphi and a slow demise.
Anyway, in the near future with agentic Large Language Models (LLMs) likely to autonomously do most of the worldās future coding, and with much of this code either unseen or unintelligible to humans, past and current coding languages really may not matter. Software may well be mostly designed with specifications and parameters set in human languages delivered in prompts to LLMs, which will then autonomously write the necessary machine code, having invented their own intermediate languages if and where necessary.
Itās going to be very different future!
Cheers,
Martin
LLMs wonāt solve the automated coding problem any better than any other solution that has been tried and failed during the entire history of software development. The issue is that specs are never complete or right, which is why every approach till now has failed, and LLMs will do no better. Writing code is the easy bit. Trying to work out what it should do is the hard bit, and that never changes and can never be automated
Over 35 years of professional software development, the main impediment to good solutions is not understanding the problem. This is language and tool independent. The ability to think and create is the skill that is required. LLMs now, and in the future, do neither of these
Ive been hearing the Delphi Is Dead calls for 20 years. Every one of them was wrong including the one that started this thread. Delphi is a fantastic product and I really enjoy using it. Ive worked with other IDEs and languages and Delphi is still my favorite after 30 years and going strong.
2 Likes
Sure Delphi is not C# or Java, and Embarcadero is not Microsoft. Plusses and minuses.
Number one, I think we are a healthy and steady-ish community.
And encouragingly, I think we are building more connections between ourselves across the globe all the time.
The last year in particular I feel has been very successful.
There was the International Pascal Congress in 2023, and itās planning for 2026.
FPC has their own set of events : Conferences and Events - Lazarus wiki
And it seems like noone has managed to tell the Brazilian community that Delphi is in trouble.
40 speakers, on 6 stages, for 600 attendees !
And in the rest of the Delphi world there was also the first Delphi World Summit, EKON, ItDevCon, and our Symposium.
On a pragmatic level, I assume Delphi is a bigger seller than C++Builder currently ā¦ and in 2024 Embarcadero sunk significant effort into bringing the C++ product up from embarrassingly dated to close to bleeding edge.
And incorporated a very high quality code intelligence product in Visual Assist (VA).
Surely this should provide some hope that a similar degree of effort can be found for the Delphi product, over an extended period.
The existence of C++20 features in the VCL/FMX ecosystem should throw the relative āmodernityā of the two languages into sharp relief. C++ 20 and 23 provide :
- views
- ranges
- concepts
- (modules)
- coroutines
- senders & receivers
- optional and expected types, with monadic operators
- pattern matching
- structured bindings
- constexpr / consteval
- and more
Pascal could have a chance to come late to the party with many possible new (to Pascal) programming ideas ā¦ and be able to find elegant new idioms by seeing what other languages have done with the ideas, and try to learn from their uglier mistakes.
In some cases, it is only when the established powers leave something for dead, that those with innovation and ambition can take it and make something great of it. Will this be the case with Delphi/Lazarus?
1 Like
Hi Misha,
You may have under estimated the new agentic LLMs coming out. They really can be great at problem analysis and very creative in making solutions.
I am a specialist dentist (endodontist) and I got Google AI Studio to create a diagnostic web app (HTML, CSS, Javascript) using just simple symbolic AI to discriminate between acute pain from a dying nerve in a tooth and jaw muscle pain; very often these two pathoses are confused by clinicians and the wrong treatment is provided with damaging consequences. I was amazed at what the LLM creatively thought up and implemented in code without me designing any part of the background reasoning or the code design. And it worked well. I do not know of any reference project the LLM could have known to design this solution and I know the AI/Dental field well. Quite remarkable, if a bit scary.
So donāt underestimate the agentic AI LLMs coming on stream. They are probably already cleverer than us in many ways! And who knows what is in the FAMMGA + ByteDance, TenCent labs to yet be released? Any you need to remember that 50% of people are at or below average. AI only has to be half perfect to beat half of the coders still giving huge benefit to the coding process.
Cheers,
Martin
1 Like
It is an interesting thing. Coming from Pascal on a PDP11 in the late seventies, Turbo pascal, all versions, then Delphi, having had to use Cobol, Data General Basic, C, and others, I think I did a bit of Fortan but Iām too old to remember now.
The current language and environment is fantastic. Other than the fact that the latest IDEs are a lot less stable than they should be, even 12.2, it is a habit now to save (Ctrl-s) after every change, so I donāt lose anything when the IDE crashes, which it does often. But things are changing. Scaling is getting better, but still nowhere correct, but then again Microsoft canāt get it right yet, so who can.
New development from my point of view is much easier done in Delphi. Prototyping is so much easier than any other environment I have tried.
Native exeās kill all other options. The whole C# and Jave environment seem to me to be similar to the DLL hell of past times.
If you know my email, (Alister Christie, I think you know what Iām talking about) you know what I am dealing with from a historic point of view.
All new development is in Delphi, and I really canāt think of a better environment.
Ian, I applaud your efforts to get Delphi, or even just Pascal, out there. Education is the key.
We are embarking on a major new development (for New Zealand) and Delphi was the only choice.
You have my support.
2 Likes
I can definitely report that my experience with DOS programming - and even using eg Word or AmiPro - back in the day was to be constantly saving my work ā¦ and then every so often changing the name of the file slightly and saving again, as a means of preserving forward progress. 
This deeply affected me for a very long time. Decades. 
Thank you, David, I really appreciate it. You also know, I hope, you have my support too. Iām here for the wider Delphi community and I am always, first and foremost, a Delphi developer before anything else - and I was that LONG before I started wearing an Embarcadero shirt. 