Leaderboard

Just start to be realistic and first of all fix bugs Knock out all platforms expect Windows family (for rest that's what lazarus/fpc is for) and focus only on Windows and VCL + native support for ARM. Performance is poor, look how much optimization they do in C# in many aspects c#(net 9.0) code is already much faster than Delphi. Don't delude yourself that the shrinking business and extinct delphi programmers(sad but true) will make this a wishful thinking gig.... Sadly, no, either you have a product for everything (that is, for nothing, with 1000 bugs more or less serious), or you focus on a narrow field and polish it like a diamond. Dear coders, stop mas****ating to this extinct product with disastrous management by old dinosaur boomers. Come down to Earth. Maintaining all those features with all new is definetely too much for them. It only looks good in theory and marketing slogans.

Since mobile platforms are changing so fast. It really depends at what moment you compare. At anytime the state of Delphi can be better or worse than last time. In 2100 when the mobile platforms are going to be stable like desktop you can expect a better integration with Delphi.

By now, like I write in another thread, I use Certum (since two days ) and they are compatible with the Windows certificate store. Of course they install they own drivers, but without calling them directly all applications view the certificate and can use it.

top request 1.IDE improvement. code formatter.denbuger[like interface can not see menber,mutithread] LSP codeinsght code imp. 2.RTL ,improvement..[base tcp component then can develop all above tcp do not use Indy!] high performance socket base->HttpClient/HttpServer/websockets/others encrypt base-->common like aes des and so on. RTL performance.today delphi 'code performance(use many buildin RTL function)even lower then .net 3. More Languane future like async(await like ),generic function(not class),interface can have field(C# imp it many years ago),more language suger. 4VCL some layout component

1. Debuggers, debuggers, debuggers - multithread handling is abysmal today 2. Make HighDPI actually works as intended - it is useless in a team as is - unless everyone runs the same scaling 3. Generics constraints for enumerated types to enable the use of conversion to/from ordinal values, use set operators, etc. 4. Native ARM64 compiler for Delphi for Windows (and Linux/Raspberry PI, but Windows has prio) 5. A 64-bit IDE that ensured that EMBT was dogfooding 64-bit VCL and RTL and raise the quality As for AI, I wouldn't mind an AI that could look at code and suggest improvements - perhaps as a part of the static code analysis, or one that could explain "what does this code do". I don't really need or want an AI to generate code. If it is a standard, a lib should already exist. If it doesn't and there is no standard that covers the need, I'd be happy to have an AI outline alternative approaches - but given that the wide scope of parameters that goes into a design, I think it is unrealistic to expect that it would come up with the ideal suggestion without us writing a huge requirement. The output of VLLM generators is extremely dependant on precise and accurate specification statements, and writing those are almost as hard as writing good code. As for AI and privacy - keep your secret credentials separated from your code. Other than that, I don't think that many of us write code that truly needs to be hidden for secrecy reasons, although it is obvious that it will be necessary to ensure that privacy permeates the use of AI.

Apparently not anymore. I'm glad to see the industry coming together.