Xerox Alto [1] had GUI, Smalltalk 72 dev env and TCP/IP networking in 1973. How much real progress has there been in software technology since then?

[1] https://en.wikipedia.org/wiki/Xerox_Alto

Considering I can do extraordinarily more complicated tasks today, and a whole more of them, a lot.

But we also missed a lot of good stuff that we used to have.

Is that an improvement in software, or just a result of hardware getting faster?

Very much an improvement in software too.

Compare a 1980 compiler with the optimizations GCC and LLVM make.

Or v8 with the best interpreter in 1990.

Or video codecs (almost non existent then). Or cryptography code (immensely improved). Or the capabilities of InDesign compared to a 1990 DTP. Or Photoshop compared to crude editors in the 80s. Or modern GCs versus crude 80s and 90s GCs. Or a modern DAW compared to 1990s midi apps...

If you want what you had with Visual Basic in 1996, you're in luck: you can just use C# with winforms, which is like a much better version of VB6.

Unfortunately, in 2019 cross-platform compatibility for native apps has only gotten harder (good luck supporting windows, mac, linux, android, and ios), which is why people tend to default to making web apps (or hybrid solutions using frameworks like React Native) rather than native apps, and this has perhaps also created a vicious circle by reducing interest in what cross-platform options there are (QT and javaFX).

> which is why people tend to default to making web apps (or hybrid solutions using frameworks like React Native) rather than native apps

Not sure why people keep making the claim that native is dying. More and more people are doing their computing on their smartphones and the vast majority of the top 100 apps on both iOS and Android are built in Objective-C/Swift/Java/Kotlin. It seems to me like the main people pushing the whole web app/PWA renaissance are web developers who resent the fact that the web is dying.

Check the Lazarus IDE (https://www.lazarus-ide.org/), it's very similar to Delphi with an integrated code editor, visual GUI designer and debugger - you can debug on the target platform including Linux, Windows, Mac, Raspberry Pi, and so on.

I learned Visual Basic 6 from a cousin when I was 8 years old. It was such a joy to use, and made me fall in love with programming. It was so intuitive to use as an 8 year-old. And while Dijkstra might disapprove, the Basic language of VB6 was really simple and easy to grasp as a kid.

The coolest part was distribution. Under the File menu, you had an option to build a .exe file. I would build it and put the .exe file on 1.44 MB floppy disk, and share my little VB6 programs with friends. Such wonderful memories!

Dijkstra had issues with the older BASICs where the only way to control flow was with GOTOs and people flinging them all over the place. Visual Basic 6 (and QBasic before it) has Algol-like structures like subs, functions, while/repeat loops, etc so it would satisfy the issues Dijkstra had with older BASICs (if he'd dislike it for other reasons is another matter though).

Most people just repeat Dijkstra's headlines without having even read his text, let alone understanding it.

I know we're talking UI frameworks, but...

I think we lost something important with the shift from workgroup to client/server (especially ODBC).

I banged out a lot of dBase II, FoxBASE, MS Access (in-house) apps, back in the day. I've not felt as productive since. I miss it dearly.

Software based on that still exists and it’s a hell to maintain.

To which I can reply 'come back in 20 years and tell me about your experience maintaining something based on npm and friends'. At least those old crufty systems had the good grace of not having a best-before date measured in weeks. They might lack many other graces but for maintainability there is something to be said for a modicum of temporal stability when it comes to development environments.

+1, would agree with again

One of the things that we sacrificed was what today we would call "end-user programmability" but which should really just be called "using a personal computer." Hypercard was a UI builder, a way for regular people to program beyond the strict limits of the software given to them, and a hypermedia platform whose "applications" could be easily distributed. To do most of the things that Hypercard could do on a personal computing system of the early 90s on a system today requires someone to learn a full fledged programming language

Everyone of these posts I Ctrl-F the first Delphi comment

Can you imagine trying to explain to a developer in 1995 what you need to use to produce an app feature equiv in 2019