5 ms·

An interesting project for devs is Glamorous Toolkit (https://github.com/feenkcom/gtoolkit https://github.com/feenkcom/gtoolkit) as an experiment/exploration of

An interesting project for devs is Glamorous Toolkit (https://github.com/feenkcom/gtoolkit https://github.com/feenkcom/gtoolkit) as an experiment/exploration of better dev tooling.

liminal 5y ago
That IDE looks amazing, but it looks like I'd need to be proficient in Smalltalk to use it for any language. Watching the video on the site and it's impressive what the presenter is doing with live representations of code.
whartung 5y ago
And, frankly, this just reinforces my overall opinion of Pharo. Pharo is really amazing, they've been working on it, actively, forever, and make great strides. But I'm convinced that Pharo is an amazing development environment and eco-system designed to make amazing development environments and eco-systems. All of the folks are putting work in to improving IDE to make a better IDE. They're on their, what, fourth(?) iteration of the "widget" set that folks still find difficult to use and under documented. As recently as Pharo 7, when I last looked in any depth to the matter, their rendering model is straight out of the 80s. Vastly less capable than Java2D or Canvas in the browser. It's my understanding it's on their todo list, but I don't know how far they've advanced on this. What focus there is on applications seems to mostly come from Seaside, the web server application framework. If you look at their "success" page, most (not all) are web apps. It would not surprise if someone, somewhere, is fusing Pharo and some Electron-esque rendering front end and using it to create a "desktop application". And it's not unreasonable to say "Well, web is where it is, desktop is dead." There's certainly arguments in that direction. In the end, it's moot. The devs have their itches to scratch. Nature of the game. Their itches are not my itches, and an "advanced" rendering model and modern widget set is something I want to use, not write. So, I'm at their mercy waiting for whenever that day may, if ever, come. If was interested in writing IDEs, then perhaps Pharo would be a place for me. Instead I'm interested in writing hobby desktop apps, and it's not there yet.
- mikro2nd 5y ago
  There is certainly some truth to what you say. Threading is... just not (afaik it's a sort of green-threads thing) and UI is, as you point out, a crapshoot. However... ;) If I'm going to write code for my own amusement, I want to work in Smalltalk and a Smalltalk environment. Ever since I first encountered it in the early 1980s, it's been my environment of choice. IMHO there's still no other programming environment that approaches the flexibility, liveness, debuggability, white-boxness of Smalltalk. To paraphrase Guy Steele, "All languages will evolve until they become Smalltalk". (Steele said something of the sort about Lisp.) That said, and jokes aside... Where Smalltalk beats every other language environment I've ever worked with or seen is at exploratory programming: Programming where you don't quite understand the problem, the domain and/or the solution space. It's more a tool for thinking about problems and how to solve them than something for a fixed, final answer. I suggest, further, that a very great deal of software development is exploratory in nature, so using a Smalltalk is a win, at least until you've understood everything in sufficient depth to "just implement a solution". At that point go to something with better runtime performance, deployability, etc. by all means. Caveat being that sometimes (quite often, really) the exploration never, ever reaches a conclusive terminus, so you may want to stick with a Smalltalk. tl;dr: Use Smalltalk to explore and understand, and then finally to deliver an executable, test-driven specification.

An interesting project for devs is Glamorous Toolkit (https://github.com/feenkcom/gtoolkit https://github.com/feenkcom/gtoolkit) as an experiment/exploration of better dev tooling.

liminal 5y ago

That IDE looks amazing, but it looks like I'd need to be proficient in Smalltalk to use it for any language. Watching the video on the site and it's impressive what the presenter is doing with live representations of code.

whartung 5y ago

And, frankly, this just reinforces my overall opinion of Pharo. Pharo is really amazing, they've been working on it, actively, forever, and make great strides. But I'm convinced that Pharo is an amazing development environment and eco-system designed to make amazing development environments and eco-systems. All of the folks are putting work in to improving IDE to make a better IDE. They're on their, what, fourth(?) iteration of the "widget" set that folks still find difficult to use and under documented. As recently as Pharo 7, when I last looked in any depth to the matter, their rendering model is straight out of the 80s. Vastly less capable than Java2D or Canvas in the browser. It's my understanding it's on their todo list, but I don't know how far they've advanced on this. What focus there is on applications seems to mostly come from Seaside, the web server application framework. If you look at their "success" page, most (not all) are web apps. It would not surprise if someone, somewhere, is fusing Pharo and some Electron-esque rendering front end and using it to create a "desktop application". And it's not unreasonable to say "Well, web is where it is, desktop is dead." There's certainly arguments in that direction. In the end, it's moot. The devs have their itches to scratch. Nature of the game. Their itches are not my itches, and an "advanced" rendering model and modern widget set is something I want to use, not write. So, I'm at their mercy waiting for whenever that day may, if ever, come. If was interested in writing IDEs, then perhaps Pharo would be a place for me. Instead I'm interested in writing hobby desktop apps, and it's not there yet.

mikro2nd 5y ago
There is certainly some truth to what you say. Threading is... just not (afaik it's a sort of green-threads thing) and UI is, as you point out, a crapshoot. However... ;) If I'm going to write code for my own amusement, I want to work in Smalltalk and a Smalltalk environment. Ever since I first encountered it in the early 1980s, it's been my environment of choice. IMHO there's still no other programming environment that approaches the flexibility, liveness, debuggability, white-boxness of Smalltalk. To paraphrase Guy Steele, "All languages will evolve until they become Smalltalk". (Steele said something of the sort about Lisp.) That said, and jokes aside... Where Smalltalk beats every other language environment I've ever worked with or seen is at exploratory programming: Programming where you don't quite understand the problem, the domain and/or the solution space. It's more a tool for thinking about problems and how to solve them than something for a fixed, final answer. I suggest, further, that a very great deal of software development is exploratory in nature, so using a Smalltalk is a win, at least until you've understood everything in sufficient depth to "just implement a solution". At that point go to something with better runtime performance, deployability, etc. by all means. Caveat being that sometimes (quite often, really) the exploration never, ever reaches a conclusive terminus, so you may want to stick with a Smalltalk. tl;dr: Use Smalltalk to explore and understand, and then finally to deliver an executable, test-driven specification.