331–338 of 338 posts
Earlier quoted context omitted.
I suspect that part of it is unusual syntax--if...fi, anyone?--and part of it is unusual syntax in non-bash things, i.e., Awk looks like moon runes, and to replace ; with \; in sed you need a grand total of SIX backslashes.
> if...fi, anyone? And here I thought everyone wanted ALGOL-like syntax...
Earlier quoted context omitted.
> if...fi, anyone? And here I thought everyone wanted ALGOL-like syntax...
As someone who dabbled in C-like languages, it threw me for a little bit.
1) "C-like" languages are often described as having "ALGOL-like" syntax. And indeed, C was based on languages significantly influenced by ALGOL-60 (though ALGOL-60 used BEGIN/END for blocks, not curly braces).
2) "IF/FI", "CASE/ESAC" and similar were in ALGOL-68, and Stephen Bourne drew them from there when he wrote the Bourne shell (with which bash was eventually intended to be compatible).
I've used Yeti[1][2] to write an Android app for personal use (on Nook Touch) that I couldn't force myself to write in Java. It made the task much more fun for me and allowed me to finish after I got a serious case of "author's block" trying to code it in Java. Yeti is ML-like and runs on JVM 1.4+ [1] http://dot.planet.ee/yeti/intro.html [2] http://mth.github.io/yeti/
Now this is interesting to me! I found Yeti last year when looking for a practical strict ML in which to write a small utility. It didn't fit my requirements at the time due to the JVM dependency but I made a note to try it later for some JVM task. I see that it has had a new release since, which makes me want to try it more still. Could you share a little more feedback about what using the language was like? Did you…
I haven't used SML nor OCaml. [Though I seem to believe I prefer SML over OCaml in terms of visual appeal/readability, FWIW.] I wanted to code for Android, and OCaml-Java requires too new a JVM (from what I understand, Android is stuck with JVM 1.4 [or 1.5?]) IIUC. Before I found Yeti, I tried to use Kawa (a Scheme/LISP for JVM), but it worked too flakily for me, unfortunately. Then I found Yeti; I stumbled into one bug at some point, but the author quickly fixed it, and I didn't have any problems since then! And it's typed, so a plus vs. Scheme.
edit: you can view my code at: https://github.com/akavel/bookshelf/tree/main/src/com/akavel -- but please note I didn't care to make it pretty or documented!
edit 2: also, I liked Yeti's manual & documentation very much.
I've recently fallen for Standard ML. It's been a steep learning curve due to all the compilers and implementation choices but what a mind bending approach to software development. OCaml seems to be the obvious next step but I'm having fun with toy programs in Standard ML right now.
What do you find mind bending about software development in SML?
I find the type system to be easy to work with (unlike C, C++, Java) and I'm finally seeing the virtues of a static typed language. Also, having to specify types while defining functions forces me to think more in a good way. It's a far more methodical approach than I'm used to in Ruby for example.
Earlier quoted context omitted.
Now this is interesting to me! I found Yeti last year when looking for a practical strict ML in which to write a small utility. It didn't fit my requirements at the time due to the JVM dependency but I made a note to try it later for some JVM task. I see that it has had a new release since, which makes me want to try it more still. Could you share a little more feedback about what using the language was like? Did you…
As I wrote above: "It made the task much more fun for me and allowed me to finish after I got a serious case of "author's block" trying to code it in Java." In other words, I did like it! :) I haven't used SML nor OCaml. [Though I seem to believe I prefer SML over OCaml in terms of visual appeal/readability, FWIW.] I wanted to code for Android, and OCaml-Java requires too new a JVM (from what I understand, Android is…
Funny that you mention Kawa: I've been meaning to try it out along with Yeti. That you had problems with it does not sound very encouraging, but then, it has had a major version release since.
When it comes to ML, I prefer SML to OCaml. I think Yeti, Standard ML and Scheme appeal to the same "small core language" mentality. If you decide to give SML a go, I'd recommend SML/NJ for development and as your REPL — especially while you are getting used to the language since it has better error messages — and MLton for production binaries.
Earlier quoted context omitted.
It'd certainly be interesting to relate a multi-decade software project from the outside :-) I will have to talk to my father about this... and explain why folks on the internet must hear his stories. In the interim, would anyone happen to have an example of a good secondhand account of a software project? A basis or, at the least, an inspiration for how to proceed.
Not just a software project. More of hardware. But a good read, IMO: The Soul of A New Machine by Tracy Kidder. About the race to build a new minicomputer, by a team at Data General. https://en.m.wikipedia.org/wiki/The_Soul_of_a_New_Machine Prize winning book.
Excellent to bring up!
Earlier quoted context omitted.
Not just a software project. More of hardware. But a good read, IMO: The Soul of A New Machine by Tracy Kidder. About the race to build a new minicomputer, by a team at Data General. https://en.m.wikipedia.org/wiki/The_Soul_of_a_New_Machine Prize winning book.
This is actually my favorite engineering book of all time. Excellent to bring up!
[ Many of the engineers state that, "They don't work for the money", meaning they work for the challenge of inventing and creating. The motivational system is akin to the game of pinball, the analogy being that if you win this round, you get to play the game again; that is, build the next generation of computers. ]
Earlier quoted context omitted.
What do you find mind bending about software development in SML?
Coming from dynamic languages I find Standard ML to offer a much more methodical approach to designing software. The view that SML brings is that your program is a language of its own and the Algebraic Data Types you define are your program's syntax. I find the type system to be easy to work with (unlike C, C++, Java) and I'm finally seeing the virtues of a static typed language. Also, having to specify types while d…