Raspberry Pi 2 arrives with 6x the speed, Windows 10 compatibility

[crossy]

Eh? teaching HTML and CSS is hard, because the languages are god awful clusterfornications of evil.

I wouldn't disagree in the slightest there - then again C++ figures pretty large in my pantheon of "languages probably popular at SatanSoft"

Python is a nightmare because they broke it all in 3, so first timers get stuck when they google because the answer they are copying doesn't work. Python is however a great teaching language in many ways, but having taught whitespace sensitive languages before, I find them to require a lot of IDE tooling to help users really understand, blank being control is a hard concept. My point is, having a universal runtime, treating the device as a 'runs any language' device is actually remotely helpful.

I hope we can see things like google's code blocks or TouchDevelop actually start to use the Pi now.

TouchDevelop looks pretty cool - I just love write-once, run-everywhere fast systems - the "everywhere fast" part excludes Java. Just annoyed that I can't remember the name of that project I was thinking about. Basically you dropped blocks in - Visio-style - to make the structure of the program then opened a block to put in it's "attributes", so a loop block has start, end and iteration. It was also pretty good at encouraging the idea of developing your own blocks that could then be shared - black-box style - with the rest of the class.

Although one thing I will agree with you on - teaching C#/Mono - is definitely preferable to C++ or, heaven forfend, the C language that I cut my teeth on. I'd just be quite concerned about locking schools into yet another Microsoft-only ghetto.

W10 on Pi2 is "Embrace", I'm waiting for the other shoe - Extend - to drop...

[TheAnimus]

I really rate TouchDevelop, because of how it uses concept keyboards. Someone who spends just one week playing with that for one hour a day can and will suddenly learn a regular programming language quite quickly.

Are you talking about Blocky? https://code.google.com/p/blockly/ I sort of like what they are doing, but, and I know this sounds bad, I dislike how crap it looks. It also doesn't have half as much cool support for heavy lifting as TouchDevelop does. I had a 10 year old Vietnamese kid, who spoke english poorly, make a flappy bird clone (seemed fitting) in about one hour the other day. That kid I hear is now hooked on being a programmer, which will offer an amazing step up in quality of life compared to his parents (who I've known for a few years).

Have you not seen what they've finally done with C#? It was an ECMA standard before, by that I mean the language was more open than say Java, but now the whole next gen compiler is completely open source, but so is the coreCLR which is currently building on linux now (as of yesterday!). I honestly can't see any reason for teaching Java now, I also don't think there is much general purpose software that is worth writing either, as I've been saying for years, Java Lang is a dead end and has been since 2005.

[DanceswithUnix]

Have you tried your kids on something like google code or TouchDevelop ?

They have tried all sorts, though I don't remember doing TouchDevelop (they may have a school though).

They tend to end up back on Scratch, the pull of being able to draw something and animate it is really strong. Whilst it is pretty limiting in most ways, they are quite fluent with ideas of setting up observers and handling events so while I think Scratch is now holding them back it hasn't been time wasted.

My daughter has a Thymio which is pretty good: https://aseba.wikidot.com/en:thymio, though again when it comes to uploading programs it uses something which isn't quite a standard language and so far she has just used the simple drag and drop programming interface which is "on this sensor saying this, do this" event model similar and as limiting as Kodu. Tying back to what you were saying before, Thymio is more expensive than a Pi yet seems better value.

So my idea is to get the Pi set up with the Python Minecraft API and write a bit of code to build a floor out of stone and a wall out of wood, maybe add a window. Then with a starting point to follow leave them to it and just see what they do. Probably place blocks in an unbounded loop taking the box down, but that is educational in it's own way

[TheAnimus]

All of these things are massively better than the Pi is my point, and don't require any outlay, because *everyone* either has a computer (Blocky/Scratch/TouchDevelop) or a tablet (TouchDevelop)

My favourite is actually TouchDevelop, because it's quick to do stuff, if using a scaffold like the flappy bird clone, far faster to get actually useful playable game made. Reminds me of Click and Create in that way.

However I think TouchDevelop is better because of how it gets people used to written rather than graphical layouts and constructs. I'd really recommend having a click through it, worth an hour of anyones time:
https://www.touchdevelop.com/app/#li...rials:overview

I like how they have the three levels of 'expertise' too. It's a bit harder in a way, because most of the tutorials start with a blank canvas that you bring content into, so they will need to have a 5 minute attention span. So I don't think it's as suitable for under 8s, however it's no worse than Logo in that regard.

What I really like about TouchDevelop is that it runs on Desktop and Tablet or Phone, so kinds can easily share their creations. Kids that want to be graphic artists can just do that, make new bitmaps for existing logic, kids that want to change stuff round, can do that easily too.

[crossy]

Have you not seen what they've finally done with C#? It was an ECMA standard before, by that I mean the language was more open than say Java, but now the whole next gen compiler is completely open source, but so is the coreCLR which is currently building on linux now (as of yesterday!). I honestly can't see any reason for teaching Java now, I also don't think there is much general purpose software that is worth writing either, as I've been saying for years, Java Lang is a dead end and has been since 2005.

Agree with this. I was going to "do" Java (because it's good for job applications) but to be honest it seems like the once mighty Java is now a bit of a dead-end. All the exciting stuff is being done elsewhere. C# seems to be eminently usable on Linux (Mono) and Android now, so it just went from "Microsoft curiosity" to "genuinely interesting" in my books. Actually the language looks pretty well worked through too - more or less as if they looked at Java and were careful not to make the same mistakes.

But job 1 at the moment is to get the Python3 skillz up to speed - need that for current job, C# might be useful for the next one though. One thing does occur though about the new Pi2 - if you'd got a mixed selection of Linux and W10 ones and perhaps the odd real desktop then it'd be a good environment to show the benefits of doing platform agnostic apps. Which, I'll suggest, is definitely a good idea in the current (and future?) environements.

[DanceswithUnix]

Agree with this. I was going to "do" Java (because it's good for job applications) but to be honest it seems like the once mighty Java is now a bit of a dead-end.

My last job change was about 3 years ago, and was triggered by the old company switching from 'C' to Java. Recruiters were telling me back then that Java was pretty much dead, and I don't think Oracle going after their biggest user in Android has helped matters.

Have you seen the Tiobe graph? Big pinch of salt application required and mentally apply a huge smoothing function on the oh so spiky results but still interesting:

http://www.tiobe.com/index.php/conte...pci/index.html

Look at the downward slope on Java It is still big, but that says to me that there are a large pool of Java programmers with a shrinking amount of work to keep them employed, which won't do well for salary.

C# seems safe for the foreseeable future, and probably would get you some interesting work.
If you want to keep a roof over your head, add together the C++ and C figures because if you can do C++ well then you can do C. Look at "Very Long Term History" table further down. I learnt 'C' in 1987, if you told me then I would be still using it decades later I would probably have laughed at you, but C and C++ still seem to be the heavy lifting languages of choice used for all modern operating systems, tools and databases.