Machine Learning IDE in Alpha

Machine is an IDE for building machine learning systems using TensorFlow. You can sign up for the alpha, but first, have a look at the video below to see what it is all about.

You’ll see in the video, that you can import data for a model and then do training (in this case, to find a mustache in an image). You’ll see the IDE invites an iterative approach to development since you can alter parameters, run experiments, and see the results.

The IDE syncs with “the cloud” so you can work on it from multiple computers and roll back to previous results easily. We don’t know when the IDE will leave alpha status (or beta, for that matter), but the team’s goal is to release a free version of Machine to encourage widespread adoption.

If you want to learn more about TensorFlow, you are in the right place. We’ve also covered a bare-bones project if you’d rather get started that way. You can also find some good background material going all the way back to the early perceptron-based neural networks.

Codebender Rises from the Ashes

If you were sad that Codebender had bit the dust, cheer up. A site called codeanywhere has acquired the online Arduino development environment and brought it back to life. In addition to the main Codebender site, the edu and blocks sites are also back on the air.

Not only is this great news, but it also makes sense. The codeanywhere site is a development IDE in the cloud for many different programming languages. The downside? Well, all the people who said they’d be glad to pay to keep Codebender alive will get a chance to put their money where their mouth is.

Continue reading “Codebender Rises from the Ashes”

The Raspberry Pi Becomes a SCSI Device

SCSI devices were found in hundreds of different models of computers from the 80s, from SUN boxes to cute little Macs. These hard drives and CDROMs are slowly dying, and with that goes an entire generation of technology down the drain. Currently, the best method of preserving these computers with SCSI drives is the SCSI2SD device designed by [Michael McMaster]. While this device does exactly what it says it’ll do — turn an SD card into a drive on a SCSI chain — it’s fairly expensive at $70.

[GIMONS] has a better, cheaper solution. It’s a SCSI device emulator for the Raspberry Pi. It turns a Raspberry Pi into a SCSI hard drive, magneto-optical drive, CDROM, or an Ethernet adapter using only some glue logic and a bit of code.

As far as the hardware goes, this is a pretty simple build. The 40-pin GPIO connector on the Pi is attached to the 50-pin SCSI connector through a few 74LS641 transceivers with a few resistor packs for pullups and pulldowns. The software allows for virtual disk devices – either a hard drive, magneto-optical drive, or a CDROM – to be presented from the Raspberry Pi. There’s also the option of putting Ethernet on the SCSI chain, a helpful addition since Ethernet to SCSI conversion devices are usually rare and expensive.

Officially, [GIMONS] built this SCSI hard drive emulator for the x68000 computer, developed by Sharp in the late 80s. While these are popular machines for retrocomputing aficionados in Japan, they’re exceptionally rare elsewhere — although [Dave Jones] got his mitts on one for a teardown. SCSI was extraordinarily popular for computers from the 70s through the 90s, though, and since SCSI was a standard this build should work with all of them.

If your retrocomputer doesn’t need a SCSI drive, and you’re feeling left out of the drive-emulation club, the good news is there’s a Raspberry Pi solution for that, too: this Hackaday Prize entry turns a Pi into an IDE hard drive.

Thanks [Gokhan] for the tip!

A Real Hacker’s IDE

We don’t use a GUI IDE, but if we did, it would most certainly be something along the lines of [Martin]’s embedded-IDE project. We’ve always felt that most IDEs are just fancy wrappers around all the tools that we use anyway: Makefiles, diff, git, ctags, and an editor. [Martin]’s project makes them less fancy, more transparent, and more customizable, while retaining the functionality. That’s the hacker’s way — putting together proven standard tools that already work.

The code editor he uses is QScintilla, which uses clang for code completion. The “template” system for new projects? He uses diff and patch to import and export project templates. Because it uses standard tools all along the way, you can install the entire toolchain with sudo apt-get install clang diffutils patch ctags make on an Ubuntu-like system. Whatever compiler you want to use is supported, naturally.

We can’t see a debugger interface, so maybe that’s something for the future? Anyway, if you want a minimalistic IDE, or one that exposes the inner workings of what it’s doing rather than hiding them, then give [Martin]’s IDE a try. If you want more bells and whistles that you’re not going to use anyway, and don’t mind a little bloat and obscuration, many of our writers swear by Eclipse, both for Arduino and for ARM platforms. We’ll stick to our butterflies.

Ask Hackaday: How Do You Python?

Python is the Arduino of software projects. It has a critical mass of libraries for anything from facial recognition and neural networks to robotics and remote sensing. And just like Arduino, I have yet to find the killer IDE for Python. Perhaps I just haven’t tried the right one yet, but it could be that I’m just doing Python wrong.

For Years I’ve Been IDLE

IDLE with interactive shell that has highlighting and code completion

I’m a Linux-only type of a guy so using IDLE for Python is a natural fit. It’s in the repositories for super quick and easy install and there’s basically zero configuration to be done. Generally speaking my preferred development environment is text editor and command line compiler. IDLE is just one step above that. You get a separate window for the shell and each Python file you’re working on. Have IDLE run your code and it saves the file, then launches it in the shell window.

For me, there are two important features of IDLE’s shell. The first is that it keeps an interactive session open after you run your Python code. This means that any globals that your script uses are still available, and that you can experiment with your code by calling functions (and classes, etc) in real time. The second desirable feature is that while using this interactive shell, IDLE supports code completion and docstring support (it gives you hints for what parameters a function accepts/requires).

But simplicity has a tough time scaling. I’m working on larger and larger projects spread over many files and the individual nature of IDLE editor windows and lack of robust navigation has me looking to move forward.

The Contenders

I’ve tried perhaps a half-dozen different Python IDEs now, spending the most time on two of them: Geany and Atom. Both are easy to install on Linux and provide the more advanced features I want for larger projects: better navigation, cross-file code completion (and warnings), variable type and scope indication.

The look of Geany brings to mind an “IDE 1.0” layout style and theme. It’s the familiar three-pane layout that places symbols to the left, code to the right, and status along the bottom. When you run your program it launches in an interactive terminal, which I like, but you lose all IDE features at this point, which I despise. There is no code completion, and no syntax highlighting.

I have been using Atom much more than Geany and have grown to like it enough to stick with it for now. I’d call Atom the “IDE 2.0” layout. It launches with a dark theme and everything is a tab.

Atom has symbol view that isn’t shown all the time. CTRL-R brings it up and it uses a search style but you can also scroll through all symbols

Atom depends heavily on packages (plugins that anyone may write). The package management is good, and the packages I’ve tried have been superb. I’m using autocomplete-python and tabs-to-spaces, but again I come up short when it comes to running Python files. I’ve tried platformio-ide-terminal, script, and runner plugins.  The first brings up a terminal as a bottom pane but doesn’t automatically run the file in that terminal. Script also uses a bottom pane but I can’t get it to run interactively. I’m currently using runner which has an okay display but is not interactive. I’ve resorted to using a “fake” python file in my projects as a workaround for commands and tests I would normally run in the interactive shell.

Tell Us How You Python

It’s entirely possible I’ve just been using Python wrong all these years and that tinkering with your code in an interactive shell is a poor choose of development processes.

What do you prefer for your Python development? Does an interactive shell matter to you? Did you start with IDLE and move to a more mature IDE. Which IDE did you end up with and what kind of compromises did you make during that change. Let us know in the comments below.

Hackaday Prize Entry: Real Hard Drives In The Raspberry Pi

‘Boy, I wish the Raspberry Pi had a SATA port’. This is the plea that echoes through the Internet, and for once, the Internet is not wrong. A SATA port — or any connector to a big, dumb spinny disk — would be a great addition to the Raspberry Pi ecosystem.

[AlanH]’s entry to the Hackaday Prize is the exact opposite of what everyone wants. The NetPi-IDE is a Parallel ATA IDE disk emulator that turns an inexpensive Raspi Zero into a big, dumb, unspinny hard drive. Drop this machine in your Windows 98 Starcraft battlestation, and you have a hard drive that you can ssh into.

As with any build involving a lot of data, bandwidth is important. The highest bandwidth interface on the Pi’s GPIO ports is the SPI interface. [AlanH] is hanging a Lattice MachXO2 FPGA off the SPI port and using that to emulate a disk. In the future, he’s going to move to the much more open Lattice iCE40HX, compatable with the Open Source IceStorm synthesis chain.

The feature set for this project includes proper IDE disk emulation with sizes ranging from 10 Megabytes to 8 Gigabytes tested so far. If you need anything bigger, you don’t need an IDE drive. A DOS redirector allows mounting any arbitrary directory to a DOS drive letter, a virtual network interface turns this project into The Cloud™, and a serial console is mapped to unused IDE registers, allowing any host system to login to the Pi without any external cables.

While it’s not what everyone wants in a Pi, this is an exceptionally cool project. PATA drives are getting old, and the systems supporting them are too. If you want to keep those retrocomputers running, we have to start planning now, and there’s no better way to do that than with cheap hardware and Open Source toolchains.

PlatformIO and Visual Studio Take over the World

In a recent post, I talked about using the “Blue Pill” STM32 module with the Arduino IDE. I’m not a big fan of the Arduino IDE, but I will admit it is simple to use which makes it good for simple things.

I’m not a big fan of integrated development environments (IDE), in general. I’ve used plenty of them, especially when they are tightly tied to the tool I’m trying to use at the time. But when I’m not doing anything special, I tend to just write my code in emacs. Thinking about it, I suppose I really don’t mind an IDE if it has tools that actually help me. But if it is just a text editor and launches a few commands, I can do that from emacs or another editor of my choice. The chances that your favorite IDE is going to have as much editing capability and customization as emacs are close to zero. Even if you don’t like emacs, why learn another editor if there isn’t a clear benefit in doing so?

There are ways, of course, to use other tools with the Arduino and other frameworks and I decided to start looking at them. After all, how hard can it be to build Arduino code? If you want to jump straight to the punch line, you can check out the video, below.

Continue reading “PlatformIO and Visual Studio Take over the World”