Responses from Parallella user survey

private_responses

The responses from our recent open response Parallella customer survey are in. Thank you to the 300+ of you who filled out the survey! Your feedback will help all us improve Parallella for all 10,000 Parallella users. The open response question was simple:

“What do you need in order to get going with your Parallella project?”

As a result of this survey we will be taking the following actions:

  • Improve existing examples and continue to build out repository
  • Organize all of our documentation into something more digestible
  • Continue to make Parallella easier to use
  • Accelerate the build up of Parallella daughter cards and accessories
  • Stretch time (more about this later…)

Cheers,

Andreas

4 Comments

  • B-J Velthuis says:

    Hello Andreas,

    Thanks for sharing this information back with us.

    You have often referred to the overwhelming success of the Raspberry Pi. Undoubtedly this is in a great part due to the undoubtedly enormous time commitment of both Liz and Eben Upton and their entire team to spread the knowledge on the various and many uses of the Pi. Documentation of examples and an enthusiastic style of presentation is key. Even the Korean based company Hardkernel was quick to pick up on the great idea of ‘MagPi’. Some of the more ordain everyday uses of the Parallella board could well be accomplished by, say, a Pi 2 or an Odroid C1, or any of the many other SBCs, so as a community, we should follow the Pi Foundation’s and Hardkernel’s lead to improve the Parallella Project’s documentation base. The Parallella Chronicles is obviously a good start in this direction, but as a community we still have a long way to go and catch up. Shodruky did an excellent job in giving examples of what we could do. But, more every day use examples of the parallella board, that may not necessarily involve the epihany chip at all, may well be followed more readily by others, most of whom may lack the (parallel) programming skills altogether.

    At some point the Parallella community forum may become too extensive and unwieldy for newcomers to find their way. We really need educators and journalists to reshape much of the already available information to make it more patable for others, be it many of the existing 10.000 or new Parallella owners. Perhaps it might be a good idea to get in touch with the good people of the Pi foundation? I am sure that, say, Andrew is familiar with the plethora of Acorn magazines from several decades ago. This really helped to create an active community. And whatever happened to the suggestion very early on, based on an article from the O’Reilly camp, that someone might be writing a book? Springer has made a huge success out of their use R! series, I think. Perhaps someone could convince those publishers that is the time for a ‘use SBCs’ series. I know, this is thinking out of the box. But it might be a way forward.

    Best wishes,

    Berend-Jan

  • A super computer needs a super application. If bandwidth is a small feat, let’s create a virtual host that run LINUX, ANDROID, and Arduino. The Parallalla learning curve is too steep for the general technology user. The ideal market task is to have this product in every dorm room and in every High School lab. A language such as PYTHON that is portable and well supported in the Open Source community should be considered. I recently purchased pcDuino. If 3 to 4 documents are on hand, with the software pre-installed, boot,up and connecting to the internet can happen within a few hours.

    Has any one used The Zynq Book?

    Andrew

  • Len Reinhart says:

    If you had a book like “Exploring Beaglebone” you could attract a lot more of the Raspberry Pi community. Then with a nice book and a Parallella board the interested could sit down and follow the examples. There is probably enough material on the forum, you just need to get permission to use it and an editor. A book like that would make it more accessible to the universities too. It wouldn’t be such an investment of time to teach a course based on the Parallella board. Don’t restrict it to focus on the parallel processor, the FPGA is good subject matter too. Same with the dual ARM core processor and the various hardware like the A/D converter.

  • Sean Halle says:

    Adoption is driven by promise and retarded by barriers. The many do-it-yourself steps act as barriers. For example, burning the SD card is a barrier, as is ordering a power supply, and installing in a cabinet with a fan. Each one of these steps represents a risk that I might do it wrong. Altogether they are each simple, but as a mass, the risk and time required until you have a prompt on a screen accumulates to a non-trivial compound barrier.

    What Raspberry Pi and Arduino have done is eliminate nearly all such “inconsequential” barriers. They are turn key. Literally 2 minutes after opening the package for my first Arduino I had the LED blinking. That is what I call low barrier! I have the intent to tackle getting the parallela up and running, but frankly, I just don’t have the time needed to do all those little things.

    What I need to reduce these barriers is a package sold on the site, that bundles the HDMI cable, the power supply, the enclosure, the fan, and a burned ready-to-go SD card that boots. And another package that bundles a display, mouse, and keyboard. I want to just buy them and get going. Even better, offer the whole thing assembled and ready to go. It comes with display, keyboard, mouse, power supply, and Parallela already installed, with a burned SD card, inside an enclosure with a fan, a powered USB hub, and a wifi transceiver. Just plug the wires and get a prompt. 5 min time from open package to seeing a prompt. Many of us would pay for the service of providing that hand holding. We just want to see it work.. and don’t have the time.

    In addition, believe that the best way to increase uptake of the Parallela is to make it the equivalent of a super computer arduino. Imagine embedding it into projects, and they gain the ability to recognize your face, or can track individual objects, or can learn your preferences. Any project that can be done with an Arduino should be doable with a Parallela card.. but with machine learning.. and with video processing.. and with voice recognition available. Make it a viable embedded micro controller with huevos.

    So, to gain adoption, on one side is reduction of barrier by offering packages that do the little steps for us. On the other side is offer a demo for a product that excites us.

    For example, imagine demo software takes video images via a USB webcam, learns our face, and greets us by name when we sit down in front of it. That alone would be cool just to show our friends. And then the pieces can be used in our own projects, such as making a door bell that recognizes our friends, and alerts us to the identity of the person at the door. We could make a sentry for our kid’s playhouse that recognizes their friends, and talks to the kids, and so on.
    Or imagine if it understands voice commands, and learns our preferences. We can then use that in our own projects as well. Suddenly the Parallela can turn things on and off via voice command, and recognize the speaker by video or voice, it can generate things to look at on the display. It can learn by example. And it is embeddable into things that we make.

    Heck, I did a robot with my son, and he kept asking to be able to talk to it, and wanted to be able to teach it things by manipulating it, rather than coding. The Arduino doesn’t have anywhere near the computation needed. But the parallela does.

    Each of those things mentioned has a known software solution, with open source available. Micro controllers just lack sufficient processing power to make them mainstream. Having those kinds of things available gives some real, everyday, value to having a parallela card. Real things that we would like to have, and attract us, motivate us to find even more uses.

    Honestly, the idea of yet another stand alone computer was maybe a cool idea, but it isn’t practical. The excitement doesn’t last. After the initial thrill, then it becomes a chore to move over the the parallela system. You don’t have your files, you don’t have your setup. You have to administer two separate systems. You have to learn Ubuntu. It just doesn’t encourage long term use. And, for what? What do you do with it?

    Much better would be to have the parallela as a slave, with my personal computer being the main environment. The IDE runs on my machine, the display is on my machine, the keyboard is my machine. The parallela is only used for its compute power. And ideally, as something I can embed into projects or products. Something that has an amazing human interface because it has the processing power to do all that on a battery, or in a fanless case.

    In the end, it needs to be minimum barrier and also useful. The less time to get going, the lower the barrier. The easier to get it to do something neat, the higher the attraction. Embedding it into projects to give them voice and video and learning makes it useful.

Leave a Reply