Re:Hardware Hackathons Are Hard
by nicholas73 on Friday January 31, 2014 @03:15PM
(#7158914)
As a hardware engineer, let me define what hard is when it comes to hardware vs. software. I wouldn't say hardware is harder in any way conceptually although of course both have degrees of difficulty. The main difference is that the "activation energy" of each step is greater, so you have to be more confident, deliberate, and determined when you are building.
A step like assembling components and wiring together a breadboard is small but non-trivial amount of effort. So you won't find many people who just play and see what happens (which probably won't work anyways).
A step like designing a PCB is as long as designing any piece of software, except you don't get to see the results until you're done. Well unless you have a fancy expensive commercial software that does simulation as well as layout. But it's still far from quickly throwing up a webapp and getting motivated with insta-results.
You might get some inspiration, then have to horribly dig through Digikey or some other catalog to find your parts, then wait for them to be shipped.
You might also find that some parts are not cheap. Not unaffordable for an engineer, but maybe you'd prefer the nice dinner over a slight probability of a future product. So time you could spend building sometimes end up being used to decide whether to spend the money.
Basically, in hardware you get bogged down in details that have little to do with designing, whereas in software you can just sit down and go.
Re:Hardware Hackathons Are Hard
by leoedin on Friday January 31, 2014 @03:22PM
(#7158957)
If you want to appear to have done a lot in a hardware hackathon, you need to do as little hardware as possible. Take a few preexisting modules and start writing software as soon as possible, because developing new electronics - even simple circuits - is hard and incredibly slow paced. Even once you've got 4 or 5 figures worth of lab equipment stacked up next to you, you're still going to be tearing your hair out because something isn't working like it should, and the tools available are very, very limited.
I suppose it might be fun to spend a weekend combining a few of the cheap sensor chips available with an rPi to make a web-enabled sensor of some variety, but the scope for what you can do is so much smaller. The novel stuff will inevitably be in software. If you want to do something really interesting it will take weeks or months (and often years). Chips will die unexpectedly. That seemingly simple power supply circuit won't work. You'll short something out and the blue smoke will taunt you. It can easily take most of a day just to connect up a fairly simple prototype board. You can spend a week slowly building a wiring harness.
Re:Hardware Hackathons Are Hard
by cushychicken on Friday January 31, 2014 @12:24PM
(#7157789)
While I think his conclusion is pretty correct, I have to disagree with some of his premises, particularly where he says prototyping is difficult (or more difficult than software). My reasoning is mostly based on the fact that a great product (or the proof of concept for one) can be based around something like UI/UX improvement, and that's pretty easy to achieve with modern technology (Nest and Fitbit both being great examples of this - simple concepts well executed in reasonably simple hardware platforms). There's tons of capacity to make some really cool HW prototypes in a weekend, or even an afternoon - we do live in the age of the $50 Linux Dev board, after all.
Re:Hardware Hackathons Are Hard
by joezydeco on Friday January 31, 2014 @01:14PM
(#7158109)
There's the catch, don'tcha think?
We've reached the intersection of cheap/easy development boards combined with a business philosophy of "slap this API on this premade stack glued to this other stack running on this VM and we're ready for our Series A".
Adafruit and Sparkfun and the like with all the cheap and easy-to-connect modules have enabled us. They've led us to believe that hardware hacking is easy. Just wire this camera to this accelerometer to this Rpi and we're in business!
Does anyone really expect to be etching PCBs in a hackathon? Is a simple breadboard setup good enough to win?
Re:Hardware Hackathons Are Hard
by logicallee on Friday January 31, 2014 @01:40PM
(#7158301)
Out of pure curiosity to see where you're coming from - when you say "I have to disagree with some of his premises, particularly where he says prototyping is difficult (or more difficult than software)", have you ever imagined, designed, and built 1) a piece of software 2) a piece of hardware?
I don't have preconceptions about what your answers will be, just curious.
Re:Hardware Hackathons Are Hard
by cushychicken on Friday January 31, 2014 @02:11PM
(#7158533)
I've done both - I have a background in EE, verging on CompE. I've only ever done hardware professionally, however.
I think my view of a prototype is more in line with what an industry would view as a proof of concept - enough hardware to show that your vision is possible. I wouldn't want to show anything I built over a weekend to a VC in an effort to get them to fund me, unless I absolutely had to.
Re:Hardware Hackathons Are Hard
by CompleteMoron2 on Friday January 31, 2014 @01:07PM
(#7158057)
they arent hard - you know whats hard?
Realizing the winners always already had a prototype (software or hardware) and really just practice their sales pitch for two days for the win
But hey - we now all know you dont win by actually starting the day of the hackathon
A step like assembling components and wiring together a breadboard is small but non-trivial amount of effort. So you won't find many people who just play and see what happens (which probably won't work anyways).
A step like designing a PCB is as long as designing any piece of software, except you don't get to see the results until you're done. Well unless you have a fancy expensive commercial software that does simulation as well as layout. But it's still far from quickly throwing up a webapp and getting motivated with insta-results.
You might get some inspiration, then have to horribly dig through Digikey or some other catalog to find your parts, then wait for them to be shipped.
You might also find that some parts are not cheap. Not unaffordable for an engineer, but maybe you'd prefer the nice dinner over a slight probability of a future product. So time you could spend building sometimes end up being used to decide whether to spend the money.
Basically, in hardware you get bogged down in details that have little to do with designing, whereas in software you can just sit down and go.
I suppose it might be fun to spend a weekend combining a few of the cheap sensor chips available with an rPi to make a web-enabled sensor of some variety, but the scope for what you can do is so much smaller. The novel stuff will inevitably be in software. If you want to do something really interesting it will take weeks or months (and often years). Chips will die unexpectedly. That seemingly simple power supply circuit won't work. You'll short something out and the blue smoke will taunt you. It can easily take most of a day just to connect up a fairly simple prototype board. You can spend a week slowly building a wiring harness.
Half of the time is spent removing and remounting the chips and inspecting the joints with a microscope
Half of the time is spent querying the device using the uC as your host, and fearing that the problem is in your test software and not the solder job.
If your goals require you to drill a hole and you don't have access a drill your options are fairly limited!
We've reached the intersection of cheap/easy development boards combined with a business philosophy of "slap this API on this premade stack glued to this other stack running on this VM and we're ready for our Series A".
Adafruit and Sparkfun and the like with all the cheap and easy-to-connect modules have enabled us. They've led us to believe that hardware hacking is easy. Just wire this camera to this accelerometer to this Rpi and we're in business!
Does anyone really expect to be etching PCBs in a hackathon? Is a simple breadboard setup good enough to win?
Yeah, I'd say so. As I've mentioned to other folks in this thread, I feel hardware hackathon entries are more "proof of concept" than "prototype".
I don't have preconceptions about what your answers will be, just curious.
I think my view of a prototype is more in line with what an industry would view as a proof of concept - enough hardware to show that your vision is possible. I wouldn't want to show anything I built over a weekend to a VC in an effort to get them to fund me, unless I absolutely had to.
But hey - we now all know you dont win by actually starting the day of the hackathon