Author Topic: Lets discuss how to make OSHW more successful (inspiration by FOSS design cycle)  (Read 8426 times)

0 Members and 1 Guest are viewing this topic.

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3781
  • Country: de
The key to 'success' is actually make a finished product.
I like the whole idea of open source products, but most of them are just simply not finished products (very far from it)
To many (major) bugs, incompatibilities, horrendous interface and practically no service.

You basically are being send into the woods with half finished gear and good luck with it.
I think the mean reason is that most open source projects are being created by tech heads and engineers and mostly only serve a very specific niche.

How you organize your code, bug tracking, meetings etc etc are all just very minor details and formalities.
Unfortunately these things are also what people tend to focus on the most.
I have tried collaborating in a few projects, but quit for exactly these reasons.
A lot of times people also don't have time enough or aren't dedicated enough to really finish an idea.

The thing is, there is a big difference between developing a "product" (as in something that can be sold) and scratching own itch, where the possibility of making a bit of money on the side is only a bonus. Both are possible outcomes but it is important to distinguish between them.

And re: "but most of them are just simply not finished products (very far from it) To many (major) bugs, incompatibilities, horrendous interface and practically no service."

Sadly, that holds equally well about traditional (not FOSS/OSHW) products as well. Case in the point - the recent Feeltech mess with their self-bricking generators. Or Red Pitaya. Or basically any software that accompanies a hardware device (software tends to be an afterthought outsourced to the lowest bidder).

Service? Good luck getting any from a commercial vendor, even big names - like that Alain guy with his broken Logitech speaker set in another post here when he contacted Logitech - "Out of warranty? Take a hike, we don't do any service."

And here you are actually paying money for that, unlike with most FOSS/OSHW projects! So that is not really a good argument.
 

Offline niner_007

  • Frequent Contributor
  • **
  • Posts: 256
  • Country: us
The collaboration is easly done in software development by push and pull request and both the original auther and the contributor can easly verify the result by compiling the source code.
One note I want to add here. It's not really as easy as that. The contributed code has to be tested, that it has not broken or regressed the existing code, in some cases not just functionally, but performance as well. The new code has to use the same style and be consistent with the existing code. Licensing has to be consistent. New unit tests and functional tests have to be written for the contributed code, new documentation has to be written. Lots of process too, it can make things slow and bureaucratic. But necessary.

Perhaps a popular hardware project should have a reference development implementation (that can easily be ordered too), designed for extensibility and prototyping in mind, and that should allow easier verification. It should also have spice simulations, as much as possible to also verify things in software as much as possible.
« Last Edit: May 23, 2018, 11:28:47 am by niner_007 »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf