I'm curious what industry you work in. Rarely have I run into a large problem that was tempting to solve with Stackoverflow/Github-driven development. I find the limiting factor is never "we don't have the code" and more "we don't know what code we need".
I certainly agree that there are many important aspects to software development, other than just code.
But I don't want to confuse the point, and accidentally be thinking "plagiarism of code isn't important, because code isn't important" (or the limiting factor).
For one reason, code is important, and valuable...
There's plenty of evidence throughout many types of software development to suggest that a lot of money has to be plowed into simply producing bulk of code -- excluding domain or market understanding, requirements analysis, holistic system design, etc.
A typical growth startup wasn't in a hiring frenzy for great analytical minds, but simply to scale up production of raw bulk of code (and connecting together legally-obtained off-the-shelf pieces). And they were often tracking quantitative metrics on that bulk, and consciously aware of how much all those code-monkey salaries were costing on the balance sheet. They paid it because code bulk was very necessary, the way they were working.
There's also plenty of evidence of the popularity of copy&paste reuse for many years. Including from StackOverflow.
Copying entire projects and subsystems has been much-much less common, partly due to the aforementioned not wanting to get caught doing it. But "AI" laundering is paving the way. See all the blog posts and videos about "I created an app/game/site in an hour with AI".
We can also look at startup schools of thought, where execution is widely regarded to be everything (ideas are a dime a dozen). There is plenty of thinking that churning out code is one of the most time-consuming or frequently-blocking parts of execution.
In both startups and established corporate environments, there are normally big lags between when a need for specific code is identified, and when that code can be delivered. Doesn't that look like a limiting factor, or at least it must be important.
I think that's enough reason that we not dismiss the value of code, nor dismiss the importance of plagiarism of code.
> accidentally be thinking "plagiarism of code isn't important, because code isn't important"
Based on my life experience from the last 15 years, you should assume that any "open source" code you leave online is going to be plagiarized heavily. It's unfortunate, but the hard truth for a billion different reasons.
In general, the legal peril of contaminated copyrights made closed-source desktop coding less lucrative. Most if not all current desktop applications from major commercial vendors is at least in part server hosted, or online subscription based.
It is simply a shift from shovel-ware to a service model, and does what DRM did for the music businesses.
Personally, I often release under Apache license as both a gift and a curse... Given it is "free" for everyone including the competition, and the net gain for bandits is negative three times the time cost (AI obfuscation would also cost resources.)
The balance I found was not solving corporate "hard" problems for free, and focusing on features that often coincidentally overlap with other community members. Thus, our goals just happen to align, and I am glad if people can make money recycling the work. =3
Why bandits are not as relevant as people assume:
"The Basic Laws of Human Stupidity" (Carlo M. Cipolla)
That’s awesome that you were able to make it work. But not everyone has the ability, either logistically or financially (or both), to enforce copyrights on open source code. So we should assume there will be some big bad actors who will absolutely take advantage of that.
Indeed, but the Apache license is quite different from BSD or GPL v3.
In my opinion, less restrictions on all users naturally fragments any attempt to lock-down generic features in the ecosystem.
Submarine Patent attacks using FOSS code as a delivery mechanism is far more of a problem these days, as users often can't purchase rights even if trying to observe legal requirements. The first-to-file laws also mean published works may be captured even if already in the public domain.
It is a common error to think commercial entities are somehow different from the university legal outreach departments. Yet they often offer the same legal peril, and unfeasible economic barriers to deployment.
Thanks, it's an interesting perspective. I may have underappreciated the code volume/time requirements to execute especially in a startup setting.
I think I see where you're coming from. I'd imagine in some applications, there is very niche/specific implementation on Github that is both expensive to reimplement and not permissively licensed. I can see the "AI laundering" angle there.
