There are more (much more) examples of young projects dying than old projects. This is called the Lindy effect: that which has survived tends to survive. Taleb first used the term Lindy effect but it has been noted before.
Taleb suggested that if something non-perishable has survived for a long time its expected remaining survival is just as long, for example if a book has been in print for 40 years it is expected that it wil remain in print for another 40 years.
The point is that projects, books, and other non perishables don't have a life expectancy like biological organisms, they're actually more likely to live on if they've endured a long time.
It's almost as if existing longevity is a sign of project community/support structures/resilience [1], and what matters for such an assessment is not a knee-jerk pointing to the existence of counter-examples to show that non-zero risk exists (as if anybody said anything about the risk being zero), but the relative probability of a fresh && much less used project dying vs a widely used mature project that has already proven it can survive for a long period of time...
It's almost as if you didn't read the first sentence in your own link, which states that this "is a theorized phenomenon", not some established law of practical software development nature. So instead of knee-jerk posting of a "proof" as a response to a criticism of a knee-jerk pointing to the existence of some risk or pointing to some straw man of zero risks you might actually realize that the main thing that matters for such an assessment is an actual assessment
It's almost as if you let the whole point fly over your head, and just hanged on the first thing you read on the link that could serve as a hail Mary attempt to build a semblance of an argument.
Yes, it's a "is a theorized phenomenon", it's not a "law of practical software development nature". The whole point is the statistical relevance of it on accessing such a risk - not it being some kind of absolute law.
It's the second time in this argument thread that you point to the lack of absolute guarantees (as if anybody said that long standing projects don't fail at all, or as if anybody said that the Lindy Effect is some absolute law), failing to see that they don't mean that long-term vs short-term project survivorship rate is the same.
You know that outcomes can have probabilities attached to them too, not just absolute guarantees, and that it's the former that are the most common tool for accessing risks, right?
>you might actually realize that the main thing that matters for such an assessment is an actual assessment
You might not actually realize that in practice any such an actual assessment will be based on risk factors, statistical observations, and handy heuristics for longevity such as the Lindy Effect.
> The whole point is the statistical relevance of it on accessing such a risk - not it being some kind of absolute law.
Someone letting the whole point fly over one's head again. You have no statistics! You're just using some empty law to support lazy thinking
> to the lack of absolute guarantees
That's just your second straw man. I know these "laws", were they true, wouldn't be absolute. You just can't understand that you don't have any data to support your claim, so when I point that out, you mislead yourself into thinking I demand absolutes
> To be fair, the Lindy effect does imply that it's less likely for a decades old project to die soon than a newer one.
True, but doesn't that imply its opposite as well? A decade old project will probably die sooner than later, because 11+ year projects are rarer than 10 year ones.
The Lindy effect can only be observed in comparison to something else, not to deduct how long a single project in and of itself will last. Which means, coreutils will probably last longer than this one, because it's been around 33 years vs 10.
The "original comment" basically made the Lindy effect argument, stopping short of naming it explicitly:
"The issue with all these efforts is whether they'll be sustained and maintained long term, or merely until the 1-2 maintainers lose interest. GNU coreutils on the other hand have been going for decades".
So there's that.
Many useful tools are "just theories", Occam's razor included.
Nope, "until the 1-2 maintainers lose interest" is a specific cause of project death, not a general observation that the risk is LOWER. Just like "a 40 years project dies because the 1-2 maintainers retire" is a cause of death for old projects would also be a similar risk, but just as useless for comparison.
No need to try to fit everything into some simplistic theory as though it's a Procrustean bed
> Many useful tools are "just theories"
As are many harmful tools, and you haven't demonstrated that the Lindy law is useful
GNU coreutils on the other hand have been going for decades.