Hacker News Reader

Comment by Y_Y | Hacker News Reader

> I think the naive estimate is if you pick something at random it is half-way through its lifespan; so there isn't much call yet to say Debian will make it to 100.

This doesn't strike me as a strong argument. That naive estimate (in whatever form[0]) is typically based on not knowing anything else about the process you're looking at. We have lots of information about Debian and similar projects, and you can update your estimate (in a Bayesian fashion) when you know this. Given that Ian Murdock started Debian 31 years ago I think more than 100 years is a very reasonable guess.

[0] see e.g. https://en.wikipedia.org/wiki/Lindy_effect

View on HN

Arguably, there is already the continuous package deprecation process that often leads to unpopular projects getting culled in the next upgrade.

In a way, Flatpak/Snap/Docker was mitigations to support old programs on new systems, and old systems with updated software no longer compatible with the OS. Not an ideal solution, but a necessary one if folks also wanted to address the win/exe dominant long-term supported program versions.

If working with unpopular oddball stuff one notices the packages cycle out of the repositories rather regularly. =3

I appreciate the Lindy effect, but I'd be very cautious in applying it in just any domain. In particular IT, where new project continually spring up to dethrone others. Another 30 years for Debian seems reasonable, but I'd probably bet against another 100. A metaculus question for the longevity of projects like debian would be fascinating.

> In particular IT, where new project continually spring up to dethrone others.

The Lindy effect says nothing about popularity, which is how I translate your use of 'dethrone' here. It observes that something's duration of existence correlates with its chances for existence in the future.

> In particular IT, where new project continually spring up to dethrone others

Yet, it's lasted 31 years, which is a pretty insane amount of time in tech. This on top of being kept up to date, good structure, really good contributions and advancements.

On the other hand, you look at centos, redhat, and oracle and their debacle. How much did they fragment that area.

And then we have Debian just chugging along.

Reply on news.ycombinator.com