[tialaramex]

Ha, amateurs. The Haiku operating system project is almost seventeen years old and still hasn't shipped a beta.

Finding excuses is vital, but it's important to have both excuses for not shipping the beta and excuses for why beta is almost here and there's no need to think the past history of delays bodes ill for the project.

Here's a HN comment back in January where the poster says Haiku's beta will be out "this quarter" (Q1 calendar 2018), I laugh at that, and another HN poster says OK, but definitely by _next_ quarter (Q2 calendar 2018).

https://news.ycombinator.com/item?id=16149192

[renox]

Well, developing an OS is more work than a game That said those working on Haiku suffer from the same NIH syndrome as the article's author but unfortunately all the efforts of the 'pragmatics' one who wanted to reimplement BeOS on top of Linux failed much earlier than Haiku.. New OS are in a bad situation: those build on top of Linux doesn't seemn to attract contributors (I blame personally a lack of imagination: a kernel isn't an OS!) but those who use their own kernel are doomed for the lack of drivers..

[waddlesplash]

> but unfortunately all the efforts of the 'pragmatics' one who wanted to reimplement BeOS on top of Linux failed much earlier than Haiku

> New OS are in a bad situation: those build on top of Linux doesn't seemn to attract contributors (I blame personally a lack of imagination: a kernel isn't an OS!)

I think these two statements are, effectively, the same thing. Why have we (Haiku) stayed around so long, and all the others died? Somehow we managed to accumulate enough critical mass to not have heat death. I think the novelty of having a complete system is an attractive one, rather than "just another Linux distribution, why use that?" indeed.

And then, you know, maybe it's possible that what we've been saying this entire time about having a kernel and base system optimized for GUIs instead of just throwing one on top of them might actually be the case. :-p

> but those who use their own kernel are doomed for the lack of drivers..

We're so doomed, I've been able to boot, run, and use Haiku on just about all PCs I've tested it with so far!

But yes, drivers are an issue. An issue we have mostly solved at this point, with the exception of hardware-accelerated OpenGL, but still...

[renox]

Saying that reusing the Linux kernel is the same as being "just another Linux distribution" is the problem: most Linux distributions are just Unix clone: if you create a VMS, a BeOS, a MacOS or Plan9 with a Linux kernel then what you have is a different OS with a Linux kernel not just a Unix clone. Using the Linux kernel doesn't force you to be POSIX compatible (which is a double edged sword: you get a lot of software but you loose your "originality").

[fjsolwmv]

TempleOS did fine.

[waddlesplash]

You know, it's almost as if a project run entirely by volunteers, who have lives outside of it, entirely on their spare time, which fluctuates wildly from week-to-week, meaning that sometimes key developers who were previously available are now suddenly not, makes it anywhere from "difficult" to "functionally impossible" to give accurate date predictions; let alone ship a release...

[staz]

even E17 did finally end up releasing despite all the jokes

[antod]

And in that time Enlightenment magically transformed from a resource heavy desktop full of eye candy effects to one of the lighter weight options out there. All it took was waiting years for the other popular desktops to overtake it in resource usage.