3 ms·

> open source groups should not be unconditionally supporting the ecosystem for a large corporation’s hardware and/or platforms. Then by this logic I'm going t

> open source groups should not be unconditionally supporting the ecosystem for a large corporation’s hardware and/or platforms. Then by this logic I'm going to stop developing against pretty much every ISA aside from maybe RISC-V, and even that's a stretch. x86 users can piss off until Intel, AMD, and/or VIA cough up some dough to my single-digit-user-count FOSS projects. ...obviously this idea, if taken to its logical conclusion, basically means the death of open source software development as we know it. The author had a decent point until not only suggesting to refuse to support "niche" platforms entirely, but to make that suggestion the "most important". As a counterexample to the author's point, see OpenBSD's support for platforms unsupported by Rust. The rationale there is not for monetary gain, but for two key reasons: 1. People want to be able to run OpenBSD on whatever hardware they've got lying around, and when they figure out how to do it they might as well help others do the same. 2. In line with the author's point about how most C developers make platform-specific assumptions about the code they write, OpenBSD targeting a bunch of "niche" platforms with oddball conventions helps catch those assumptions early - and with them, any lurking security bugs deriving from those assumptions failing.

ohgodplsno 6y ago
Sigh. The HackerNews standard of pushing some logic to its limits and going "HA, GOTCHA" is getting tiring. x86, x86_64, ARM are used by billions of people around the world. It's easy to see that there is actual value for end users by supporting these platforms. John Doe from accounting can easily use your open source software and it will benefit him. The clear benefits of supporting z/OS, aside from helping IBM push out more of their mainframes and making more money, are unclear. Supporting PA-RISC, POWER or any other esoteric architecture whose sole purpose is to make their vendors sell million dollar support contracts should not be anyone's priority, and if indeed they do want this software on their non-standard, non-widely used ISA, they can use some of these millions to fund LLVM development.
- yellowapple 6y ago
  > Sigh. The HackerNews standard of pushing some logic to its limits and going "HA, GOTCHA" is getting tiring. So is the Hacker News standard of ignoring the entire point just to nag about said supposed standard. Sigh. Said entire point, specifically, being that there are far more reasons to develop and maintain software beyond "value for end users", and that even when targeting that specific reason, developing against esoteric cases does ensure the software is more robust even for "normal" use cases. (And mind you, logic - like software - should be pushed to its limits, because that's the most surefire way to identify its flaws - like, for example, the failure to consider that nearly all computing platforms in existence and use today were and are the product of some large corporation that could be paying your bills but won't, or the failure to consider that it's users, not hardware vendors, who are generating support requests for their hardware)

ohgodplsno 6y ago

Sigh. The HackerNews standard of pushing some logic to its limits and going "HA, GOTCHA" is getting tiring. x86, x86_64, ARM are used by billions of people around the world. It's easy to see that there is actual value for end users by supporting these platforms. John Doe from accounting can easily use your open source software and it will benefit him. The clear benefits of supporting z/OS, aside from helping IBM push out more of their mainframes and making more money, are unclear. Supporting PA-RISC, POWER or any other esoteric architecture whose sole purpose is to make their vendors sell million dollar support contracts should not be anyone's priority, and if indeed they do want this software on their non-standard, non-widely used ISA, they can use some of these millions to fund LLVM development.

yellowapple 6y ago
> Sigh. The HackerNews standard of pushing some logic to its limits and going "HA, GOTCHA" is getting tiring. So is the Hacker News standard of ignoring the entire point just to nag about said supposed standard. Sigh. Said entire point, specifically, being that there are far more reasons to develop and maintain software beyond "value for end users", and that even when targeting that specific reason, developing against esoteric cases does ensure the software is more robust even for "normal" use cases. (And mind you, logic - like software - should be pushed to its limits, because that's the most surefire way to identify its flaws - like, for example, the failure to consider that nearly all computing platforms in existence and use today were and are the product of some large corporation that could be paying your bills but won't, or the failure to consider that it's users, not hardware vendors, who are generating support requests for their hardware)