Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

What's the advantage of standardizing through ISO/IEC? Better adoption in industry?

Seems like this would take away a lot of power from RISC-V International. But I don't know much about this process.



Government agencies like to take standards off the shelf whenever they can. Citing something overseen by an apolitical, non-profit organization avoids conflicts of interest (relative to the alternatives).

Random example I found at a glance: NIST recommending use of a specific ISO standard in domains not formally covered by a regulatory body: https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.S...


It's impossible to take ISO seriously after the .docx fiasco.


That’s the definition of throwing the baby out with the bath water.

Is ISO as an organisation imperfect sometimes (as in the docs case) sure?, it’s composed of humans who are generally flawed creatures, is it generally a good solution despite that?, also sure.

They’ve published tens of thousands off standards over 70 plus years that are deeply important to multiple industries so disregarding them because Microsoft co-opted them once 20 odd years ago seems unreasonable to me.


What .docx fiasco?


Office Open XML, the standard behind .docx and other zipped XML formats, was fast-tracked into the international standard without many rounds of reviews (by the same JTC 1!).


> Citing something overseen by an apolitical, non-profit organization avoids conflicts of interest (relative to the alternatives).

Of course this is a lie. But yes, governments like to claim that.


As the article says:

> “International standards have a special status,” says Phil Wennblom, Chair of ISO/IEC JTC 1. “Even though RISC-V is already globally recognized, once something becomes an ISO/IEC standard, it’s even more widely accepted. Countries around the world place strong emphasis on international standards as the basis for their national standards. It’s a significant tailwind when it comes to market access.”


Says that, but I don't agree with that. If anything it would have been less successful being picked up in discount markets if the specs weren't free for download, and I don't know what fringes they're trying to break into but probably none of them care whether the spec is ISO.


That can depend on how the spec gets made into an ISO standard. There is a process called "harvesting" that can allow the original author to continue to distribute an existing specification independently of ISO.


> Says that, but I don't agree with that

I guess you just never had to fill in a grant application where you have to justify that you are using official standards so that you can get money


I'm guessing in those kinds of situations it doesn't matter about the arch given x86 and ARM also aren't ISO standards. The manufacturers however should comply with relevant quality standards.


it doesn't matter when there is no ISO standard for a given tech. But as soon as there is one, then you have to provide arguments as to "why didn't you use the standard".


Usual lies. There are a plethora of largely ignored international standards. Making it an international standard is just one of many ways to achieve the wide worldwide acception and still has a high failure rate.


My take is that it could help tie up fragmentation. RISC-V has different profiles defining what instructions come with for different use cases like a general purpose OS, and enshrining them as an ISO standard would give the entire industry a rallying point.

Without these profiles, we are stuck with memorizing a word soup of RV64GCBV_Zicntr_Zihpm_etc all means


riscv was already gaining a profile mechanism outside of ISO, for example 'RVA23' is a known set of extensions


Hardly, see programming languages standards and compiler specific extensions.


languages are more fluid than processor architectures. I don't think they can be compared.


One would think, yet welcome to enterprise consulting, especially customers whose main business is not selling software.

You will find fossilized languages all over the place.


fossilised is often desirable or requested in some industries. Developing for the embedded market myself, we often have to stick to C99 to ensure compatibility with whatever ancient compiler a costumer or even chipset vendor may still be running.


RISC-V never had a fragmentation problem, thanks to the profiles.


I wouldn't say it never had a problem, but the profiles are definitely a reasonable solution.

However even with profiles there are optional extensions and a lot of undefined behaviour (sometimes deliberately, sometimes because the spec is just not especially well written).


The FUD keeps being brought up, but the solution here was in place before the potential issue could manifest.

It started with G, later retroactively named RVA20 (with a minor extra extension that nobody ever skipped implementing), then RVA22 and now RVA23. All application processor implementations out there conform to a profile, and so do the relevant Linux distributions.

Of course, in embedded systems where the vendor controls the full stack, the freedom of micromanaging which extensions to implement as well as the freedom to add custom extensions is actual value.

The original architects of the ISA knew what they were doing.


Maybe it helps get government contracts

“We’re standards compliant”


Sometimes it helps, sometimes it doesn't. Like when Sun Microsystems submitted ODF for standardization to ISO, it was so successful that Microsoft had to do it too for OOXML. In fact MS pushed so hard that it left a huge trail of destruction in the standards committees.

Other times, like with the "ISO power plug", the result was ISO/IEC 60906-1 which nobody uses. Swiss plugs (IEC Type J), which this plug is based on, use a slightly different distance for the ground pin, so it is incompatible. Brazil adopted it (IEC Type N) but made changes to pin diameter and current rating.


It's not like ARM and x86 are standardised by ISO either.


Governments seem to care about "self-sufficiency" a lot more these days, especially after what's happening in both China and the US right now.

If the choice is between an architecture owned, patented and managed by a single company domiciled in a foreign country, versus one which is an international standard and has multiple competing vendors, the latter suddenly seems a lot more attractive.

Price and performance don't matter that much. Governments are a lot less price-sensitive than consumers (and even businesses), they're willing to spend money to achieve their goals.


This is exactly what makes this such an interesting development. Standardization is part of the process of the CPU industry becoming a mature industry not dependent on the whims of individual companies. Boring, yes, but also stable.


Yes, and they're both massively debated and criticised, to the point that the industry developed Risk-V in the firstplace. Not to mention the rugpull licensing ARM pulled a few years back.


Yes, but if 30 years ago ARM had an ISO standard they could point to, that would have probably helped with government adoption?

(It's still a trade-off, because standards also cost community time and effort.)


Relatedly, 30 years ago someone attempted to turn the Windows 3.1 API into an ISO standard:

https://en.wikipedia.org/wiki/Application_Programming_Interf...

It didn't become one, but it did become standardised as ECMA-234:

https://ecma-international.org/publications-and-standards/st...


Well, Wine shows that Win32 is the only stable ABI, even on Linux.


>On May 5, 1993, Sun Microsystems announced Windows Application Binary Interface (WABI), a product to run Windows software on Unix, and the Public Windows Interface (PWI) initiative, an effort to standardize a subset of the popular 16-bit Windows APIs.

>In February 1994, the PWI Specification Committee sent a draft specification to X/Open—who rejected it in March, after being threatened by Microsoft's assertion of intellectual property rights (IPR) over the Windows APIs

Looks like that's what it was.


they are de-facto…


It ticks a checkbox. That's it. Some organizations and/or governments might have rules that emphasize using international standards, and this might help with it.

I just hope it's going to be a "throw it over the fence and standardize" type of a deal, where the actual standardization process will still be outside of ISO (the ISO process is not very good - not my words, just ask the members of the C++ committee) and the text of the standard will be freely licensed and available to everyone (ISO paywalls its standards).


> the ISO process is not very good - not my words, just ask the members of the C++ committee

Casual reminder that they ousted one of the founders of MPEG for daring to question the patent mess around H.265 (paraphrasing, a lot, of course)


This allows RISC-V international to propose their standards as ISO/IEC standards.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: