Re:Obscure election will decide the fate of the open internet
by tialaramex on Friday September 30, 2022 @05:30PM
(#33041058)
The ITU and IETF are very different. Although these are often lumped together as "Standards organisations" we should squint even at the word "Organisation" when it comes to the IETF.
The IETF isn't an organisation - like the ITU or a Boy Scouts troop - it's an Activity, like dancing. Humans can participate in the IETF if they want, corporations cannot (corporations also can't dance) and since it isn't an organisation the IETF has no members and writes documents reflecting the consensus of participants.
In this particular case an even more important difference is their approach to standardisation itself, the ITU produces "de jure" standards. The organisation produces documents which say how something shall be and that has the force of law. ITU members are sovereign entities, so they have police forces and armies, in principle they can choose to "enforce" the standards they've agreed, and only in practice do we end up relying on the fact that you might as well implement say I.762 as make up some other way to do it.
The IETF produces documents how participants agreed things ought to work, but sometimes that's how they really work ("de facto" standards) and sometimes it isn't. For example the IETF published a document about how early browsers wanted Cookies to work, which briefly mentions how Netscape Navigator's first implementation did it, assuming this will quickly be superseded. It wasn't, the IETF RFC about cookies is only really useful today as either an aspiration or because it obliquely documents what you actually need to do, mimic Netscape (and guess which shortcuts work and which do not).
The ISO and ITU approach works great where the main objective is to agree something and what it is exactly is secondary. For example the A-series paper standards from ISO are very nice, but if ISO had standardised what is now "US Letter" instead that would still be pretty useful even though it's an ugly shape without the nice properties of the A-series. However as you get into technical details this approach works much less and less well.
In contrast the IETF approach works better for these details, because of the objective of consensus rather than an insistence on democracy, it will usually identify when there actually just isn't any agreement and standardise only what's actually agreed - the cookies example I mentioned is an example where this didn't work, but it's not that common.
This has led to some conflict with "traditional" standards organisations which feel like they ought to just be able to write a document saying "No, it works like this" and have that take effect. Whether its ECMA, or the ITU, or even ISO this doesn't tend to go well for them but you can imagine it's frustrating even as they operate a liaison mechanism, to find that somehow the IETF is good at this while they can't comprehend how it even works.
The IETF isn't an organisation - like the ITU or a Boy Scouts troop - it's an Activity, like dancing. Humans can participate in the IETF if they want, corporations cannot (corporations also can't dance) and since it isn't an organisation the IETF has no members and writes documents reflecting the consensus of participants.
In this particular case an even more important difference is their approach to standardisation itself, the ITU produces "de jure" standards. The organisation produces documents which say how something shall be and that has the force of law. ITU members are sovereign entities, so they have police forces and armies, in principle they can choose to "enforce" the standards they've agreed, and only in practice do we end up relying on the fact that you might as well implement say I.762 as make up some other way to do it.
The IETF produces documents how participants agreed things ought to work, but sometimes that's how they really work ("de facto" standards) and sometimes it isn't. For example the IETF published a document about how early browsers wanted Cookies to work, which briefly mentions how Netscape Navigator's first implementation did it, assuming this will quickly be superseded. It wasn't, the IETF RFC about cookies is only really useful today as either an aspiration or because it obliquely documents what you actually need to do, mimic Netscape (and guess which shortcuts work and which do not).
The ISO and ITU approach works great where the main objective is to agree something and what it is exactly is secondary. For example the A-series paper standards from ISO are very nice, but if ISO had standardised what is now "US Letter" instead that would still be pretty useful even though it's an ugly shape without the nice properties of the A-series. However as you get into technical details this approach works much less and less well.
In contrast the IETF approach works better for these details, because of the objective of consensus rather than an insistence on democracy, it will usually identify when there actually just isn't any agreement and standardise only what's actually agreed - the cookies example I mentioned is an example where this didn't work, but it's not that common.
This has led to some conflict with "traditional" standards organisations which feel like they ought to just be able to write a document saying "No, it works like this" and have that take effect. Whether its ECMA, or the ITU, or even ISO this doesn't tend to go well for them but you can imagine it's frustrating even as they operate a liaison mechanism, to find that somehow the IETF is good at this while they can't comprehend how it even works.