RE: International identity data standards?

RE: It’s only when you get into connecting completely different systems together that you start to wish your “date.toString()” had worked the same way on both sides

+1

From: Justin Richer <jricher@mit.edu>
Sent: September 3, 2019 8:24 PM
To: Liam McCarty <liam@unumid.org>
Cc: public-vc-comments@w3.org; public-credentials@w3.org
Subject: Re: International identity data standards?

I’m going to be pessimistic here, and if you want rainbows and flowers, avert your eyes. So the way I see it, here’s the problem:

There has been a lot of work to create international standards for identity data. So many standards across so many domains that it’s easy to lose track of them all. But every effort to create The One True Ontology always, without fail, falls short of the most important feature of a standard: adoption.

Standards are only truly standards if people use them, and it’s far too easy to create a new data schema on the fly. Or even just create data without a schema, and figure that you’ll clean it up later if you need to align things. As has been said elsewhere in this thread, there are huge efforts and entire companies that do exactly that.

Don’t believe me? A simple question, then: how would you represent a date field? ISO8601? The combined timestamp or just the date portion? Are you using the “T” field separator or a space because it looks nicer (don’t laugh, I’ve seen it)? And what calendar system is this going to be in? Maybe that’s too weird, so how about epoch seconds? Epoch milliseconds? From which timezone? Or do we care about timezones at all? And is Taiwan a timezone? Depends on which country you ask from, on that one.  And if you ask most developers, it’ll be whatever “date.toString()” prints out for them.

Data is messy, normalization is hard. But coming up with a data format is easy. It’s so easy that everyone just kinda does it, even to the point of creating defacto standards that people get stuck with for years. You see, the hard part isn’t creating them, the hard part isn’t even using them, the hard part is getting everybody to use the same one for the same thing, consistently.

This is going to be true of any data standard no matter how good it is. I can tell you, this crowd loves schema.org<http://schema.org>, but most of the internet has never heard of it and doesn’t care about it or JSON-LD. I’m not making a judgement on how good any of the technology is, I’m simply saying that most of the world simply does not care that it exists because it by and large solves problems that they don’t see as problems. It’s only when you get into connecting completely different systems together that you start to wish your “date.toString()” had worked the same way on both sides, and at that point it’s usually more practical to just write a data cleanup and translation rule set instead.

— Justin


On Aug 23, 2019, at 3:26 PM, Liam McCarty <liam@unumid.org<mailto:liam@unumid.org>> wrote:

Hi all,

Is there work being done to create international standards for identity data? For example, it would clearly be valuable to have standards for common data points like name, address, phone number, etc. If not that, it'd be helpful to at least have standardized mappings between different regional formats.

I've done some preliminary research and discovered groups like the NIEM (National Information Exchange Model, which is U.S.-based) and UPU (Universal Postal Union), but not anything more comprehensive. If international standards already exist, could someone point me in the right direction?

If not, creating international identity data standards seems like a natural extension of the work on DIDs and VCs. Would love to help kickstart that if people would find it useful.

Liam McCarty
Co-Founder of ePluribus<https://epluribus.io/>, Unum ID<https://unumid.org/>