this post was submitted on 30 Jun 2024

121 points (98.4% liked)

Programming

16996 readers

100 users here now

Welcome to the main community in programming.dev! Feel free to post anything relating to programming here!

Cross posting is strongly encouraged in the instance. If you feel your post or another person's post makes sense in another community cross post into it.

Hope you enjoy the instance!

Rules

Follow the programming.dev instance rules
Keep content related to programming in some way
If you're posting long videos try to add in some form of tldr for those who don't want to watch videos

Wormhole

Follow the wormhole through a path of communities [email protected]

founded 1 year ago

MODERATORS

[email protected]

121

8 versions of UUID and when to use them (www.ntietz.com)

submitted 2 months ago by [email protected] to c/[email protected]

29 comments fedilink hide all child comments

top 29 comments

sorted by: hot top controversial new old

[–] [email protected] 18 points 2 months ago (1 children)

A lot of people in this thread who don't fully understand how UUIDs work...

[–] [email protected] 5 points 2 months ago

You're not kidding.

[+] [email protected] -15 points 2 months ago (1 children)

There is an IETF standard for UUIDs? Do we need an IETF standard for UUIDs? I've been coding since the '90s and never thought a UUID to be complicated or contentious enough to need a standard. I guess it makes for a pretty unique icebreaker to say you've contributed to an IETF standard, if you get invited to those sort of parties.

[–] [email protected] 23 points 2 months ago (1 children)

If you're generating UUIDs from different languages, libraries, etc. you want to be sure there doing it the same way.

[–] [email protected] 10 points 2 months ago (1 children)

My face, screaming in horror, but in words instead. I've only really worked with projects in homogenous languages on the application side, so hadn't considered that. Thanks for taking the time to reply.

[–] [email protected] 2 points 2 months ago

Look the issues with java.util.UUID and Postgres.

[+] [email protected] -15 points 2 months ago (3 children)

Personally, I always regarded UUID as one of those overcomplicated and frankly unneded "enterprisey" standards (similar to SOAP and XSD, XSLT and various other XML techonologies). After reading this article my opinion didn't change.

Also... do they even know what "version" means? That they choose that word over "type" or any other alternative says it all.

UUID Version 7 (v7) is generated from a timestamp and random data.

Use v7 if you're using the ID in a context where you want to be able to sort. For example, consider using v7 if you are using UUIDs as database keys.

Please, do NOT rely on that and just add to your tables a field with the actual timestamp.

[–] [email protected] 6 points 2 months ago

He's not suggesting to replace timestamps (nor database sequences). They're unique identifiers, and they happen to include a timestamp.

[–] [email protected] 23 points 2 months ago (1 children)

You don't quite understand. One of the major drawbacks of UUIDs over monotonically increasing id's is the lack of ability to sort them. Not just for manual querying, but for index operations, caching, data locality etc.

It's very handy and is a big part of the reason why Twitter developed Snowflake IDs, which are basically like UUIDs v6 and v7.

The UUIDs specs are quite easy to understand and definitely not "enterprisey".

They chose "version" because they are just that, versions. Improvements over the original design that benefit from new insights and technological improvements. We're lucky they had the foresight to include a version number in the spec.

[–] [email protected] 15 points 2 months ago (2 children)

They chose “version” because they are just that, versions. Improvements over the original design that benefit from new insights and technological improvements. We’re lucky they had the foresight to include a version number in the spec.

No they aren't. A higher version of UUID isn't "newer and better", like the word "version" implies. It's just different. It's like they called a car "vehicle version 1" and a motorbike "vehicle version 2". The common use of "version" in the software world would mean that a motorbike is a newer and hopefully improved version of a car, which is not the case.

The talking pumpkin is 100% right that they should have used "type" or "mode" or "scheme" or something instead.

[–] [email protected] 16 points 2 months ago (1 children)

"Version" is definitely used commonly to describe two different ... versions of the same thing, without implying that one is better than the other or supercedes it. There are two versions of the PS5, one with and one without a disk drive. There are many different versions of Windows, like Home or Enterprise. You can get hardcover or paperback versions of many books. Etc. Etc.

[–] [email protected] 1 points 2 months ago

In normal English, when not using a number, sure! But in software, with numbers versions it almost universally means chronological releases of something.

There are many different versions of Windows, like Home or Enterprise. You can get hardcover or paperback versions of many books.

Great examples! Those are both called "editions", not versions. Thanks for proving my point 😄

[–] [email protected] 7 points 2 months ago (1 children)

Version 5 of a software, device, vehicle or such isn't necessarily better than version 4, and no official definition of the word "version" require this, either. If I may make another anology: You may pick one of 5 different versions of an outfit to wear, and even though they were labeled in the order they were made, from 1 to 5, none are inherently, objectively better than any other. In the case of UUIDs there are versions that are meant to supercede others, but also simply alternatives for different use-cases. Anyone with access to some up-to-date information can learn what each version's purpose is.

[–] [email protected] 2 points 2 months ago

Version 5 of a software, device, vehicle or such isn’t necessarily better than version 4

Yep, I can attest to that! I used to play Minesweeper Adventure version. Then Microsoft decided to do a complete rewrite and literally ruined the game. It was way slower and way buggier, and on top of that they also lost all my progress. So, well done Microsoft - now instead of seeing more ads (which was undoubtedly why they did the rewrite) I now don't see ANY ads (because the game is just horrible now and not worth playing anymore, even if it didn't have any ads!).

[–] [email protected] 6 points 2 months ago

Uhum, why not?

[–] [email protected] 11 points 2 months ago (4 children)

Reject UUID embrace ULID.

[–] [email protected] 2 points 2 months ago* (last edited 2 months ago)

I prefer CUID

Just to clarify: Yes, I do know not all use cases are appropriate for CUID. But in general when generating ID, I'd use CUID2

[–] [email protected] 1 points 2 months ago

I vote for nanoid.

[–] [email protected] 12 points 2 months ago (2 children)

At the company I work at we use UUIDv7 but base63 encoded I believe. This gives you fairly short ids (16 chars iirc, it includes lowercase letters) that are also sortable.

[–] [email protected] 1 points 2 months ago (1 children)

base63? I'd guess you'd mean base64?

Anyways, doesn't that fuck with performance?

I'm using this in production: RT.Comb - That still generates GUIDs, but generates them sequential over time. Gives you both the benefits of sequential ids, and also the benefits of sequential keys. I haven't had any issues or collisions with that

[–] [email protected] 2 points 2 months ago

It's Base62 actually, misremembered that. It's to avoid some special characters iirc. And no, performance is fine.

We're using this: https://github.com/TheArchitectDev/Architect.Identities

[–] [email protected] 1 points 2 months ago (1 children)

I'll be borrowing that little trick

[–] [email protected] 2 points 2 months ago

https://github.com/TheArchitectDev/Architect.Identities

Here's the package one of our former developers created. It has some advantages and some drawbacks, but overall it's been quite a treat to work with!

[–] [email protected] 13 points 2 months ago

Interesting 👀 https://github.com/ulid/spec

[–] [email protected] 57 points 2 months ago (2 children)

They should make the versions UUIDs instead of integers so that we don't make assumptions about their ordinal relationships.

[–] [email protected] 3 points 2 months ago (1 children)

Yea, should have been V-00000000-0000-0000-0000-000000000008 instead

[–] [email protected] 2 points 2 months ago

Yes and no. They had to put the version identifier somewhere to avoid sorting problems or parsing problems, so I think that putting somewhat in the middle is a good tradeoff.

[–] [email protected] 4 points 2 months ago

Or maybe an abbreviated hash of the text of their specifications?

[–] [email protected] 15 points 2 months ago

I didn't even know it was an ietf standard. Let aline there were versions. Apparently it's only since may this year that there are 8 versions. Before it were only 5.