Why is an object's ID random by default rather than a string converted from the name?

[ZeroAurora]

I started using Appwrite to build an app recently and noticed that nearly every object id is randomly chosen by default.
By testing on Appwrite Cloud, I understand that orgnization names and project names have chance to collide. But it's not the case for objects like databases, collections and more. These objects are isolated per-project and won't collide with objects in other projects.
I would recommend the sane default id for these objects to be a converted id from the name, i.e. convert Unicode characters to equivalents and whitespaces to hyphens etc.. But please tell me if this is a choice by design. Thank you!

[wellkilo]

Random IDs are a deliberate safe default, but they are not mandatory. For resources whose API accepts an ID, you can supply a stable custom ID yourself instead of ID.unique(). Appwrite's validation permits a-z, A-Z, 0-9, ., -, and _, with a maximum length of 36 and no leading special character.

For example:

const slug = 'customer-orders'

await databases.createCollection(
  databaseId,
  slug,          // custom collection ID
  'Customer orders',
)

I would not make an automatic name-to-ID conversion the platform default, for a few reasons:

• names are mutable display labels, while IDs are commonly used in code/config and should remain stable across renames;
• transliteration/Unicode normalization is locale-dependent and can create collisions;
• two display names can normalize to the same slug;
• predictable IDs may leak business naming and make enumeration easier (permissions still protect data, but random defaults avoid creating that expectation);
• generated resources and concurrent creation need a collision-safe default.

The usual pattern is therefore: use ID.unique() unless a stable semantic ID is useful, and if it is, run an application-owned slug/validation function and handle the conflict explicitly. Do not derive IDs from mutable names unless renames are intentionally decoupled from IDs.

Current Appwrite endpoint descriptions explicitly state “choose a custom ID or generate a random ID with ID.unique(),” and duplicate-ID errors recommend the same split. If this explains the design and the supported alternative, please mark it as accepted.