On Friday, October 10, 2025, [email protected] wrote:

> On Fri, Oct 10, 2025 at 12:14:20PM +0200, Christian Barthel wrote:
>
> [...]
>
>> > But only whithin one file, I gather.
>> 
>> No, I don’t think so.  I copied an ID from one item of todo.org
>> into todo_2025.org and this is what the *Messages* buffer says:
>
> Ugh. This sounds potentially expensive.

>
> [...]
>
>> > See, it is complicated [...]
>
>> Yes, I think there is an option that deals with duplicating
>> templates.  See ‘org-clone-subtree-with-time-shift’ where you can
>> specify to delete the ID.  Or it does it automatically... I don’t
>> remember [...]
>
> You conveniently left out the case where you *do want* the
> duplication :)

You mean the "non-unique UUID" case that according to you is a
"philosophical monster"? :)

By the way, the source of the function does something like this:

--8<---------------cut here---------------start------------->8---
;; file: org.el.gz
...
                  (and idprop (if org-clone-delete-id
                                  (org-entry-delete nil "ID")
                                (org-id-get-create t)))
...
--8<---------------cut here---------------end--------------->8---

I think it might be the place where I deduced that org-ids are really
to meant to be unique and started using them.  (but as I said, it
would have been better to use the CUSTOM_ID and specify the
meaning for my particular org-mode setup)

>
> [...]
>
>> Yes - manually copying an UUID results into a non-unique UUID .
>
> ...a non-unique UUID -- a philosophical monster 8-)
>
> I just wanted to draw attention to the fact that UUIDs are not
> magical per se -- sometimes they are useful, sometimes not.

I appreciate your messages and helping me to clarify my thinking.
And I agree with you that UUIDs are sometimes useful, sometimes
not.

> I know it's popular here, but to me, for the sake of "marking"
> positions in documents, they are more hassle than they are worth.
>
> I prefer to locate by content/context.

What are you doing when the content/context changes?  Dangling
pointers where org reports "location not found"?

-- 
Christian Barthel

Reply via email to