On Friday, February 27, 2026 at 2:02:03 AM UTC-8 [email protected] wrote:

On Wednesday, February 25, 2026 at 1:49:43 PM UTC+10:30 [email protected] 
wrote:

It’s currently not usable for me due to the surrounding ecosystem and 
plugin support


I agree that the lack of plugin support is a problem for most users. Even 
simple plugins like auto_accounts or implicit_prices are probably in pretty 
wide usage. That said, most plugins are actually pretty simple and 
straightforward and if  someone got their alt-beancount to a stable-ish 
point and wanted to facilitate wider adoption it would be pretty easy for 
them to start porting popular plugins, especially now in the AI assisted 
age. And could even make that an explicit thing "open a bug report for a 
plugin you want converted to my alt-beancount and I'll do it for you so you 
don't have to learn rust/lua/zig"

For example, your own effective_date plugin is roughly 300 lines of python. 
For someone who was actually competent in lua they could probably port it 
in an hour or less. It would probably be a weekend's work to port the top 
20-30 plugins. In a different thread someone posted about rustledger which 
seems to have taken that approach and claims to come bundled with 20 
plugins that have been ported to it.


Very much agree: porting plugins is trivial with AI.

IMHO, where it makes a difference is kicking the tires. Taking TurboBean as 
an example, if it came with excellent, fast, Lua plugin support, and *in 
addition* had as a fallback, slower but working support for existing Python 
plugins, the latter would help kick the tires on it, and eventually, 
adoption. The idea of “I can download and run alt-beancount against my 
ledger in 1min with no modifications” is alluring.

I could probably try telling my AI agent to “do what it takes to port over 
my ledger to TurboBean so I can try it out”, but an author telling their AI 
to “allow my alt-beancount to work with existing plugins” provides 
predictability and efficiency.

In the future, there is likely to be many more variants and not just 
ledger, hledger, and beancount, where it makes sense to solve for easy 
inter-operability. Which segues right into Simon Guest’s thread :).
​

-- 
You received this message because you are subscribed to the Google Groups 
"Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/beancount/4ae191a2-1a90-4ab2-98e9-b89c8cfdf62dn%40googlegroups.com.

Reply via email to