On Tue, Aug 26, 2025 at 4:30 AM Chary Ev2geny <[email protected]> wrote:
> On Saturday, August 23, 2025 at 2:38:32 AM UTC+2 [email protected] wrote: > > > TBH I'm really not making the most of the data that's there. > I feel like there's so, so much more that can be done to pull useful data > from it. > > > Can you may be share your idea of what can be pulled, but is not being > done? May be others can develop something? > Well there are many things, making pivot tables tables (e.g., expenses per period), parallel tracking of mirror accounts (like IRA$ allocated being used up), make currency conversions easier, make it possible to combine multiple ledgers into one, etc. all previous ideas we've discussed on the lsit. But I think the area that stands to have the most impact is to come up with a really simple, generic and clean way to convert between tables (e.g., spreadsheets) and Beancount data with as little specification as possible. I do have a few places where I'm doing this, but I think this can be made more generic and commonly used. It provides a natural channel for people who aren't users to participate. But I'm severely time-constrained, trying to grow a small family and on the > work front the deluge of machine learning related input is impossible to > keep up with (+ the time to put in work itself). I prioritize heavily but > Beancount-related improvements rarely make it to the schedule, there's just > too much else to do. > > > It normally gets a bit better with the time availability when kids grow. > But for now you have build a very solid foundation with the quality of > your code and, which seems to be much more rear on the open source hobby > projects, with the quality of the documentation, so, the community can > carry on. > That's nice of you to say, thank you; but consider this: at the current pace of change we'll soon be able to say something like "Hey Gemini, please convert this entire project as close as possible to its original form but translated to Rust" and the output will be near identical but faster and type-checked. We're almost there... this *will* happen in the next ten years I think. So I suspect once humans start to leverage the models a lot more than they are now (most people are still operating on what they know), the nature of software itself will start to shift a bit. I predict a faster pace and more chaos. It's unclear as of yet whether the faster pace will win over the extra chaos (I've learned never to underestimate complexity so I really don't know). -- 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/CAK21%2BhPn1cMei9tADYDFWpf51E54R2AhBfkNcc-vqHqOKWBimg%40mail.gmail.com.
