Plain Text Accounting: An Opinionated View
ฝัง
- เผยแพร่เมื่อ 2 ก.ย. 2022
- Today we dive into the world of "ledger-like" plain text bookkeeping tools, such as ledger, hledger, and beancount. Pete shares his view on the category of tools. We talk about what they're good for, what they're not good for, and ruminate on the philosophical underpinnings beneath them.
- เกม
Great video!
I liked the balanced discussion in terms of pros/cons on each side.
There is a reason the industry has standardized on certain tools. They are the easiest to use for the largest amount of people.
Communication and collaboration amongst people *is* the magic.
CLI tools are not what most can work with these days (said by a programmer who recently dove deep into tmux, vim, CLI tools and LOVE them for personal use).
I think the whole philosophy behind plain text is really 'portability'. Meaning ease of taking your data with you someday when you decide to move to a different toolset.
I have recently landed on markdown files for all my notes, journaling, and analysis because of this portability. But even with markdown, there are formatting and syntactic differences between tools.
It is crucial to be able to easily consume the data into the next application. Rather than merely being able to export from the first one.
Great video, but I wish you had went into the benefits of PTA a little more. I don't think anyone could watch this video and come out wanting to try it, which is a disservice. For me, the four main benefits are privacy, price, control, and speed. I don't own a business and my investments are all in retirement accounts, so my accounting is very straightforward. With PTA, I can take the CSVs from my banking websites, pump them through an importer I wrote, bang out some descriptions rapidly in the IDE of my choice (with autocomplete), and version control with git. There's no such thing as not being able to get a report exactly the way I want it, or trusting my passwords to some third party. All free! And while I agree that the compile-error-debug cycle is delayed, I'd add that the errors beancount throws are understandable, unlike that awful GnuCash gibberish you highlighted.
Different strokes for different folks!
I think choosing any software package is about balancing tradeoffs, and some of your own criteria have tradeoffs with each other. I think perhaps the difference between you and I is we disagree specifically on the "speed" part of the equation, and I have an additional requirement which is minimizing input mistakes. For me, that's probably one of the most important. There's no question that I was and am faster using GnuCash than Beancount, particularly once stock transactions got in the mix - I found it trivial to get beancount into a state where it was making nonsensical inferences about how many lots I had, and lost literally hours trying to debug that. And if speed is truly a priority, there's just no comparison with any product that supports bank feeds. In terms of both speed and accuracy, having bank feeds in a product represents an order of magnitude difference between manual entry, or even importing CSVs. I'm willing to trade off price for that, and I'm also willing to trust a reputable partner to not steal or mishandle my bank feed credentials. If you're not, I totally understand that - but even in that situation, I'd still generally hesitate before unleashing PTA on someone who wasn't interested in doing their own scripting.
All of this, of course, is just my opinion.
@@TeaLeavesProgramming Thanks for the thoughtful reply! To be honest I can't really argue against any of that, except to add that I'd suggest hledger to someone who wasn't comfortable scripting. (But if you go with hledger, you lose fava.)
Just starting my journey with hledger. My current hold up is how to get nice descriptions from imported bank CSVs. How do you accomplish the automated importing with nice descriptions?
I used GNU Cash for a good year and it was great, but once you fall off the waggon it hard to get back into keeping your records accurate.
please do more opinionated videos! I couldn't care less about accounting, yet I found value in your logic.
Thanks for your support!
Just stumbled along this video and it was upmost delightful!
thanks for taking the time to create this video
Incredible video, thanks for this amazing content.
Thanks for your kind words!
Thanks a lot for this video!
The appeal to me is how it deals with multiple currencies.
If there is other software with a nice GUI that nicely supports multiple currencies I would like to know about it.
What you are calling a "Philosophical disagreement" can be solved by a simple syntax highlighter with an editor like vscode. Almost all the programming languages are written in plain text, your editor needs to know how to highlight syntax errors. There is no need to compile the code and find your errors after. That does not happen anymore 😉
I don't disagree, and as a software developer I'm certainly used to the idea of using tools to make up for the deficiencies of existing tools. But the fact that I can do that doesn't mean the deficiencies don't exist!
The philosophical question is: If in order to use PTA effectively you *require* the use of special editors, syntax highlighting, etc...then what's the actual point? You're basically just quibbling over file format, which is invisible 99% of the time.
@@TeaLeavesProgramming my point is that the editors are not special. They are as ubiquitous as the Plain text itself.
On the contrary, Microsoft xl which is the first choice of amateur CA has become more exclusive.
In this day and age when accountants are learning python. Its really hard to see need for an editor as a deficiency.
@@jsksingh88 We simply disagree. I think if you asked 100 accountants if they would give up Excel or Python, at least 99 of them would give up Python. Possibly 100. And, again, if you're using Python, you can store your data in an actual database, which provides fast random access, access control, better validation, and is, from a programmatic standpoint, EASIER TO PARSE than plain text. The value of proposition of plain text as a journal format is just...weird.
@@TeaLeavesProgramming its not python vs excel !! The comparison we’ve been discussing so far is plain text vs proprietary accounting tools including excel.
I think, plain text as a format is very powerful!! It takes away power from the vendors in favor of the users.
Markdown has done the same thing to expensive word processors and HTML editors.
Good discussion though !
So, what do you use and why? Also, what do you recommend for common scenarios, that is, levels of complexity, with or without business/non-personal accounts, simple/complex assets, with or without international complications, …?
Not counting software that I use just for purposes of review - I'm always trying things for the channel and just to keep up to date - for several years I found GnuCash to be the right balance between complexity and flexibility - and the price, $0, was of course right. One benefit of GnuCash for personal books is that it's purpose-built for tracking stock investments, which is something a lot of us do. It has AR and AP capabilities, but they're a bit primitive.
As my business activities have grown, I've needed to do more invoicing, and the lack of bank feeds in GnuCash became more problematic. I tried all of the "big three" low-end cloud services (Quickbooks Online, Xero, and Sage Business Accounting), found them all to meet my needs pretty well, and chose one of them for myself (they're all roughly equivalent, frankly, so I don't want to get more specific than that in a TH-cam comment.) One drawback of this is that since those are really business accounting packages, they're not really equipped to handle stock trransactions. You can do so (and I do!) but you have to do some calisthenics to make it work.
@@TeaLeavesProgramming Very interesting. I had used GnuCash 10+ years ago. Maybe I should try it again. The thing I most like about plain text is the ability to work effectively without graphics, as I basically live in emacs and the terminal. I probably need to get over that and broaden my horizons. Also, I didn’t know about the cloud services option; good to know. Thanks for the video, and follow up info!
Wait a second you started the videos saying "PTA doesn't support my complex financial system so it is not suitable for complex systems" yet all I heard was "It doesn't correct typos or malformed text"....
Genuine question: show me something you can't do something in PTA that you can do in some other accounting software...
If the only limitation is, "You should know how to write PTA language", that's poor argument. I would like to see an actual real world example of a transaction that you cannot do in PTA. I'm asking this because, I'm new to PTA and I would not like to invest in something that has some serious limitation.
The standard isn't "can or can't do" but "is it better or worse"? We literally used hand-written paper ledgers, which are plain text, for hundreds of years to do ALL accounting. Nobody would go back to those today. That's not because the accounting is IMPOSSIBLE, but because it's harder and less convenient.
I'd specifically call out accounts payable/receivable tracking, reconciliation, approval workflows (and limited permissions generally - allowing one person to have only read access, another person to have only write (but not read) access, and so on), and the creation and maintenance of audit trails as activities that are all harder to do well in the existing popular plain text accounting tools (ledger, hledger, beancount) than something like QuickBooks, Xero, or Manager.io.
your intro says counting like it's "1979"
Thanks for the correction!
We ARE more beautiful then average. Thank you for noticing.
How could I not? It's so evident! :-D
Personally, I haven't been able to figure out why would anybody punish themselves with this approach if there's GNU Cash available around. Save your stuff to SQLite format and create your own reports with SQL (if plethora of built-ins are not enough), or just manage huge loads of data with it. CLIs, CSVs, XMLs and others? Right there. Why bother with PTA at all? One may only wonder.
For the same reason that you like GNU Cash: familiarity.
I understand that raw text files and CLI tools feel daunting and/or unnecessarily complicated to most people, but as a software developer using a powerful text editor with completion and syntax highlighting to write transactions with a code-like syntax feels like home. I like it. On the other hand, waving my mouse around or tabbing like a maniac to get around GNUCash's dated interface feels like torture.
@@ohnoes9608 - again, GNC gives you both options - a CLI and GUI. Plus SQL / XML as a database - as a dev you know how powerful they can be. Be objective.
Personally, I've tried both approaches, but with 20k+ transactions, PTA-only is much slower for me.
OR you could do what I did and spend all of your time deciding between plain-text accounting, software-based or spreadsheets... and no time doing any actual bookkeeping or accounting.
p.s. Not recommended.
At least you avoided the next level of the trap, which is writing your own accounting software from scratch.
@@TeaLeavesProgramming Thank goodness I'm not a programmer.
That is to say that I'd have that many more unfinished projects laying around! 😁
@@TeaLeavesProgramming i have to resist...☠
TOO MUCH NUANCE 😂
I know, right?