The end of the accounting search | Hacker News Reader

The end of the accounting search

lwn.net

384 pointsbelter204 comments

I'm a QuickBooks ProAdvisor and have been one for over 20 years and the last few years has convinced me that Intuit is determined to kill the goose that lays the golden eggs by:

1) Forcing all their customers to migrate to the online version of their program. Many of the clients I've spoken to were happy with and preferred the Desktop version. The Desktop version is faster and can do more than the online version.

2) Throwing major updates at the online version that hides or moves functionality. I work with multiple clients and based on the client their version of QuickBooks can be very different from another's. Even the same client that has multiple businesses and has to use multiple subscriptions could have a different look. I don't know what drives this at Intuit but it's confusing for me and I'm supposed to be the expert.

3) Relentlessly raise prices. Intuit has been annually raising their prices in what appears to be an attempt to eventually price everyone out of their product. Some clients have to pay 200.00 a month for the advanced version for one feature they need to have for their business! The attitude from Intuit feels like "So where else are you going to go?"

For most of my adult life I've worked as an accountant that works mainly with QuickBooks but I'm worried that I'm going to have to try to find a new way of supporting myself because Intuit appears to be demonstrating by their recent decisions that they don't need to show loyalty to customers or even keep customers. If they can have a few stay on to pay ridiculous subscription fees that's all they need.

Why would people not go to Gnucash or Fresh books?

They are just not as good as QuickBooks and not many accountants are familiar with these other packages. The prices keep increasing but the QuickBooks online functionality is the best out there and it is the platform with which most other tools make sure to integrate.

Also, none of the open source alternatives support simultaneous multiuser access.

Users can slowly adapt, but most businesses larger than a sole proprietorship need multiuser capability.

This is a big one. I switched from GnuCash to QBO for a non-profit because of that.

One would think it can't be that hard to put together a webapp to freaking sum up some numbers, but nobody seem to want to do it in open source.

It could be because it's indeed so easy that it's boring, or could be that anyone with the necessary knowledge of accounting standards makes way more money doing other stuff than coding for free.

I'd assume that the overlap between capable open source developers and people who enjoy accounting enough to write their own accounting solution is fairly small.

Quickbooks is currently a necessary evil for me. I have to use it, because that is what my bookkeepers and accountants use.

If you want to change either your bookkeeper or your accountant, your new bookkeeper and accountant know how to use it.

Every time I have to open up QB, I cringe. Painful!

Agree with the points above about updates and $$$ increases. The worse!

Gnucash is hard to use.

I recently attempted to help a non profit setup Gnucash. It turned out to be cheaper to get commercial software (Xero) as they were spending too much time on upskilling in Gnucash.

They took a week to onboard to the new system and are amazing at how quickly they can do everything they need to do.

The thing that pisses off the accounting people where I work is how often everything changes.

They’re one step from just using excel because it won’t change on them.

> just using excel because it won’t change on them.

Are your accounting people quite young? Excel moves and changes stuff, though not on quite such an aggressive timeline. Modern Office is not like 1990s Office where you buy it and then it doesn't change unless you decide to change it.

This is fine, the universe's one constant is change, if we fight it we'll lose, but it's worth calling out that there will be change, just some of these outfits are taking the piss.

Many companies have tons of desginers who need to prove they are needed, so they force change for the sake of change - at the cost of users.

It might be like the thing at Google, where people want to get promoted and big changes or new launches are the only way up, regardless of the effect on user experience.

With M$ the bias is towards their certification and training business. They have to change enough to make training seem more important.

That’s just one leg of the stool. They used to be relentless at driving SQL Server Enterprise adoption. Now it’s shifted to M365 E5, plus accessory subscriptions. PowerBI, Viva and Copilot for whatever are the new legs.

> Many companies have tons of desginers who need to prove they are needed, so they force change for the sake of change - at the cost of users.

That may be one part.

But another part is companies that need to demonstrate to their uses that a version change is a big change.

Microsoft is (im)famous for this: as far as I can tell, there has been little need for the constant UI churn in windows aside from convincing users that a new version is big enough to warrant that appellation (as opposed to a service pack/patch).

You’d think that the SaaS Office would change less, but it seems to churn about as much as ever.

Designers alone can't push through a change for the sake of change as their work relies entirely on user research. Even if they do have the research needed to prove a change is required, they also have to get developers on board.

So I don't know where this conspiracy theory comes from, but it's entirely unfounded unless you're talking about a designer like Jony Ive with the power to go through multiple obstacles with ease, in which case the change isn't happening to prove he's needed but because he has a different vision.

> as their work relies entirely on user research

Definitely not the case everywhere.

It's the case everywhere. A designer (in the context of software/hardware) does not work without user research, that is core to their work. Any position that does not require user research in design is titled differently these days.

Unless the designers end up becoming the product managers/owners and stop scrutinizing the designers.

Of course no engineers would ever make bad UX decisions.

They’re probably older, on the contrary, and are still using the non-subscription desktop version instead of Microsoft 365.

Yes, everyone seems to be stuck on 2010.

I'm older. I still haven't gotten over the "ribbon" interface. It drove me to LibreOffice despite all its faults.

Are you a professional accountant?

EDIT: rather than downvoting, maybe you could engage? Following the thread back the question was about accountants using Excel and then someone weighed in that they use Libre/open office. I have never seen an accountant use libre office for accounts work, hence my question. Are you an accountant who moved from Excel because of whatever, or was your comment irrelevant?

I'm sure he's not, I am one and I have to use excel though I also hate the ribbon but I just use excel 2010 + ubitmenu and will continue to use that version or similar until I can't anymore

I had used excel 2003 until I was forced to use win 10 at which point it barely functioned anymore (i.e., it worked great in win 7)

I still miss excel 2003, it has been downhill from there

As for accounting software, they all mostly stink though serve a purpose in larger companies where you need some degree of separation of duties to reduce (at least in appearance as much internal control is theatre) error/theft risk...

I'm not, and I'm sure if I had to use Excel every day I would adapt. But both Windows and Office reached their peak at around 2003 and it's been downhill ever since. I quite liked Windows 2000 Pro as a desktop.

It is already many years ago since that I started to do certain jobs in Libre Office (back then it was still OpenOffice.org) even if I had MS Office. I mean, it was better at certain tasks and costed nothing.

In all fairness, back them I also had to use MS Office for certain other tasks, but luckily I don't need that anymore.

Now there is Confluence instead however to mangle our docs and waste our time.

Excel is distinctly different from the rest of Office for a power user because while you're right that they've reorganized things (though only one real reorganization since at least Office 95, that's pretty good), the keyboard shortcuts have remained the same for literally my entire life. It's why Excel is nonstandard and weird compared to even other Office apps: Excel is the most this-is-for-work tool in Office, and they've been loath to break things for people like...well...me.

And I don't hate the ribbon interface, I think it's totally fine--but also, in Excel specifically, I almost never click a button on it.

QuickBooks's online version, on the other hand, has terrible discoverability and doesn't have standard keyboard controls.

Those keyboard shortcuts (the ones that begin with a slash, at least) date back to Lotus 123.

The Excel UI has changed once significantly in the past 2 decades with the addition of the Ribbon.

The Excel data format has changed once, significantly, in the past 2 decades (which was forced on MS and even though MS has tried to sabotage the openness of the new formats they were required to create, its still much better than the predecessors).

It’s hard to think of any other change that wasn’t just an addition (new formulas, higher column/row limits).

The Excel data format has changed once, significantly, in the past 2 decades

But, importantly, that data format change didn't mean you were forced to stop using the old format.

> Are your accounting people quite young? Excel moves and changes stuff, though not on quite such an aggressive timeline.

Excel UI interface changes but it won't change your spreadsheet content, which is the core functionality Excel users care about.

I've been using 15 year old .xlsx files with no issues across Excel 2007, to 2021, and Excel online.

I've seen companies using files with an absurd amount of worksheets within the file, or with an absurd number of rows (up to the limit), and those haven't been broken.

And I don't remember transitioning from xls to xlsx to be that painful, but this might be untrue for people using some very specific features.

Not at the same speed as SaaS software. Jira seems to change every five minutes.

I use LibreOffice bacause you know what Microsoft.

Are you a professional accountant?

The chrome of Excel changes. The core product is COBOL like in its stability.

The problem with all of these enterprise web applications is that they all suck. I worked for a big government agency that transitioned from a mostly homegrown, mainframe 3270 based accounting system to PeopleSoft. The staff revolted.

We joked about the “Revenge of the Beancounters”, but looking at it, they were right. It took them 4-6 months to train a new person with the old system, but they were hyper-productive. With the new system, they actually needed to grow the staff because you couldn’t build out workflows that the business needed.

They did get some benefits, like improved payments and invoice tracking. (Which drove the business case) But the general financial operations suffered.

this is one of the more annoying things about the Web "apps" and the ecosystem.... virtually no control over what's being used...

I’m a business owner and use Quickbooks online and a couple years ago they just flat out changed how they calculated taxes. When I reported the bug they just kept repeating how I could completely change my process to make it work, like they had released some new feature. But they really just broke shit, and refused to ever acknowledge it. There’s no way I was the only person reporting this. They obviously don’t care

Was this in reference to payroll tax, cause I also got screwed on that from them. I paid for payroll but they didn't file any papers. They paid everything but never ever filed a 940 or 941. By the time government speed came around and the IRS notified me I was delinquent for 2 years. I ended paying 2-3k in fees, dropped there payroll that day. But sadly still suffer through their accounting.

Is that something they can be sued over?

So you are saying Quickbooks is trying to be the next SAP?

(for those who don't know/understand, the "joke" in the industry about SAP is that it's a perfect fit for you, so long as you are willing to change all your processes and workflows to fit SAP. And no, it's not really a joke)

To be fair, I think the 'joke' is that for a regular 7- to 9-figure injection of capital, SAP will do things your way. Until you realize it's easier to just change your whole business top to bottom to do things the SAP Way.

A favorite HN comment of mine from a few years back (https://news.ycombinator.com/item?id=21590482):

> Outside of fundamentally better workflow improvements, most professional fields don't randomly change their tools. If you gave a professional artist a new pencil that had to be gripped differently for no reason, they'd throw it in the trash.

> But in software, we tolerate buggy tools that change all the time for no discernible reason. We tolerate software that simultaneously targets professionals and casual users, serving both segments poorly. We tolerate software that can't be customized or adapted for specific workflows. It's tough to put into words, but if you watch a musician or a painter interact with their tools, there's a very clear difference that emerges, and over time you start to realize how much better all of their stuff is.

> In most professional artistic settings, workflow changes only happen because they have a clear benefit -- drawing from your shoulder instead of your wrist, changing your embouchure if you play an instrument. And even in those fields, it's generally accepted that over time people will end up with very specialized setups that are very consistent and refined and that remain constant for years and years.

> Only in the software industry would someone tell me that my professional tools should change because change is inherently good. Only in commercial software would an elegant, consistent interface like Markdown that allowed me to build up decades of muscle memory until my computer was an extension of my fingers and I didn't need to think about the way I typed -- only in software would that be considered a bad thing.

The converse story is all the people that complain about $enterprise_software saying things like, it hasn't had an update since 1996!

Those are people who don’t use it judging by superficial things.

There is middle ground between doing significant changes all the time and no changes in ~30 years.

I was working at a software company years back that was doing this significant interface change. I suggested we have an 'end goal' interface and breakdown changes into multiple releases over time, organised it so each version change is easier to adapt to and educates somewhat in the changes as they occur, but this approach got no traction. I still wonder if its the correct approach or its better to get it over with in one go and then try to be stable.

At least in my experience the “one and done” feels better, but make it so the new one is somehow faster or easier or something. (Not salesforce lightning, e.g.)

The other ends up feeling like death by a thousand cuts.

That also speaks to the enduring appeal of tools like vi and emacs.

yeah, imagine if a new version of vim changed keybindings.

Torches and pitchforks. Lots and lots of them.

I tend to work with ERP a lot.

I once offered to fix the tab order on an input form, so the cursor started in the right box. The manager of that team got very excited and asked me not to change anything, the team had got used to the keystrokes!

I've found that once a behavior is released in the wild, even a bug, people will build it into their workflows and rely on its presence. Frustrating but unavoidable.

https://xkcd.com/1172/

That XKCD is hilarious! I've never seen it before:

  My control key is hard to reach, so I hold spacebar instead, 
  and I configured Emacs to interpret a rapid temperature rise 
  as "Control."

Good ol' "M-x spacebar-thermometer-mode".

That’s interesting, catering to an employee(a) over catering to the customer.

Point 2 is driving me crazy. Everything moved to places that now seem illogical.

last few years has convinced me that Intuit is determined to kill the goose that lays the golden eggs by:

Is that you, Arlene?

Seriously, everything you list (and more) are things my accountant complains about. You're not alone in being a QuickBooks insider who hates QuickBooks and would love to move their clients onto something else.

It's crazy how slow it is to get anything done with Quickbooks Online. The site is slow to respond, it has very limited keyboard shortcuts, if you open it in multiple tabs to cross reference information, it's constantly logging you out of the other tab, and they discontinued the desktop electron wrapper that let you have more than one tab open.

Historically, a huge strength of Intuit was the huge population of bookkeepers and accountants who would recommend QuickBooks to clients, but I don't see how anyone would recommend it at this point.

From quickbooks to TurboTax, Intuit software has been some of the most god awful both to use and to develop against in my experience.

Argocd is pretty nice

My team’s product is mainly built on Argo Workflows, and while the idea of the tool is great, we have learned to massively distrust it. It doesn’t always do what you tell it to, or, worse, it tells you it’ll do something and then does something else.

Can you elaborate? Just started building out a new platform centered on it

Wait, argocd is owned by intuit? It's under the CNCF, is it not?

but intuit started it (I think, they bought the company behind Argo overall, but Argo CD came out after the acquisition). Being a place for companies to host projects they started but want to share is kinda CNCFs thing after all.

ArgoCD is pretty nice when you use the CLI tools. Try opening the web interface when your k8s setup has, say, 300 nodes, even on a MBP with M2 and 32 GB.

Similarly, Intuit Mint has gotten worse and slower over time as well. I used to use it to track all my spending (including manual entry for cash transactions and much manual categorization of all transactions), but I gave up a few years back because it was all taking so much time and was so slow. Now I just use it to see all my account balances in one place, but otherwise don't interact with transactions much at all.

Back before Mint allowed manual transaction entry (!), I was using it in spite of that. Until the day I logged in and suddenly three months of transactions had disappeared. They didn’t care. I’ve been trying to run away from Intuit ever since. But I’ve been unable to get rid of Quickbook, and now they bought Mailchimp, so…

Ugh.

The advertisements that are constantly trying to push you towards their payment processing and payroll services are also really terrible. There's no way to permanently turn them off.

Genuine question: do you find there is creativity and the satisfaction of creating something from nothing in accounting work? It has always seemed so dull from the outside looking in (cue the joes about accountants) but is this the reality?

Many people don’t see software engineering as creative but it certainly can be!

Creative accounting and tax avoidance do exist tho

My sister is an accountant and she genuinely enjoys working with the numbers. She finds it, not exactly fun, but kind of fun? She describes it like every situation is a puzzle she gets to solve, and then maintain. She also likes working with her customers and enjoys those interactions, like she gets a good feeling from helping her customers out. That's enough for her.

Not the vocation for me but it's very interesting to talk with her about her experiences, because they're so remote from mine...

FWIW I share your sister's take. A large part of my job involves accounting and audit/compliance, and few things stimulate me more than trying to figure out "what the fuck were they thinking" and fixing it.

Sounds like some hardcore deep backend software dev jobs (I have similar for more than a decade), but actual creativity is scaled down to few %.

Accountants are extremely limited by rigid tools they use and laws they have to adgere to. Then comes given company's 'way we do things here' limit.

Take any modern programming language, you start with unique big problem/task, have to crack hundreds little and big problems, often with use of stackoverflow and other webs, but ie hard alghoritmic problems I solve mostly on my own since nobody faced those in exact mix of technologies, data and other constraints. There are sone further limits but my alghoritms are my solid creations, nobody ever questioned them nor made me refactor them for any reason.

To outsiders it may look exactly same weird nerd stuff of course. To me, accounting has very little room for creativity, its there but almost microscopic. And the joy you describe is also present, thats there in many jobs that actually create stuff, digital or not

Note your sister doesn't just work with "numbers". She works with a structure of interrelated accounts. In a sense she is working with the structure of the business. She is observing the business-process which is a bit like looking at a simulation.

Very true and well said!

I'm a small business owner and I get a lot of satisfaction from the feeling of confidence that everything is properly counted down to the last cent. Both for business and for IRS I have a confident answer. Kinda like designing a proper DB schema for the use case, using proper data types, optimal queries etc.

I work around financial software, so I met a lot of accountants over the decades.

There is some creativity in defining where to stash this or that number so that everything works out in the best way for the business. As others said, it's a bit like solving puzzles, and it's harder than it looks - it requires a significant level of preparation, and keeping up with constantly changing regulations.

What a lot of them also get pride from is formatting their reports "just so" - so that they're readable and enjoyable by businesspeople, while still being exact, comprehensive, and meeting formal standards. To me, that sort of thing is kinda boring and trivial, but for them it can be career-defining, life-or-death stuff. In a way, it's a hyper-specialized branch of graphic design.

I think it is the naive idea that bookkeeping is accounting.

I am an accountant, I never do any bookkeeping.

Last week a few things I did were

1) worked on a way to get sales rebates to correctly reflect in the product cost. This is half accounting knowledge and half ERP.

2) presented accounts at a meeting where I was trying to use the numbers to make the business take a different strategic direction.

3) reviewed and questioned the commercial sense in other people's plans

4) worked out how to run a new process through the ERP so the accounts flowed through correctly

In other weeks I may be forecasting cash needs, doing an investment appraisal, building business plans, getting funding, writing sql to get data, building a BI system...

I like the creativity and the amazing variety of what I get to do in accounts.

Accountants are like attorneys. The shitty ones go through the motions and create problems. The great ones often solve problems before you know you have one. They get a bad rap from a stereotype perspective!

It is boring, incredibly so. There's no real creativity involved unless you're trying to stretch the rules capture a transaction in a certain way or make it look like you're doing something you're not really. Its not fun and at some point you become aware of what you're doing and you quickly supress the memory in the deepest part of your mind. Its a hugely usefull skill though that helps you with so many other things, like budgeting, raising finance, running a business, investing. The the profession itself is like being force fed cold porridge for eternity.

Accounting work is in many ways similar to a garbage collector in program runtimes (e.g. JVM). Their primary goal is to manage available resources: keeping track of current usage, anticipate future availability and prevent misuse. In that sense, there can be a lot of creativity in the optimization of that process just as software people have from tweaking a piece of code to work better.

At one point many years ago I did tax preparation. I liked to joke the secret to being a good tax preparer was to be creative. But not so creative that the IRS gets angry.

This sounds like what Adobe and Microsoft are also doing. It seems like they assume that they are somewhat invincible and therefore can do whatever they want.

The sad truth is that they actually are.

Digital feudalism is a real thing.

I canceled my subscription a month ago and moved to wave. What an awful piece of software and company. Slow, ads and promos and every interaction, slow, not very intuitive, did I mention slow? The only reason I signed up was because of a deal I got but it was not worth the gray hairs that I got.

Edit: intuit and quickbooks

Which software and company is awful? It is not clear from your comment.

This is a well-deserved Intuit complain-fest.

Intuit, tax and quickbooks. In addition to quickbooks online not being great, Just an example: I own the gmail for my name, somebody used it by mistake on turbo tax so I get their emails. I called their support and explained it to them but I have to give personal info on the account but I don’t own the account. It was like that for years until recently they fix it, although I already had created a filter for it. Similar happened with mint.

One thing they are going to start pushing via their online platform is likely things line AR and payments via Melio (their a big investor in the company) and likely push out other vendors in that space as a result. My grapevine understanding is this is less possible with the desktop version because you’d need everyone to upgrade.

That’s what I think this is all about

AR? Augmented Reality? Wtf is Intuit gonna do with that?

https://en.m.wikipedia.org/wiki/Accounts_receivable

Most users and companies I've encountered want to remain on the desktop version. Its faster, gives them more functionality, its under their control and if they want they can pay or not pay maintenance fee. But the software firms are obsessed with the cloud and so are the heads of the Finance teams, they think it makes them look technically knowledgeable to move to the cloud so this pincer movement is forcing the everyday users onto something they dont want or need. Any cost savings disappear in rapidly rising yearly subs, but by then the people that forced the move are gone and everyone's lumbered with an expensive, clunky restrictive finance system.

> they think it makes them look technically knowledgeable to move to the cloud

It's not about looks, it's about locking people to the platform and milking them in a subscription model

That's the software company, I was referring to management.

Xero do a similar pricing thing. I have to pay for the $100 a month plan to get multi currencies because I contract to an overseas company. The standard plan which has everything I need but that is 1/3 the price.

Xero's been juicing the lemon more and more aggressively every year...

One huge problem with Quickbooks Desktop is that it's hugely expensive to integrate it with anything, be it cloud-based software, or another piece of on-premise software. This inevitably results in it becoming an "information silo", and necessitates stuff like dual data entry (i.e. entering the same information twice, first in QB and then in some other software).

In contrast, QB Online has a REST API.

the online thing is one aspect, but even the desktop part is hostile. forced obsolescence when you think you are buying a "perpetual" license. disgusting.

i've not found any local solutions that i like, at a cost (perpetual license is a requirement) that i like. so i moved to xero. it's slow and annoying in lots of ways, but it's very usable. some things are less abstract than QB, ie the method of entry for those is closer to just doing a journal entry, i guess this is good or bad depending on the user.

> Even the same client that has multiple businesses and has to use multiple subscriptions could have a different look. I don't know what drives this at Intuit but it's confusing for me and I'm supposed to be the expert.

I guess that, behind the scenes, Quickbooks might run each client on its own code base or some limited multi-tenancy on a shared code base. So if your clients span those code bases, they could easily be on different versions. And the heuristics for when a code base is upgraded may depend on factors that you can’t see.

This is just a hypothetical, but that’s where my money is.

It's inevitable that Intuit will snuff out the desktop version of Quickbooks. The question is, what will rise up in its place? Will it be QBO or does something else have a shot?

I'm curious what the secret sauce is that Intuit/QB has provided advisors, like yourself, to stay on their platform for the past 20 years. What would it take for another accounting software company to get you, and others in your same situation, to commit to their platform for the next 20 years?

> “…I'm worried that I'm going to have to try to find a new way of supporting myself…”

What? I work in accounting and I was reading your post with curiosity until this. How do you support yourself—being a QB Pro advisor or as an accountant?

I am strongly considering a switch to GnuCash, but not for the reasons in TFA.

>Crashes were frequent; QuickBooks users know to make backups every few minutes to avoid losing work.[...] More recently, Intuit has discontinued support for the desktop version entirely in an effort to force all users to its online, paid-by-the-month service.

Both of these statements are not true, based on both my professional and personal use of Quickbooks desktop for over 15 years. QB uses transaction logging[0] so that transactions are not lost, even after for example a power outage. And the desktop versions are still available, but no longer with a perpetual one-time license, but rather $800 annual subscription for the Accountant's version. (This is a doubling in price over just the last few years).

So, I will be able to use my 2022 desktop version into the future at no extra cost, however some of the internet-based services (receiving payments, for example) will no longer be supported after 3 years.

I have installed and started to migrate to GnuCash because I am not going to pay $800/yr to Intuit, especially when less and less of my use is for professional purposes.

[0]https://en.wikipedia.org/wiki/Transaction_log

QB Enterprise uses Sybase internally, which is a sound choice, except they also do some kind of crazy file-locking over a SMB share, which can sometimes go wrong and cause issues. They scramble the login/password (and there are tools to do this computation so you can try to connect to the DB directly, but the schema is convoluted, and clearly decended from their old Dos/Win 3.1 roots), and then connect over the network to Sybase.

Network clients crashing can cause issues that are difficult to unwind. It should not be possible to get the database into a state where backups + verification fail if it was enforcing referential integrity, but that seems to be common.

The official tools to get yourself out of a pickle from Intuit are guesses at best. There is a cottage industry of people who will (for a great fee depending on how urgent it is), attempt to recover/correct the database for you, because they have experience peeking behind the curtain and building additional tools on the XML API.

Good luck if you are using EDI or other electronic document exchange with other customers, and you transmit sales order numbers for a few hundred orders in the morning, and then things go pear shaped and QB refuses to cooperate and open the file in the afternoon. Going back to last night's backup means carefully/manually re-ordering 4 hrs of work to make sure that everything lines up, so you're probably in for a late day.

Intuit "officially" supports a linux based server for QBE, but the community seems to thumb it's nose up at it, and offers no support for this configuration.

QB, for many years was a very good price for the SMB market. There was a LOT of pricing headroom to the next best accounting package. Someone @ intuit has figured that out over the last 10 years and they are ratcheting up the price till customers fall out -- figuring it's finally worth it to drop QB for another solution.

I'm aware of a migration from QB Enterprise to QB Online, that after signing up and paying, Intuit basically aborted the process. "Sorry, we don't support doing multiple tax rates for employees in multiple states in the QB Online version, here's your money back". Maybe they fix that eventually, but that's where it stood last year.

QB is a frustrating product, for the inertia it currently has.

QB Online works pretty well, I've used it for many years.

Intuit has an obvious disrespect for their customers - some important bugs have gone unfixed for years amid continual complaints, while the newbie support people suggest they be reported to Intuit!

I handle the accounting for a small non-profit. We need the ability to categorize our revenue and expense, which now requires the full QBO.

At $850 / year, the cost is out of line with our $30K / year revenue. But, I'm rather stuck because the organization needs to be able to transfer the work to another accountant if something happens to me, and GnuCash, ledger, hledger, etc. aren't well known. Also, we need the ability to accept payments from PayPal and Stripe without a lot of manual entry.

That's where I am. Our 2017 QB desktop was getting to the point where it felt likely to stop working so I switched to GnuCash. However, I am worried about another user learning it and being able to do it right. I do a lot of babysitting of the program as it is.

QB online would be too expensive. Xero might work. The fact that entering bills is an "advanced feature" in most packages gives me pause.

It looks like there is a market for GnuCash support contracts priced at, say, $300 a year. It could offer some consultation and hand-holding in complex cases.

My aunt was an accountant (rip).

She had one of those "not calculators" with a built-in printer. Not sure what they are properly called.

Her "not calculator" was old, stained with cigarette smoke and coffee and she did not wish for a new one.

I got paid a bit a few times to clean it up inside and out. The thing was a tank. (If it still exists, it should still work with a bit of care and attention.)

She used some form of financial software running on a terminal. I think it was from an AS/400.

She was an expert on both.

Whatever key combinations were required to enter data on the terminal had become so ingrained that if you asked her how to do it, she would have to look at what her hands did. (weird).

The UX on the terminal was shit if you had to learn it. (Since I got paid to do some data entry, I felt the pain).

But if you got used to it, it was damn fast. No lag, no mouse pointing, and it stayed entirely consistent. A bit like VI I guess.

As far as I know the system never crashed on her. The server did need care, but it would order it on its own.

At some point they brought in Windows PCz and Lotus Symphony, I think. She hated that.

She found it all slow, inconsistent and did not like the mouse one bit.

The other way around though. When someone new was hired and we were in the age of Microsoft Windows, new people were horrified by the old terminal software.

I still sometimes see people at warehouses or stores that use terminal driven software and some of them are just as fast. It is fun to see.

The majority though hunt, scratch head, peck, hit whatever undo is, and try again.

Should we make easy to use software or software as a tool for those who go through the effort to learn?

I dont think any "windows/osx/linux/" screen UI accounting system would be as fast to use.

Adding machine, perhaps?

Counter-intuitively, making terminal software is harder because you have to think hard about the UX. You have to design how the accountant (or the user) is going to use, you have to make shortcuts, you have to think about the progression, etc... This is not the case for what's happening on GUI-land where the "designers" just randomly put texts, buttons, buttons that look like text, input boxes, tables, etc... It's all there just click a few buttons to get there.

This makes intuitive sense, but isn't an idea I'd thought of before. Are there any resources you'd point to as an intro to terminal/CLI/etc UX (especially any that contrast against GUI instincts)?

GUI software for professional use takes the same amount of care to develop. I've been involved in data entry applications used for call center agents, cashiers and industrial control. A lot of care is taken about tab-order, keyboard shortcuts, proper use of F-keys, readability of everything, color-coding related functions and stability of known workflows. The advantage that GUI brings to the table is that a novice or one-time-user can make progress by just using the mouse as always. But e.g. with tooltips, a mirrored F-key toolbar (think Norton/Total/Midnight Commander), always on hotkey hints, learning a more efficient workflow is made possible.

One of the pieces of software I was most proud about was a data entry system for a P&C insurance company, the old policy subscription terminal-based system allowed for about 200 policies/hour by user, then they "upgraded" to a GUI that slowed them to about 20/hour and I built a desktop application that allowed them to go back to 200 policies/hour but with stuff like auto-filling, geolocation, error correction so the data quality was much higher.

I worked with the end users a lot to make sure the UI matched the workflow, iterated over several versions and it definitely paid off.

> I dont think any "windows/osx/linux/" screen UI accounting system would be as fast to use

Doesn't make sense for such a general assertion, you can fully replicate the keybinds of any terminal system and make all the "screen UI" strictly additive to the keyboard experience

> Whatever key combinations were required to enter data on the terminal had become so ingrained that if you asked her how to do it, she would have to look at what her hands did. (weird).

As an Emacs user, I'm like that too. I don't know which keys I need to type to perform some operations, only the movement. I'm pretty sure a lot of Emacs users are like that too.

I'm the same way with vim. Funnily enough, even though I've been using vim for 15 years and rarely ever have to think about what keys to press, when I am on display (for example in a screen share or pair programming) I'll sometimes forget how to use vim and until I can get myself relaxed a bit I'll fumble around like I have no idea how to use a computer.

Airline workers at the gate use similar terminals, and they're astonishingly fast at using them - rerouting passengers, checking flight openings, etc.

FYI, I recently updated scripts and documentation for migrating from QuickBooks to GNUCash : https://github.com/erikmack/qb-escape

Would love if someone tried to package it into a standalone migration tool, so non-technical folks could make use of it!

Please consider sending patches back if you think they would be of wider interest.

I would love to see one for Xero. The MYOB crowd might also appreciate it.

We have been working on a FOSS desktop based alternative to GNU Cash with a modern UI for the past couple of years: Frappe Books

- https://frappebooks.com

- https://github.com/frappe/books

Do check it out!

Can it do multiple simultaneous users?

Hey, Frappe Books developer here. Not as of now, but it is something that we will be adding in. There has been a lot of requests for multi user support.

This looks pretty good! Hope you guys succeed.

Really neat! This looks like something that could be immensely helpful to me. I have very minor accounting needs but am on the verge of outgrowing the Spreadsheet method.

May I ask how you fund development? What is your feeling on the future?

Thanks! We fund the development via our other product - ERPNext.

We have one full time engineer working on it. If end up using it, please give us feedback and help spread the word.

Is this basically the accounts section of ERPNext? If not, what is the difference? I'm considering ERPNext for a small home business I'm starting up, but accounts are something I've never really delved into much beyond a few attempts at personal budget tracking in the past.

I'm another happy user of GnuCash.

I use the SQLite backend for the same reason Jonathan chose GnuCash: scripting. I can script in bash with the `sqlite3` command-line tool.

I also use the reports extensively.

I personally enter every transaction because I want to know where the money is going and categorize every expense into a budget, but yes, import is important.

As soon as I can, I am going to support GnuCash better than just shilling.

Serious question - why are people stuck on QB when there are seriouly decent alternatives put there, like xero?

My minimum price to go back to QB online would probably be $150k, maybe even more. Here is why: - constant down time - constant errors - clueless offshore support - inflexible reporting - constant UI changes for no reason

Switched to Xero, havent looked back.

QB Desktop was quite good but working offline is impossible these days for any tech startup.

Industry momentum.

If you’re not an accountant, and want someone to take it over for your business, they usually start you on QB.

>why are people stuck on QB when there are seriouly decent alternatives put there, like xero?

When you're an business owners employee you don't have time or authority to change accounting software because you're overworked & thought of as a cost centre.

When you're a business' owner you do whatever your accountant says which is usually a legacy software because a lot of tax preparers & accountants aren't actually in the trenches anymore and don't know about the granular bookkeeping part of accounting and how much better software has gotten in the last decade.

When you're an tax preparers employee you don't get to talk to the client.

When you're the tax preparer director/owner you don't use software anymore.

Source: accounting startup who has doubled revenue 2 years in a row thanks to all of the above

Would you be willing to share which startup you are at?

Gnucash is awesome even if not perfect.

The problem to use it in France is that there is regulatory barriers preventing to use it officially. In theory accounting solutions have to comply with rules that are complicated to follow with your own opensource solution!

The same in Denmark from next year. What is missing in GnuCash is mainly some support for e-invoicing in

[https://en.wikipedia.org/wiki/PEPPOL Peppol]

Is anyone working on something like that?

Is it just me or is that an honest-to-goodness commercial open source support business opportunity?

Maybe there is business for developing the missing features. I would pay for that.

But there is also a requirement for certifying that the data of the software can't be altered, and you are not allowed to certify yourself. This is the bad catch 21 situation. Even if it was possible to ensure that data can't be modified on your computer, something like android protected storage, I guess most people would not like that because if you use open source it is to have control over your machine.

So the only solution is the "cloud"/hosted solution by a provider, but again, if you use open source in the first place, probably it is because you don't want some external people to have control over your data.

Oh, that's weird. I've never heard of anything like that in the US.

In theory, a pretty simple scheme of publicly posted cryptographic hashes would prove to anyone you send the data to (presumably tax authorities? Auditors?) that you haven't altered data. Maybe periodically post a running hash of the list of transactions. Even if that works in theory, I'm guessing it's hell to get it actually certified.

GnuCash may not support concurrent users, but it does support concurrent programmers as its XML config files can be checked into git. From my programmer perspective, revision-controlled accounting makes sense. You can fix bugs and change names, merge and split accounts, all with history, comments and easy reversion. And unlike with h?ledger, you have a nice GUI for managing most day to day tasks.

Does the SQLite format support concurrent users? Or would it need better conflict resolution support in the application?

Checking the XML data into text-based version control is a really interesting idea. you could probably do something similar with the SQLite data, using a tool like Dolt.

SQLite itself doesn’t support concurrent users.

The FAQ disagrees with you: https://www.sqlite.org/faq.html#q5

> But only one process can be making changes to the database at any moment in time

You can open the same SQLite DB from multiple processes and then issue UPDATEs and INSERTs "simultaneously", and SQLite enforces the single-write rule internally with locking - using shared-memory on the same PC or file-locking on NFS-alikes.

Not open source but free with many features, https://www.waveapps.com/ can fit the bill for many small businesses.

However, it may not scale the way QB does, and folks you bring on won't have had experience with it's way of doing things.

That being said, with QB's changes, chances are that's true of QB folks as well (ha, ha).

I've used waveapps for my consulting business and it works great

Related ongoing thread:

Plain Text Accounting - https://news.ycombinator.com/item?id=36022005 - May 2023 (31 comments)

I use beancount for my personal finances and Odoo as well as ERPnext for my company.

ERPnext is my favourite while Odoo currently supports the taxation in my country (Germany) a bit better...

Any tips for getting ERPNext up and running? I just spent a weekend of yak-shaving, and I get to the point where nginx is routing my requests to the “bench start” process, but refuses to serve a page.

I currently just use a Cloud offering, but I also have setting it up on my own server on my ToDo list

May you have more luck than me! If you figure it out, I'd love to read about it. The official documentation... leaves something to be desired.

Frappe Cloud is by far the easiest way to run an ERPNext instance if you're okay with shared hosting. It's a Frappe framework-based app hosting service from the developers of ERPNext and it supports data file imports from other instances of ERPNext:

- Frappe Cloud: https://frappecloud.com

In addition to the Frappe Bench setup process, Frappe provides instructions for self-hosting ERPNext with Docker and Kubernetes:

- Docker: https://github.com/frappe/frappe_docker

- Kubernetes: https://helm.erpnext.com

There's also a third-party Docker image from pipech that's easier to setup, but it's not always up to date:

- Third-party Docker (pipech): https://hub.docker.com/r/pipech/erpnext-docker-debian

Bitnami packages ERPNext in single-tier (AWS, Azure, or GCP) and virtual machine (local computer) offerings:

- Bitnami: https://bitnami.com/stack/erpnext

To anyone reading this, if you only need software for accounting and not an entire enterprise resource planning (ERP) suite, do yourself a favor and follow the advice in TFA to use GnuCash.

Thank you, I’ll check these out.

I resisted Frappe Cloud because I want to own my ERP code and run it on my own server.

I recently started using Beancount, and liked the concept of plain text accounting so much, I want an ERP with the same spirit.

Edit: I'm building this at plaintexterp.com.

I'm a huge fan of Manager.io [1], which is free for single-user desktop use. They make money from their cloud plan and from a self-hosted server licence.

It's not open source, but I can export practically everything as CSV or TSV, so there's data portability.

[1] https://www.manager.io/

Xero isn’t free, but it’s a good alternative to QuickBooks if you need one. It has a neat feature called Hubdoc, which automatically scans, renames, sorts to folders and archives your invoices when you email them to a particular email address.

The integration between Hubdoc and Office365 is clunky because for reasons I can’t fathom, they only support Dropbox and Google Drive for archiving, but we have a script that we run on our Synology NAS that copies the invoices from Google Drive to SharePoint (OneDrive).

Nothing against GnuCash or open source, we use a lot of FOSS. But we’re in Australia and we switch banks all the time. Those compounding factors would mean that we’d spend a lot of time updating bank roll integration with GnuCash

It’s transformed businesses and the business of accounting down under. Accountants are focused on helping the companies and tougher questions rather than bookkeeping. Most accounting work is automated and companies have constant control and understanding of their numbers. Not always - but far more often.

It’s astonishing how slow the transition to online accounting is in the USA.

I'm somewhat surprised that GNUCash was picked over Ledger or another of the plaintext systems out there:

https://plaintextaccounting.org

Maybe this was primarily done based on the users being comfortable with a GUI driven system?

At the other end, there are a number of accounting systems that are based on plain-text files and an associated set of programs. These systems are much easier to work with [than the non-GnuCash GUI tools evaluated] and are adaptable to most needs. Such a system could have been made to work for LWN, but they tend to lack some useful features. Check writing and report generation, for example, tend not to be well supported. The user interface for many operations is a text editor; this is often seen as an advantage, but it becomes less so as the number of transactions grows. There are advantages to a graphical interface designed for the job at hand.

> Maybe this was primarily done based on the users being comfortable with a GUI driven system?

FWIW I started prototyping a plaintext accounting 'app' that combines a popular plaintext accounting system (beancount) with the most used web frontend for it (fava) - it's called beancolage and makes use of eclipse theia (vscode-like experience): https://github.com/seltzered/beancolage

My observation is the plaintext accounting ecosystem is wonderful but a bit diffuse to so many specific usecases held by a more technical userbase. My aim for prototyping beancolage was to ask how things could look if there was a more common 'download and try' experience accessible to more people for basic workflows. Also have a talk on the motivations here: https://youtu.be/mxgoFqmSCFc

I used Ledger to manage the accounting for a small nonprofit. It worked for me, but I was the Treasurer and the only person managing the books. Multiuser would have required git or something similar to handle the updates and possible conflicts. That's all stuff I'm comfortable with as a developer but a complete non-starter for anyone else. These issues would quickly make text-file accounting infeasible for larger organizations.

When I moved out of that role, the next Treasurer wanted to use Quickbooks so I had to help import the data. That's another problem with using obscure software -- it's more difficult to hand it off to someone else who only knows the mainstream systems.

The main issue towards adoption is the ease of fetching the transactions from various banks. Sure we can write some scripts to process the PDF files etc but it is still very manual compared to, e.g., allowing transaction query via REST api by granting users API keys.

I'd take an API which dumps CSV files. It doesn't help every bank times out cookies in less than one day.

CSV, JSON, XML - any plain text format accessible via API would be better than some binary blob you need to download manually.

look at https://gitlab.com/gnukhata/gkapp/-/issues

this is a AGPL V3.0 licensed accounting software aimed at small businesses and schools and professionals.

Its currently being developed for indian market because devs are all based out of india so if anyone wants to implement for their own country, please submit a PR.

we need more accounting softwares that are free and open source

I loved GnuCash until my database was corrupted and i had no idea how to get it back into working order. I stopped tracking my finances closely after that

GnuCash keeps a year of backups (e.g. old saves) by default. It's probably still there.

My CPA wife sets up and recommends Quickbooks Desktop installed on Right Networks, a Windows Remote Desktop shop where Quickbooks and Excel is installed. The convenience of remote work without the awful Quickbooks Online. It handles backups and you do file transfers and local printing. It doesn't stop the Quickbooks subscription pricing but you get to use Quickbooks Desktop.

As a linux user and free software advocate, I die inside at your CPA wife's recommendation, but I upvote it for it's pragmatism and applicability to so many people who don't have time or desire to screw around with software. Thank you to her and you for sharing this advice.

> Scheme looks like Lisp but isn't really Lisp

Well that’s a surprising statement, i’ve heard it touted as more lispy than common lisp.

Scheme is definitely a Lisp. One of its primary innovations over previous dialects of Lisp was lexical scoping, and most later Lisps adopted that, including Common Lisp.

A lot of people use the name "Lisp" to mean specifically "Common Lisp".

> Scheme looks like Lisp but isn't really Lisp

The Lisp world is extremely confusing.

I'd wager that at this time, most people in that space can't even agree on what "being a LISP" means.

Haven't agreed in half a century, won't agree now, never will.

Javascript is lisp but clojure ain’t so!

I've never seen anyone say that MacCarthy's Lisp 1 and 1.5 are not Lisp. Or that MacLisp, or InterlLisp, BBN Lisp and other historic dialects are not Lisp. Or that Emacs Lisp or ANSI Common Lisp aren't.

The entire disagreement is about languages that deviate from the core design patterns set out in the ancestral Lisps, some of whose adherents want them to be regarded as Lisp.

This article has some real problems. Here is one.

> Such systems are unwieldy to work with and difficult to get started with. In truth, many of them appear to have been developed as platforms for consulting businesses, though surely nobody would deliberately make their software complex and tricky just to motivate clients to purchase their services.

Or perhaps some businesses have complications to deal with that go beyond what a newsletter website needs? Some businesses have inventory, subsidiaries, multiple currencies, have to report sales tax in multiple countries with ever changing requirements (hence the support and updates that this writer doesn't think are required). Some companies have local statutory books that are different than their group books.

Really it is a bit like learning 'echo "hello"' and calling oneself a programmer. Then maybe one might ask why are lwn making Linux so complicated...

> Or perhaps some businesses have complications to deal with that go beyond what a newsletter website needs?

So ... perhaps there should be a version of Quickbooks that serves the simpler needs of smaller customers better?

Like: Simple. Desktop. Requires no upgrades as long as its functionality at-time-of-purchase continues to serve well. No forcing of upgrades via dropping of old formats, etc.

I also have given up on Quickbooks.

Imagine you were relying on Quickbooks, and a major taxation change, like the European VAT OneStopShop came out. How would you deal with that without updates? What about the UKs Making Tax Digital, that would be tricky too.

That wasn't the point I was discussing, but there is an answer anyway.

Certainly charging for upgrades with useful new and critical functionality is reasonable.

But reasonableness would also price even critical functionality relative to how hard it was to actually implement.

I.e. don't hold hostage a customers' investments in a software product. Don't force customers to pay a ransom to get a simple-to-implement feature.

I think it's a legitimate complaint that packages like LedgerSMB are meant for larger businesses, with internal accounting departments, and do not have a default that works better for small businesses. Quickbooks has the concept of deleting a transaction and it resets the dependencies of that transaction. GnuCash does that pretty well, too. LedgerSMB expects a reversing transaction, so you are not really able to make mistakes the way a small business owner might. You have to start out at a higher level of accounting know-how. Putting LedgerSMB together with an external accountant/tax preparer/bookkeeper using it is a no-go. They already roll their eyes at Zoho Books and I imagine they didn't like Xero until enough people used it. Asking them to use your own LedgerSMB install, they'll just nope out.

GnuCash is fine for the solo but I don't think the external people will like it. I think LWN will do what it takes to make the reports that the accountant will need, but as it stands it is not ready for that out of the box in my opinion.

Other packages, like Odoo, treat double-entry accounting as a trade secret and so it's not really open-source. There are add-ons but their level of maturity is not clear.

If I tried to use ledger (plain-text accounting) I think even my business consultant would tell me to take a hike. I used it when I did a little contract work and it was fine, but my needs were very simple.

Still, getting off of QB and into Zoho for my business has been great. Everything can be exported if I need (to CSV files) and it handles almost all of my needs. But it's not open source.

More personal finance so only tangentially related but I thought Microsoft Money was as great program years and years ago.

I wonder why MS killed it.

Every now and again I look around for something similar.

The Sunset edition of Microsoft Money works on Windows 11. It was perfect for personal finance. Didn't make you learn double-entry bookkeeping like GnuCash.

I really, really enjoy Beancount. I use it for personal finances, and also for the accounting of an association I'm a part of.

The combination of a simple text format (that I version), extensible CLI-based querying and reporting tools, and free software makes it absolutely perfect to me.

> Getting our QuickBooks data into GnuCash took a bit of scripting using the capable (if poorly documented) Python bindings. Another set of scripts has since been written to import data from the site and from our bank account. GnuCash has a reasonably capable import mechanism for the data formats exported by banks, complete with a Bayesian system to assign transactions to accounts, but a bit of scripting makes the process quicker yet.

I assume the person who posted this to HN is not the article's author, but if the author happens to be reading, it would be very helpful to see these scripts. The difficulty of writing custom scripts for GnuCash data is one of its main limitations IMO.

Some of the scripts are available here: git://git.lwn.net/qb-escape.git

But yes I heartily agree, those scripts would be immensely helpful even just as an example/reference for somebody writing their own.

This is good for GnuCash. It could bring enough visibility and interest to bring in more skilled developers to contribute a little work into adding or polishing some of the features the article mentions are lacking.

Because someone is going to pay them?

Because LWN adopted it, raising visibility and interest. I edited my comment to make it sound less like I was saying GnuCash is buggy.

It's not true that Intuit discontinued support for the desktop version. Where did they get that idea? It's a subscription, yes, but the desktop version is alive and well. I WISH they would kill it so I could force my own users to upgrade to the online edition but no, it hasn't happened yet. We just upgraded to QuickBooks Desktop 2023.

As a business owner and software developer I'm extremely reluctant to host mission critical and sensitive data with a provider like intuit. This is why we use desktop QB.

I got the idea from the emails Intuit kept sending me saying that the desktop version was done and I'd have to buy a subscription from now on.

You do have to buy a subscription, but the desktop version is not done. There is a subscription for the desktop edition and separately one for the online edition. They are still perfectly happy for you to use the desktop edition.

https://quickbooks.intuit.com/desktop/ -- it's the "Work on desktop" options.

Annoying that the UK tax authority (HMRC) has made it impossible for open source projects to access their APIs: https://wiki.gnucash.org/wiki/GB/HMRC#MTD_APIs

Any HN users out there with experience with Firefly III (https://github.com/firefly-iii/firefly-iii)? Sounds interesting.

I love beancount and have used it for a small business and have used wave accounting (made by some friends in Canada). That said Quickbooks is our main system

https://beancount.github.io/docs/index.html

I have been using it for some 15 years now. GNUCash is great.

"Getting our QuickBooks data into GnuCash took a bit of scripting using the capable (if poorly documented) Python bindings."

I wonder if LLMs can help with these gaps in documentation and examples?

I wonder if LLMs can help with these gaps in documentation and examples?

I wonder if LLMs can do someone's taxes.

But I'm not going to be the first to try it.

I was trying to understand what box 2b was on one of my tax forms. Google wasn't helping me because I all I found was unreadable blog spam. So, I tried ChatGPT. "On form blah, what goes in box 2b?"

It descended into me arguing with the machine over whether or not 2b even existed. "GPT, I'm literally holding the physical paper in my hand and staring at the box labeled 2b!" "That is incorrect. Form blah has never had a box labeled 2b"

So, I'm going to go with: probably wanna hold off on the LLM taxes.

So, 2b, or not 2b? That was the question?

This seems weird. Did you try rephrasing the question?

Can I see the transcript?

>It descended into me arguing with the machine over whether or not 2b even existed.

Sounds like ChatGPT can already obsolete the tax office.

LLMs definitely help with examples. I use them for that all the time - for any library that had reasonable usage prior to September 2021 they often give me the exact example code I need to understand how to solve my problem.

Reply on news.ycombinator.com