Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Seeing Like a Spreadsheet (davidoks.blog)
76 points by paulpauper 7 hours ago | hide | past | favorite | 30 comments
 help



This is really two articles in one. First a (great) history of the spreadsheet.

Then it goes into the risk that comes from looking at everything in a company as numbers, attractive now that you have the spreadsheet to manipulate numbers easily, but there are many things that can’t adequately be represented as a spreadsheet without losing valuable information in the process. Finally, AI agents now make it tempting to think of everything in a business as a collection of potentially automatable tasks, which similarly risks losing what makes companies special.


This looks really good. Haven't read in full yet, but I was hoping to see him credit Ben Evans's "Office, messaging and verbs" (2015): "In effect, every person on that floor is a cell in a spreadsheet. The floor is a worksheet and the building is an Excel file, with thousands of cells each containing a single person."

https://www.ben-evans.com/benedictevans/2015/5/21/office-mes...


"Office, messaging and verbs" was excellent, thanks for the recommendation.

https://news.ycombinator.com/item?id=47579407


I think spreadsheets have been severely undervalued by software engineers and they're generally under-researched. It's definitely possible to use them in more non-obvious and interesting ways. E.g., see AmbSheets [1]

[1] https://www.inkandswitch.com/ambsheets/notebook/


You can do this in Excel but it's not all in one cell [0].

[0] https://support.microsoft.com/en-us/office/spilled-range-ope...


A recent blog post from Christoper Drum on running Lotus 1-2-3 on MS-DOS in a 286 emulator https://stonetools.ghost.io/lotus123-dos/ .

I wrote my first AI agent (well a backpropagation model, LOL) in Excel on Mac in 1988. It could only handle several thousand parameters. But it was very cool to see the model in operation.

The rich and complex history of spreadsheets inspired me to build React Spreadsheet. Along the way I deepened (and others) understanding of the complexities and intricacies of spreadsheets https://iddan.github.io/react-spreadsheet

Thanks for that ad-read and self-promotion! Maybe next time you can contribute some insight that doesn't feed your balance (spread)sheet.

Excel (e spreadsheets) is the best quantitative planning piece of software.

There is no other planning tool in the software industry that can answer “what if I changed that” as seamlessly as excel.

Planning is not about its absolute numbers but about its sensitivity to inputs and assumptions.


For better or worse...

A single spreadsheet used locally is probably the best imaginable tool for answering "what if I changed that."

That same sheet shared across an organization suddenly becomes a game of "what caused that change."


Absolutely. And the data and code being stored all in one file makes it exceptionally nimble for the planning phase. You can generally count on any stakeholder in your org being able to handle it.

Can you give an example of what you mean by "planning?"

Budget planning, presumably. How much you are going to spend and on what, and what you need to charge for your products to break even or meet a profit goal.

It doesn't have to be financial. It's anything that can be quantified.

Some random sheets I've used, neither made by me nor about business:

https://docs.google.com/spreadsheets/d/1m08haqvTiXKIh4c7y4uM...

https://docs.google.com/spreadsheets/d/1Fo_-HebVr_9PruE94LgT...


I don't know how true it is today, but many a rollercoaster has been designed/planned in a spreadsheet. g-force and speed analysis, making sure there aren't any "blackout" points, etc. It allows you to iterate quickly and automatically appreciate the ramification of design decisions.

can Numbers by Apple ever catch up ??

This has interesting speculation on the future business impact of AI, extrapolated from Excel:

"This will be genuinely extraordinary for what organizations, particularly the best organizations, can achieve. But if each previous ideology of the corporation illuminated something real about its character and potential, each also, in the fullness of time, deformed it. The financial ideology was blind to what could not be quantified; and the AI ideology, I suspect, will be blind to what cannot be made legible as a workflow."


Hence the title and its hearkening back to seeing like a state - I would guess one of the author's related views is that a rigid, high-modern codification scheme will always miss the magic stuff that fills in the cracks. And you can't go without that without eventual unforeseeable consequences. It's a techne versus metis distinction I think

> And yet you will struggle to find people who love the spreadsheet.

who doesn't love spreadsheets? the average corporate employee holds a death grip on google sheets even if you spend $1m on software that theoretically should keep them out of it.

i've seen countless instances across engineering/data, product, marketing, and recruiting where data is smuggled out of an HRIS/ATS/CRM/ERP to create static structure, improved personal tracking, note-taking, data analysis, realtime team collaboration, etc. all wrapped up in a mini database.


I wrote this, hope people enjoy it!

Thanks I enjoyed it a lot.

I’ve been thinking about how AI will change the way companies are organized. It’s hard to believe that today’s corporation is the ultimate organizational form, there’s just too much stupidity on display.

How will companies compete in the future when they’re all just an AI wrapper?


Maybe you’re talking to the wrong people? Management consultants spontaneously express their love for excel without being prompted. I’ve even seen it at parties.


They are also very good at it. Coders suck at using excel. Honorable mention for the finance folks who know both excel and vba because they deal with both sides.

We suck at excel because we recognize that it has a bad data model and avoid it. So when we want to calculate something we pick something with better structure. something more pleasant to use than the spreadsheets "it's a big bag of cells" approach.

Really, spreadsheets are fine, they probably hit that sweet spot for easy to get something together and deep enough to express complex needs. But I have to admit, now that I have better tools I don't enjoy doing work in them anymore.


I feel it's the extreme of "static vs dynamic languages". In Excel, even variables (cells) are dynamic, not fixed names in a lexical scope.

The reactive programming aspect is genuinely good; I wish my business logic could be expressed declaritively and the system just reacted automatically.

I also find it fascinating to consider the looks-like-a-spreadsheet-but-statically-typed-and-scoped world (airtable is a step in this direction, for example).


I really feel for Dan Bricklin. He should have been richly rewarded for his innovation.

He was, for a time, until others made even better products. It would've been terrible if he had exclusive IP over the idea of "digital spreadsheet".

Was anyone using a spreadsheet to drive automation for testing earlier than 1988?

I have some reason to believe my team was the first within Apple SQA to lean heavily into that, but I’d love to hear of earlier examples.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: