For a crisis-response, an approach may reach maximum impact within days and really only be needed for a short duration (a week to a month). Migrating a system with intense user/volunteer churn (and keeping two databases synced as you transition) can be tricky. Handling permissioning at the same time is also a challenge.

At FindTheMasks, we expected the government to have supply chains and epidemiological response (testing/contact tracing) together within two or three weeks. When that didn't happen, we kept bolting on new tools and strategies to double the lifetime of the endeavor, expecting the government to get its poop in a group by then. 20 weeks later, supply chains are somewhat improved, though the entire situation is compromised by the lack of overwhelming success at the epidemiological scale.

It's hard to go from Google Sheets to a web frontend to Postgresql and a snappy shared frontend. That is the problem statement to solve for imho (interface, schema, data portability).

As you tack on extra tabs and handle more fields, it isn't too hard to approach that limit.

If we do trip that threshold, we will be forced to degrade some functionality.

As a bonus, I don't believe that Sheets allows anyone to see exactly how many cells are in use, so one lives with a probabilistic risk of lockup.

If we were continuing to handle large volumes of intake, a switch to Airtable would be in the works. Fortunately, the PPE situation is at least somewhat better than it was in March, so we are simply keeping the lights on and wheels turning.

Airtable-wise, I'm parroting our devs on the subject. We had perpetual/nascent plans for an Airtable migration (and did so when FindTheMakers came to life) which have been tabled for now. We believe we'll make it to sunset on Sheets, but just barely.

Is there some special meaning of "record" in their vernacular?

I haven't used either before but cryptpad seems solid. Here's a demo sheet https://cryptpad.fr/sheet/#/2/sheet/edit/DirOb8rZr+6gfi5PCrS...

I sat there thinking "I can't believe there isn't a better option for this sort of thing." Kind of surprised Zoom does not have a native feature for meeting participants to share contact info with each other during the meeting.

A custom web app for this purpose can scale better by being designed with the knowledge each user only edits their own row and nothing else needs to be recalculated. And of course it can have a tailored UI.

Unfortunately I don't know of any way for a non-developer to match that. There's no fundamental reason IMHO that there couldn't be a nice rapid, code-optional CRUD app generator, like Microsoft Access for the 21st century. Maybe this is what Microsoft Lists will do? AirTable is a general-purpose web-based RDBMS frontend, but I don't know if it scales well to many simultaneous editors either, I don't think you can tailor the UI as much as I'd like, and I hear folks complaining about database size limits.

You can get Google Sheets to scale better, but it's a bit awkward:

* Give folks a Google Forms link to edit their own information. This atomically claims and populates the row.

* Give folks a publish link (see File | Publish To the Web) to view everyone else's information. This link scales well because it doesn't let them update anything and (more annoyingly) it doesn't show changes until they hit refresh.

That's a big number, but it's not as big as I'd expect from something from Google. I'd assume the average number of people working on any Sheet is 1 or 2, but I'm surprised that haven't been able to scale it up to 1000 or more by now.

The technical limitation is 2-4 times what the former product manager believed was a human limitation already. Scaling beyond that doesn't seem like a priority.

Google Docs/Sheets/Drive function a lot like Stardew Valley multiplayer, in that pretty much anyone can do anything, and this allows really fluid communication but it simply won't work with 100 people.

Sheets supports 100s of people to an extent (even more) but I wouldn't want them to optimize for this. I think if I were in charge of it I would be exploring ways that they can find UI paradigms that allow more seamless interaction and realize the dreams that I think Google had when they built Wave.

* What if you could embed live video/audio in a Doc?

* Sheets comments kind of suck.

* Is there a way to use comments that makes them more visible without ruining the sheet view? (Maybe not.)

* As this article says, people are often using a Sheet as a very hacky app dev platform. Is there a way to embed the sheet data in a Doc (or something Wave-like) that has a more seamless communication pattern?

* Is there a way to build small apps on the fly without messing with a bunch of macros?

Though I think a lot of this is about communication. I don't think Google should be thinking about scaling, they should be thinking about those use cases where you're talking to a small group of people and you need a little custom app that can't take more than 30 seconds to put together or just talking would be more efficient.

That being said, this article seems to completely ignore the fact that there are already well-established solutions to this problem. Off the top of my head, wikis and GitHub/Lab pages offer very nice solutions to creating a scale-able information-base with built-in features to help address some of the challenges I raised above....

It's a flippant answer, sure, but spreadsheets are the index cards, sticky notes, and yarn approach to data organization. Certainly they are easy and approachable to use, but any sufficiently complex data set in general is going to be more "chaotic mind corkboard mess" than "organized".

Spreadsheets are very useful data hammer, but not all data is a nail; to borrow another aphorism.

That said, there's a wide gulf in the general user approachability between spreadsheet applications and more bespoke/"better" data organization options. Decades of experiences and mistakes learned from tools like DB2, FoxPro, and Access have shown us we have a hard time finding an "easy medium". AirTable thinks they have some modern ideas. Microsoft's PowerApps teams have their own modern approaches (and as almost the direct the inheritors of FoxPro/Access, presumably the most experience to learn from). Simon Willison's work with Datasette (SQLite DBs wrapped in simple view/management webpages) has had some interesting ideas. Though I don't think any one approach will solve it, it might take a whole shelf of tools, and at the end of the day maybe as a species we'll never have as clean of a user experience as the worst spreadsheet.

https://www.theverge.com/2020/8/6/21355674/human-genes-renam...

edit: HN link

https://news.ycombinator.com/item?id=24070385

I wonder if that's why the new version of macOS Numbers defaults to a read-only mode — to ease the burden when the sheets are in collaboration.

I would love to see Elixir based solution just to see how much performance can be squeezed out of it, whatsapp seems to handle huge load just fine

Archve.is link: https://archive.is/N00i5

This got me thinking what would it take to build a decent collaborative editor that would work for hundreds of people.

The low hanging fruit would be adding a read-only view of a static copy, because that's what people actually needed.