paulfitz

(disclosure: Grist employee) You're exactly right https://github.com/gristlabs/grist-core/blob/7f1f8fc9e60a2e1... - sorry the import choked!

Open core is definitely a tension (disclosure: I work at Grist). Specifically for Grist, we differentiate by offering data portability and autonomy, and a full-featured open-source app is part of how we do that. I invite you to check the repo out, you'll see we're not holding back. We did want to flag the project as open core from the start, since we will build proprietary services and features around the core app (which again is very featureful!), and don't want that to come as a surprise down the line.

Yes it does, changes are broadcast to all connected clients and applied in place. Disclosure: I work on Grist.

https://undici.nodejs.org/ "Undici means eleven in Italian. 1.1 -> 11 -> Eleven -> Undici. It is also a Stranger Things reference."

For google-trends it translated "Is there anything cat-related that people are excited about?" to something very reasonable that worked (answer: popcat apparently). Nice!

Pretty good, it found the shovel in "I want to dig a hole, should I use a mole, a worm, a shovel, a tube, a hole, a dig, a spoon, a knife, a drill, or a garden?"

Grist is an sqlite-based spreadsheet format that captures relations well and discourages funky structure https://www.getgrist.com/blog - it works well for passing relational datasets around (disclosure, I work on it)

I hear what you are saying. Spreadsheet/database hybrid vigor :)

One simple thing you can do is read a table as a data frame with:

  pandas.read_json("http://localhost:8484/api/docs/<document-id>/tables/<table-name>/data")

When running locally, you can use whatever you like, pandas or anything else, which is fun (we've experimented with some options for Jupyter integration ourselves). In our hosted service, for security we have a locked down python sandbox with white-listed libraries. Definitely on our roadmap to liberate the full power of python libraries in hosted!

Hey thanks! We're excited too.

Yup, and in the meantime there's a docker image https://github.com/gristlabs/grist-core#opening-and-editing-... (hi Cyprien! waves in French)

Ah, we do have a general purpose importer from sqlite lurking but not published, I'll bump its priority. In the meantime, you can use the hosted or local rest api from python or javascript to generate a grist spreadsheet https://support.getgrist.com/rest-api/. On the backend, for serving individual user spreadsheets, we use node/express workers connecting to a pool of sqlite files, persisted via the sqlite backup api https://sqlite.org/backup.html https://github.com/mapbox/node-sqlite3/pull/1116. If you want to dig in, you could poke around at https://github.com/gristlabs/grist-core

Yes, a new table in your spreadsheet is stored as a new sqlite table, plus some metadata about it in some other administrative tables. You get all of this as a standalone, round-trippable file when you download your spreadsheet.

It was a combination of the value to others we could liberate by doing that, and the value it would bring to us. As Grist took shape, it felt like a nice self-contained tool (and format) that could be useful at many points in a data pipeline, given its editing, visualization, and summarization functions. And spreadsheets have a long tail of features users want, so it makes sense to open it up to collaborative development, especially given how tech-savvy our users can be.

For formulas: could it be that Grist formulas are per-column rather than per-cell? This can take a little getting used to, but we think it leads to neater spreadsheets, and there are other features that can take up the slack that mixed-content columns used to (like multiple views within a page/tab). Sorry you hit editing glitches - thanks for letting us know, and if you have the time I'd like to hear more here or via the "give feedback" button in the app. Cheers!

cto here - I signed up to work on Grist because I like the UI, like it as a spreadsheet format, and think it has potential for decent version control - something lacking in data projects, and that I've been interested in for way too long - here I am ranting in 2013 https://okfnlabs.org/blog/2013/08/08/diffing-and-patching-da...

In Grist spreadsheets you can store a dictionary in a cell (here's a random example https://support.getgrist.com/examples/2020-08-invoices/#ente...). It is sometimes handy though it gets ugly if overused I'd say. Grist just became open-source btw https://github.com/gristlabs/grist-core#grist (disclaimer, I work there)

You're right! I package public trained models for natural language queries (https://github.com/paulfitz/mlsql/), and the models are bad at saying they don't know. I'm optimistic though. There's significant year-on-year improvement (driven by real progress in NLP), and the training datasets are getting more interesting. There are now conversational datasets (e.g. https://yale-lily.github.io/cosql) where the model is trained to ask follow-up questions, and an explicit goal is "system responses to clarify ambiguous questions, verify returned results, and notify users of unanswerable or unrelated questions". That could be a big win.

Grist spreadsheets have a page where you can review all formulas as code https://support.getgrist.com/formulas/#code-viewer - I find this handy when figuring out someone else's spreadsheet, or my own once I've forgotten how they work, and would love to see this become a standard spreadsheet feature (disclosure: I work on Grist)

Neat! Looks like that site also has Lotus 1-2-3 https://www.pcjs.org/disks/pcx86/apps/lotus/123/1as/

You can download VisiCalc from Dan Bricklin's site at http://www.bricklin.com/history/vcexecutable.htm and run it using dosbox. Tip: a formula is something like "+A1" instead of "=A1".

Nice! Would be interested in how this compares to e.g. handsontable[1]. What do you think of audreyt's ethercalc[2] myliang?

[1] https://github.com/handsontable/handsontable

[2] https://github.com/audreyt/ethercalc

How about sqlar as a container format? https://sqlite.org/sqlar.html A regular sqlite database file, with anything you like in it. Mountable as a file system with sqlarfs. Written by the sqlite guy.

(DeepMind is part of the Alphabet group, true, but is based in London, with some research people in Canada and some applied people in California.)

Sign me up for your product or service, would definitely be interested in seeing code for this.

Volume up and down are used as modifiers. Sounds crazy but it works.

For people like me trying out the .deb on linux: the executable to run is (somewhat cheekily) "code"

Tully Foote has posted this on Willow Garage's blog: "Willow Garage has decided to enter the world of commercial opportunities with an eye to becoming a self-sustaining company. This is an important change to our funding model."

http://www.willowgarage.com/blog/2013/02/11/willow-garage-ch...

There is also Dan Bricklin's socialcalc https://github.com/DanBricklin/socialcalc which has some relatively active forks including audreyt's ethercalc https://github.com/audreyt/ethercalc

You can see what the spreadsheets look like at http://www.ethercalc.org/

For just viewing and editing spreadsheets without social features, you can ignore the servers included, the javascript widget works just fine by itself.

Dan Bricklin co-authored VisiCalc, the first spreadsheet program for the masses.