[fjsolwmv]

No. An domain-specific SaaS app takes all the power away from the people writing their spreadsheet programs. Only one app is needed, a hosted spreadsheet with backup and versioning an replication. The only domain-specific business is in writing plugins.

[peatmoss]

I agree that purpose-specific tools aren't the answer, unless the question was, "can I have a straight jacket?"

Flip side, my feeling is that, if an analysis was worth doing, it's probably worth doing right. And to do it right, it should probably be done in a way that one can show their work. I like where the R community in particular has gone in dropping the barrier to reproducible analysis to roughly the same place as Excel.

[Jesus_Jones]

My friend's startup has a project that matches up with a lot of these requirements, called Coda. It's easy to query data and build visualizations, has a spreadsheet data model, and has an easier programming model. Excel did a great job getting people to "program" their spreadsheets though.

I also wonder about google sheets capabilities. It does work so well for many people working on one spreadsheet, except for undo. The major flaw I've seen is undo, because in google sheets it was global undo, not personal undo.

[ellius]

I wrote this in another thread, but I think the next move should be to create an Excel-compatible spreadsheet program with a few additional constraints inspired by programming. Make primary keys required, each column needs a type, etc. These are concepts I’ve successfully explained to non-technical coworkers to great effect. By enforcing them in a tool, I think you would get about 80% of the benefits of true software without all of the overhead that entails.

[_rerr]

This actually exists inside excel since 2010, it is called get & transform (known before as power query) and lives in the data tab. It allows the user to mash data from different sources (databases, csv, excel files, current workbook) and perform sql-like opérations : filter, join, etc. In a wisiwyg-record-macro fashion. I find it very intuitive and the steps are easily reproducible as each transformation step is translated into a language derived from F#. It is a game-changer for business users that dare using it, as it is rigorous and precise like a classical SQL flavor, but also usable with zero training.

Since it was a very powerful tool but struggled getting traction, they are using it as the core query tool for power BI (alongside DAX for great dashboarding).

Well done Microsoft.

[mtpn]

Airtable does something like this. I don’t know how excel-compatible it is, but the core of data types and something similar to primary keys are enforced in a pretty intuitive way, and basic formulas work fine. Haven’t done much advanced stuff with it.

[barrkel]

There's a reason chefs don't use multi-tools; purpose-built tools always beat generic tools when you need to maximize efficiency on a task.

[njarboe]

I would say a chef's knife is the ultimate kitchen multi-tool. Having one great knife you really know how to use for cutting everything is much more efficient than a bunch of different knives.

[AimHere]

Of course the multitool is the single point of failure. If the chef isn't careful, that one great knife can be the vector that means ALL the customers get salmonella, not just the ones who ordered the chicken!