[kkfx]

Mh, beside the issue of a spreadsheet itself [1] and the fact it's shared to the world [2] the real computation must be another: since crystal balls are out of service but luck is blind and bad luck sees very well... The right computation, witch happen to be pure logic NOT reducible to math, a thing soo many people forget is IMVHO "how likely is a diesel/gasoline shortage in the subsequent 1-8 years future?", "can I eventually charge LOCALLY on my own energy (p.v. typically) a BEV?", "can I sustain to buy a new BEV at current prices every 5-8-10 years?".

The answer to this three questions is not numeric, but it's the key. Having an EV to protect themselves against fuel shortage is meaningless if you can't charge it autonomously, being able to buy one but unable to sustain it's CapEx is a desperate action, that tend to have equally desperate consequences, otherwise if you are not living in a country who produce LOCALLY enough oil and have a population enough ready to fight if needed to get it an EV is not economically convenient, but it's a guarantee to being able to travel. The importance of such expensive guarantee it's very personal, depending on so many parameters that's more easy to estimate in person instead of trying summarize all of them for a generic solution.

Really of topic but IMVHO more HN-ish is the means the author choose to share: if in 2022 someone willing to do something AND SHARE IT to the community, a technically sound one, resort to such tech... Well... That means we are really in a sorry state. It's NOT a critic to the author, nor an indirect encouragement against sharing, is instead a sore consideration of the actual state of computing and widespread tech knowledge.

[1] spreadsheets should NOT EXISTS at all, they are a tentative from another era to provide a flexible calculator for tabular data to the masses and such tentative was and is a FAILURE: instead of simplify it complicate things and instead of empower users if force them toward very bad paradigms.

[2] witch prove the very limited scale of such collaborative model, perhaps to be confronted with the SCM model(s) AND the idea of sharing "active" documents (formatting + code) like org-mode files in Emacs or Jupyter notebooks etc where any user get the doc (so the logic) and instantiate it alone on in a small cohort witch is the sole example of live collaboration limited effectiveness behind the WOW effect.

[sokoloff]

Spreadsheets are one of the most popular and widespread computing platforms out there. Excel is almost surely the single largest no-code/low-code install and monthly-active-users base.

I don’t see spreadsheets as a failure nor evidence of poor state of computing or technical knowledge. If anything, I think they’re useful, practical, extensively used, and evidence that computing is working for people.

[shagie]

> Excel is almost surely the single largest no-code/low-code install and monthly-active-users base.

I'd argue even further that Excel is the most common flavor of LISP installed.

  =IF(A1 = 1, A2 * 3, MAX(A3, A4))

is remarkably similar to:

  (if (= A1 1) (* 3 A2) (max A3 A4))

The difference is mostly infix vs prefix and the naming of variables. Not only is Excel the single largest low code install, but its also a gateway to LISP if you introduce it to people correctly.

[kazinator]

Never mind IF, Excel now has LET and LAMBDA.

There is an n-ary AND function also; the documentation doesn't mention whether it's short-circuiting or not, and it doesn't return the last value if they are all true, just a Boolean true.

[shagie]

I haven't had the opportunity to use excel for a problem that would require a lambda yet.

But for those interested https://www.microsoft.com/en-us/research/blog/lambda-the-ult...

My two examples were more of a "show a clear correspondence between the syntax of excel and lisp" for the "if someone knows basic excel, you can introduce them to lisp."

[kazinator]

If that worked for C programmers on Unix who use GNU Make, that would be cool.

   $(if $(filter foo.o,$(OBJS)),this,that)

[kkfx]

If you want AutoCAD offer a real lisp, AutoLISP, some users use it regularly without even knowing that's a real and powerful programming language. That's anyway do not make autocad, nor in general that kind of CAD something better then an archaic crap since the '60s. Parametric CAD are the modern and technically sound solution since the early '70s, even if in some sector they never get traction due to technicians ignorance and reactionarism.

That's a good proof of inability to evolve of sooooo many people...

[kkfx]

Many things are very widespread and are absolutely crap... Indeed MOST widespread and heavy used things happen to be crap.

Their failure is the fact that instead of simply they complicate life. And their users fails even to comprehend that since they do not know any other option. Oh, BTW the "no code/low code" is the biggest failure: users programming of classic system is the technical success, something simple enough that all users use without issue and still technically sound and effective. Dummy low-code/no-code modern environment who have already touched incredible horrific peaks like RENAMING a gene because excel convert it in a date prove only one thing: masochism is common and popular...