Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
Uncategorized
24
Posts
13
Posters
3
Views
-
Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
> there is also a large class of users who, while they are not computer specialists, would be willing to learn to interact with a computer in a reasonably high-level, non-procedural query language. Examples
of such users are accountants, engineers, architects, and urban planners
https://dl.acm.org/doi/epdf/10.1145/800296.811515How quaint!
Anyway, tell me again how LLMs are making non-techies able to talk to computers. 👀
@rysiek COBOL (Common _Business_ Oriented Language) tried to do away with programmers too. "Let the business decide what they want the computer to do!" My ass.
Fast-forward 60 years: companies pay a 50% premium for entry level devs who can spell "mainframe" because _absolutely no-one_ can be arsed to code in COBOL.
I bet there are dozens of languages more like this if you count niche corporate ones that tried and failed miserably to get rid of devs.
No nerds, no computing. Period.
-
undefined 77nn@goto.77nn.it shared this topic
-
1. We need specialists to talk to computers.
2. Specialists are expensive, can we avoid them?
3. Ah, what if instead of using difficult programming languages, we used something closer to natural language!
4. Wait, natural language is imprecise, which leads to Bad Things.
5. Okay, natural language, but Precise™!
6. Wait, that's just a programming language and non-techies hate it.
7. A new type of techie emerges, specializing in the new tool.
Tee hee "prompt engineering".
@rysiek
Scotty: "Computer, write me a prompt for this other computer...." -
Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
> there is also a large class of users who, while they are not computer specialists, would be willing to learn to interact with a computer in a reasonably high-level, non-procedural query language. Examples
of such users are accountants, engineers, architects, and urban planners
https://dl.acm.org/doi/epdf/10.1145/800296.811515How quaint!
Anyway, tell me again how LLMs are making non-techies able to talk to computers. 👀
This idea doesn't even fail sometimes, you just end up with really wonky languages dominating certain fields with statisticians using R, engineers using MATLAB script, and just about every field making such elaborate excel sheets it may as well be a programming language.
Of course, those all have the advantage of still being deterministic and free of hallucinations, but the innards of those scripts and spreadsheets look like hell and the interface is wonky while LLMs are shiny-in, shiny-out.
-
@craignicol @rysiek
That's one of the hard parts, for sure.- What problem are we solving?
- Are you sure that's actually the problem?
- Is this the best framing for it that we can think of?
All these are far harder than the actual encoding, IME.
-
Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
> there is also a large class of users who, while they are not computer specialists, would be willing to learn to interact with a computer in a reasonably high-level, non-procedural query language. Examples
of such users are accountants, engineers, architects, and urban planners
https://dl.acm.org/doi/epdf/10.1145/800296.811515How quaint!
Anyway, tell me again how LLMs are making non-techies able to talk to computers. 👀
It's not that unrealistic. I knew a couple of secretaries at a former employer who kept getting asked to do reports, bought themselves SQL books so they could generate reports directly.
As a software developer I'd sometimes ask their advice on quirks about the database.
Their skills got them promoted to "administrative assistant" positions.
But the problem has never been that computer languages are hard to learn. It is that knowing how to solve abstract problems by organising and manipulating information is a skill that people don't want to pay for.
-
@rysiek
Scotty: "Computer, write me a prompt for this other computer...."@spytfyre @rysiek Voice recognition and prompt engineering? This can only go well: https://www.youtube.com/watch?v=HbDnxzrbxn4
-
Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
> there is also a large class of users who, while they are not computer specialists, would be willing to learn to interact with a computer in a reasonably high-level, non-procedural query language. Examples
of such users are accountants, engineers, architects, and urban planners
https://dl.acm.org/doi/epdf/10.1145/800296.811515How quaint!
Anyway, tell me again how LLMs are making non-techies able to talk to computers. 👀
@rysiek You need to go to SQL (school) and take a SQL (sequel) to not break your SQL (skull).
-
@rysiek Not only that, but I remember MS Access being sold as making databases so easy, your CEO can use them - and having to clean up the aftermath.
It used a query language similar to, but not quite the same as, SQL :headdesk:
-
@rysiek Not only that, but I remember MS Access being sold as making databases so easy, your CEO can use them - and having to clean up the aftermath.
It used a query language similar to, but not quite the same as, SQL :headdesk:
-
It's not that unrealistic. I knew a couple of secretaries at a former employer who kept getting asked to do reports, bought themselves SQL books so they could generate reports directly.
As a software developer I'd sometimes ask their advice on quirks about the database.
Their skills got them promoted to "administrative assistant" positions.
But the problem has never been that computer languages are hard to learn. It is that knowing how to solve abstract problems by organising and manipulating information is a skill that people don't want to pay for.
> But the problem has never been that computer languages are hard to learn. It is that knowing how to solve abstract problems by organising and manipulating information is a skill that people don't want to pay for.
Exactly this is my point. And that no technical solution will ever exist for this particular social problem.
-
-
@rysiek BASIC was my first programming language!
...wait, it's worse. I have forgotten the language, other than it having GOTO and GOSUB.
@KatS add IF, RETURN, FOR and NEXT and that was basically all of BASIC.
Really great for writing spaghetti code. Complicated almost like assembler, but much slower!
-
Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
> there is also a large class of users who, while they are not computer specialists, would be willing to learn to interact with a computer in a reasonably high-level, non-procedural query language. Examples
of such users are accountants, engineers, architects, and urban planners
https://dl.acm.org/doi/epdf/10.1145/800296.811515How quaint!
Anyway, tell me again how LLMs are making non-techies able to talk to computers. 👀
@rysiek I had exactly the same idea: there was era, when databases, including SQL, were marketed as solution for everything and supposedly, no one will have to do programming any longer. And if we were about to do programming anyway, we would do it by clicking by mouse and drawing algorithms (which of course never happened...)
(except there was parallel rise in spreadsheets, which did almost the same thing, but differently...)
-
-
Remember how SQL (still called SEQUEL) was supposed to make non-techies able to talk to computers?
> there is also a large class of users who, while they are not computer specialists, would be willing to learn to interact with a computer in a reasonably high-level, non-procedural query language. Examples
of such users are accountants, engineers, architects, and urban planners
https://dl.acm.org/doi/epdf/10.1145/800296.811515How quaint!
Anyway, tell me again how LLMs are making non-techies able to talk to computers. 👀
@rysiek hahaha the ERP system the company I work for is using just switched their whole DB-system around to be able to give their customers an AI-supported drag and drop GUI "no-code" alternative to SQL.
It is limited in its abilities and soooo painful to use. Before SQL queries have been free. Now "code" queries cost as well as the "no code" alternative.
They promised their AI would transfer the old queries to their new DB structure.
IRL only the most simple ones have been transferred and I had to redo most queries.
Everyone in the "community" (aka customers who can't easily escape) is hating it while they are celebrating it as the new hot shit.
Feed RSS
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@LaVi
neve!? che roba è? C'è il mandorlo che ha fiorito da un mese, sabato ho visto un albicocco in fiore, una vite sta già germogliando, i tulipani hanno emesso il bocciolo, i narcisi sono in piena fioritura e ho l'orto infestato dall'acetosella in fiore.
-
@ALFA @tizianomattei @LaVi @Uilebheist @ju
A Caprona, una sera di febbraio,
gente veniva, ed era già per l’erta,
veniva su da Cincinnati, Ohio.Mi fa impazzire che sia scritto in terzine (ne parlavamo con @oblomov giorni fa)
-
Efficace, il servizio di #PresaDiretta sulla fuga degli italiani dall'Italia, brava #ElenaMarzano!
-
Ho un'altra idea carina per questo coso usando una striscia WLED, domani se ho voglia vedremo
-
@aeva thank you aeva :3
-
@ju @ALFA @Uilebheist @matz sì, be', se andiamo sul protoindoeuropeo ci si può infilare pressoché qualunque cosa :D
-
hey me, hope you're doing well. out there in the future.
i've been coding in react a lot lately. i think the unified data flow really makes sense in wrangling in this frontend complexity.
i'm a little sad by all the build tooling though. It feels more complicated to scaffold a basic UI now than it was ten years ago. I'm sure ten years from now we'll have some sort of reactivity system without all the complex overhead.
right? right?
-
"1 giovane laureato su 10 se ne va", #PresaDiretta #ElenaMarzano
