For too long, web browsers have dominated the market for applications that are consumed by document readers, glad to see some competition at last.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
21–30 of 30 posts
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#22Earlier quoted context omitted.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
Yes. There is an immediate tactile sense that everything is slower. This matters hugely to me because I regularly work with sheets that are already large and slow
That said, I was always hoping there was some option out there proving it to be possible! Here’s to hoping it comes soon if not, because I personally don’t understand why a spreadsheet needs OS calls to function performantly.
For simpler apps, I have hope for “local first web development”, which seeks to move everything possible onto the client device and then do asynchronous syncing - thus handling the latency issue you mentioned. I’m guessing this is basically already what Google sheets does
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#23Related to this, I recommend Felienne Hermans' talks on Excel as a functional programming language - https://www.youtube.com/watch?v=0yKf8TrLUOw .
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#24For too long, web browsers have dominated the market for applications that are consumed by document readers, glad to see some competition at last.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#25For too long, web browsers have dominated the market for applications that are consumed by document readers, glad to see some competition at last.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#26Related to this, I recommend Felienne Hermans' talks on Excel as a functional programming language - https://www.youtube.com/watch?v=0yKf8TrLUOw .
Not related to this, excel has just introduced Python support and it's a pretty wild take on it all. You can have Python code in any cell, and all the cells share a namespace for their variables. GL HF everyone. Ideal: Functional. Reality: Stateful spaghetti
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#27Related to this, I recommend Felienne Hermans' talks on Excel as a functional programming language - https://www.youtube.com/watch?v=0yKf8TrLUOw .
Not related to this, excel has just introduced Python support and it's a pretty wild take on it all. You can have Python code in any cell, and all the cells share a namespace for their variables. GL HF everyone. Ideal: Functional. Reality: Stateful spaghetti
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#28For too long, web browsers have dominated the market for applications that are consumed by document readers, glad to see some competition at last.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
Yikes, I hope not. Powerful, native desktop apps are what got me interested in computing, not the Web.
Browsers are a good use-case for kiosk-based computing, I don't disagree, and there are many use-cases for effectively booting directly into a browser (i.e. ChromeOS) but when it comes down to wanting pedal to the metal, I always reach for native technologies. It provides both a better developer and end-user experience.
For instance, you asked what a spreadsheet program might need the OS for. Many spreadsheets are used as databases (even if they shouldn't). A high-speed asynchronous I/O API is pretty critical for responsiveness, without going through the interpreter and GC of a browser.
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#29For too long, web browsers have dominated the market for applications that are consumed by document readers, glad to see some competition at last.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
Re: CellLVM: A proof-of-concept LLVM to Excel spreadsheet compiler
#30Earlier quoted context omitted.
Interesting dichotomy! Ultimately I see the browser as a document viewer. As someone who’s trying to kill desktop apps once and for all with the progressive web app revolution, I’m very curious: what excites you about getting this in excel instead of, idk, Excel.com? Do you think there’s a fundamental performance difference that matters?
I like having the application on my machine so I can use it without permission or contact with the remote server, at least after licensing is completed. I find myself working offline more than you might imagine, and I also run desktop software that is 10 to 20 years old pretty regularly. I shudder to imagine what will happen to today's generation of SaaS software once it is past commercial viability. Probably it will…
it's the convergence of Java's "program once, run everywhere"
conceptually, this abstraction is similar to docker or k8: you're fundamentally disconnecting the metal from the OS from the app runner.
it's pretty much a useful abstraction with lots of parallels.
imagine the browser as the final unified app store.