41–50 of 50 posts
Earlier quoted context omitted.
If you're the M Kromberg, any chance you could ask J Scholes to whip up a dfns version of Iverson's 1960 Simplex algo for us?
Sadly, John Scholes passed away in 2019. https://aplwiki.com/wiki/John_Scholes
Earlier quoted context omitted.
Real time processes are indeed critical in some contexts. And, arguably, we should be putting more effort into some of those contexts. For the APL family, that's often a call for code which works with external interfaces. In many potential APL environments a handful of well designed custom primitives would get you where you need to go. But no programming language is adequate for all roles, and that includes APL. (It…
Everything involving a user is a real time thing. But that's now how the software industry views it, somewhere along the line the user became a peripheral rather than the other way around.
Earlier quoted context omitted.
genuinely curious -- do you not feel the burden of being locked in with cpython, gcc/clang, rustc, ...? or do you maintain forks of those?
I don't feel locked in by the license of those projects. They have permissive licenses. Anyone can maintain them. I don't need to personally maintain a fork in order for the possibility to maintain one to exist.
proprietary programming systems typically come with escrow agreements, providing access to the source code to customer e.g. if the vendor stops supporting the product. that wouldn't be materially different from another large corporate picking up e.g. llvm/clang if apple/google stop being interested.
Someone in the HN comments called out that J is the proof that Iverson's "Notation as a Tool of Thought" is a failed idea. Even Iverson himself, the inventor of "Iverson Notation" (proto-APL) as a better math notation, the inventor of APL, who used it to design the IBM 360's processor and then turned it into a programming language, abandoned it for ugly ASCII scribble (J) because that was more convenient. I guess the…
The inventor of APL, Ken Iverson, also made the language J which is very much influenced by APL but does not use the special characters of APL. Think of it as ASCII-APL. It is a very nice language to program in and they have a decent user community. It is open source and is incredibly fast and looks at the first glance as cryptic as APL. https://www.jsoftware.com
https://en.m.wikipedia.org/wiki/APL_syntax_and_symbols#Monad...
Earlier quoted context omitted.
I don't feel locked in by the license of those projects. They have permissive licenses. Anyone can maintain them. I don't need to personally maintain a fork in order for the possibility to maintain one to exist.
what is your realistic estimate for the capability of any non-corporate group to pick up maintenance of one of the open-source projects? proprietary programming systems typically come with escrow agreements, providing access to the source code to customer e.g. if the vendor stops supporting the product. that wouldn't be materially different from another large corporate picking up e.g. llvm/clang if apple/google stop…
I appreciate you mentioning escrow agreements. That's a valid point. However, "typically" is not "guaranteed". It looks like MicroAPL, a company so involved with APL that they put it in their company name, no longer sells or develops their APLX product (even though the company appears to still exist). They stopped selling and developing it in 2011. Was it guaranteed that Dyalog would take the project over and host the binaries and documentation in 2016? I don't think so. Good on Dyalog for doing it. The fact that nothing was guaranteed is what makes their actions meritable.
Yet, as far as I can tell, it's only the binaries that are hosted. It appears to receive no further development, no support for new architectures, etc. The thing which is materially different is that the source code is not available, there is no guarantee that it ever will be, and no one, outside of maybe one of two people, have legal authority to use the source. Your point about escrow stands and is valid. It's also fundamentally different than a project having an open license.
PS: sorry, I didn't directly answer your question. You asked for an estimate of a non-company continuing development of one of the current freely licensed projects you listed. For something like LLVM, I think pretty high. A lot of people are involved in it, a lot of people prefer it to GCC, and it has been worked on by many people for many years. But knows. Maybe the companies pull out support and it dies. Maybe not.
That's also a separate question: what's the chance of an open project getting continued support. My lament is that many APLs aren't open to begin with and are not guaranteed the chance to get support.
Someone in the HN comments called out that J is the proof that Iverson's "Notation as a Tool of Thought" is a failed idea. Even Iverson himself, the inventor of "Iverson Notation" (proto-APL) as a better math notation, the inventor of APL, who used it to design the IBM 360's processor and then turned it into a programming language, abandoned it for ugly ASCII scribble (J) because that was more convenient. I guess the…
Lots of people do Python so it makes sense that a lot of leaderboard solutions are in python! Hardly rigorous (n=1) but I got on the leaderboard for day 2 using K. Other people have before in array languages too.
Does it? If you saw a bike racer riding an old steel commuter bike with mudguards and pannier rack, because lots of people ride those, wouldn't that be weird? Bike racers ride lightweight carbon fibre bikes and they do so because other bike racers ride them and they could not be competitive with something much heavier and less aerodynamic. They use every advantage they can get.
If there was a programming language which was unusually good for rapid data crunching, over several years of AoC wouldn't you expect it to become the dominant language on the leaderboard?
Earlier quoted context omitted.
Lots of people do Python so it makes sense that a lot of leaderboard solutions are in python! Hardly rigorous (n=1) but I got on the leaderboard for day 2 using K. Other people have before in array languages too.
> " Lots of people do Python so it makes sense that a lot of leaderboard solutions are in python! " Does it? If you saw a bike racer riding an old steel commuter bike with mudguards and pannier rack, because lots of people ride those, wouldn't that be weird? Bike racers ride lightweight carbon fibre bikes and they do so because other bike racers ride them and they could not be competitive with something much heavier…
Each year probably only a single digit number of people using an array language actively try for leaderboard spots, so if anything it's impressive any of them manage compared to the 100x or 1000x number of python/etc programmers also competing.
Earlier quoted context omitted.
> most people vastly prefer easy and DWIM over simple and precise Most people also vastly prefer not having to work on difficult problems. Iverson was clearly trying to provide a notation that would help the (admittedly smaller) group of people who find solving difficult problems interesting and valuable.
In the beginning of the computer age all problems were difficult. If not because of the problem itself then because of the limitations of the machines. It's amazing they did as much as they did with so little, your average Arduino makes the best hardware from those days look puny.
Earlier quoted context omitted.
> most people vastly prefer easy and DWIM over simple and precise Most people also vastly prefer not having to work on difficult problems. Iverson was clearly trying to provide a notation that would help the (admittedly smaller) group of people who find solving difficult problems interesting and valuable.
If you're the M Kromberg, any chance you could ask J Scholes to whip up a dfns version of Iverson's 1960 Simplex algo for us?
John once did a live demo where a negative number (like ¯17.01) suddenly appeared in his APL session log. He looked puzzled for a moment, continued with his demo script for a little while, and arrived at the statement ⎕DL ¯17 (delay negative 17 seconds, and return the actual amount of time waited). He exclaimed "Oh, THAT's what that was!" and continued with his demo.
John had a fantastic sense of humour, if you have not seen them I recommend taking a little time off to watch some of the recordings at https://johnscholes.rip/video/