Re: was: Some info about Python: Back to batteries

hi there! :)

really related, everyone must read it at least twice!!!!4

(a bunch of translation links are there, at the bottom of it, as its not an easy reading, and please add more translations to this gem to spread the word, in case if you are fluent in klingon, elf or whatever like so. thx! :) )

(i already dont strictly care about the title mess behind the main topic, but im still about the same... :D )

The existence of Python doesn't make Lua useless as a general purpose

language.

Sadly, I have to admit something though: today if I have to choose a

backend stack for a serious Web project it will probably be Python-based,

and I use Ruby on Rails at my current job.

thats just a personal choice, well, a popular one, especially in the business sector... ive selected github://kernelsauce/turbo to be my server engine, cuz it knows everything that would be a pita to make it all on my own, and theres the more serious openresty for the lazy ones who need something more, but i would go for less if it turns out that turbo isnt really good enough :D

actually, and in general, i hate bloatware cuz i dont wanna touch what i cant learn. otherwise, i would break something silently (or loudly) and shoot my leg. sure thing, there are some more rare components that are based on well defined purposes, that can be accepted even as-is, cuz they are stable, without significant garbage included, and with pretty, smart, tasty and sane codes behind them, like lua, luajit, curl, unicode, linux, libressl etc... in contrast with the things from the hype driven development bloatcollectors, where no1 knows what happens behind the scenes, and they arent really reliable... i also dont like language level package managers cuz of this, too many traps are included for a high cost. sure thing, i can do instant miracles, customers are happy, i got paid and run away before it start to burn in flames.... thats why i wanna break free from my web dev role... (nope, im not like those ppl, ive just described.)

just imagine a case when u wanna build something that lasts, and that can be relied, and where u can do whatever u want. that requires a different mindset than what we are currently oppositing lua with... nobody have time to keep away bitrot from things like fancy new n+1 web frameworks, they come and go, wordpress likes to catch on fire under every second constellation, and uve got no idea why it happens, cuz of those modules that are JustWorks™, cuz the wp must be updated against vulners and what not, but the modules are often developed on a hit n run manner, and im still not about the amount of those modules that ppl will use... (no, hopefully i didnt have to play much with wp and the like...)

but yea, tasty opensauce! u can do eeeeverything u can dream about! u only need to learn a dozen of languages, frameworks, read the commits, follow the mailing lists be on irc know how to divide with zero and ur there! ... ohh, theres the community! just send a ticket and profit! (spoiler: kinda often nobody cares https://www.monkeyuser.com/assets/images/2016/10-expectation-vs-reality-open-source.png ) i dont even really like to ask things on forums, either cuz its just a simple thing to save myself a few clicks when i see experts who could sacrifice a fraction of that time it will be to discover the rabbit's hole, when they already wasting their time there, and then i get rtfm&stfw, or when its about complicated stuffs, then tldr&silence comes in exchange for the time of telling possibly awesome basic ideas/methods/whatever...

so yep, making important things must be reliable, that requires ppl to be able to read and touch the code base that they use. its even better profitwise in the long run, and there are always enough project ideas out there in the wild to pick one thats not just a one shot project, but such as ppl can develop their customer circles around their own beloved code bases...

so back to my initial topic, a web server isnt that much of a big thing, compared to a freestanding gui that can make all of my dreams happen, and tekui is a thing that compares to nothing else out there (as far as i know right now), its just not overengineered, it can be read up to the last letter, it can serve all my basic needs out of the box, and it is flexible enough to make me able to implement such graphical things that doesnt exist in any general gui toolkits, without making my nose bleeding with its initial code base - even if it isnt 100% polished, and even if its nearly abandonware (maybe thats a good thing; otherwise its just considered to be basically feature-complete, and its main author is responsive, just busy in general). its a gui that can be touched, that doesnt really bleeding, that wont run away with its development governed by a huge community (that makes a bloatware) and its already really much powerful and flexible. i love it! :)

on pythonland, ive seen things like wx binding (it was 2.x only, when i already wanted stuffs from 3.x, it relies on an external lib, based on c++ thats just an another untouchable beast, its bloated, and what not, but i must note that i was nowhere to my current knowledge in those times, and it still cant draw a single pixel on a canvas, if im right :D ) and tkinter (that came from the book that ive initially used to learn python from) is a slow beast that supports 4 different programming languages that i dont care about (i wont learn all of them to be compatible and i wont clean it up and i dont wanna follow the crowd) and its written in some kinda klingon dialect (tcl)... the nearest to my needs were racket and io languages so far now, but neither those can really competite with my current funds... (im looking up everything i can find in the wild, as i dont wanna be on the wrong ship, cuz its hard to migrate, but ive spent here a few years now, and nothing could say that lua isnt what i was looking for or that there is anything that could really beat it...)

otherwise, the universal graph handler, that im talking about on a regular basis, would be kinda much painful with the fundamentally different dataset types in python, while the miracle behind it is that every data structure is a graph, so a graph handler can handle anything, and can be the connection between all the graphs, like currently its for matching lua tables with the filesystem, thats an awesome thing, it can be used for iterate, search, compare, repalce, dump, serialize, merge and what not anything that is based on graphs, that is actually any kinda data. even the ast is a graph. (no it doesnt know all of these right now, but these can be done without the need of refining this much different beasts when ive got an idea for one of them). the universal parser will be the next level of this, as that will make a graph (or any special case of it), as everything is just a bytestream on the disk. python would be a huge pita for such tasks, while it would also be really weak for it, c wouldnt be enough flexible, cuz its compiler based, and if there is something big to link against, then its also a slow task, so no live-coding is possible there, even if that isnt as much of a strict requirement... aaand, for an ai (not the kind what ive explained in my previous mail today :D ) the most important thing is to be able to read understad(?) and modify its very own codebase, thats again a pita with such stuffs as python that is huge bloated and has too many sugar included... taking a lotsa various modules would make a huge mess cuz of the various coding-styles, and thats not really a good stuff, but with an universal parser+graph handler, its not that hard to extract essence from here and there in case of heavy lifting or reshaping things without tearing the roots (that gives an another kinda pain when it must be done more than once). using the right tool to the right job brings in just that serious pain that ive talked about above, on besides with the article i started my writing with. its good for some cases, but actually not that much... especially when u wrote something in python but next time u need it in bash and then c++ and then in br*infuck and that way u must reimplement all of ur precious gems that once were considered to be ready.

really much ontopic:

one thing that really makes us love lua, is that its really flexible, we have different style, need, u name it ... and that n+1th implementation is really necessary to keep this spirit, making too much standard stuffs could slowly take this away. lua is the just enough fund, and this is good, and this is why we love it. otherwise ive got no idea what area of coding falls far from lua, there are multiple gaming engines, there are various guis, there is torch for current "ai" stuffs, there is gpu rendering, we have an awesome browser (luakit), phone centers and self-driving cars use lua, cern as well, we have everything that can be wished for networking (even wireshark and nmap) we have lua-users and luarocks to find these gems, we have nice resources (thats started right with the best refman out there), and we have every single c lib!!!!! we have uncomparable power by the nice c api, all the goodies included in luajit, the flexibility of the metatables, that i dont even use cuz i already has enough flexibility without the need for going farer, but i say they are nice to have...

lua is bound to c99, and this is the right choice for the right flagship that is belongs to, above that, much pain comes. and actually it is a hacker paradise and fund, if it would require any less hacking, then it would become an untouchable bloatware sooner or later, and ppl who need a language to be hacked will choose lua even tomorrow, and there are no much ppl around who hack python internals or who would try to fork and modify c++...

i think that one thing to spread the word is to express better that lua isnt just a language to embed cuz it cant make things done on its own... the other one is to laugh at the hype out there and enjoy the party here! :D

uhm, sorry for the amount... bests to all of u folks! :)

Sent with ProtonMail Secure Email.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

On 2020. January 19., Sunday 18:43, Pierre Chapuis catwell@archlinux.us wrote:

Here it is. We're asked "why should we enable Lua as a general purpose

language?" Pierre used Lua's natural speed to make it a better CGI

language than PP&R. Imagine how many more of us would do this if Lua

had tested and curated packages for DOM, XML, and the like, so that you

never have to spend time evaluating competing softwares, and whatever

software you write is likely to have its dependencies available

anywhere the Lua Standard Library exists.

The existence of Python doesn't make Lua useless as a general purpose

language.

Sadly, I have to admit something though: today if I have to choose a

backend stack for a serious Web project it will probably be Python-based,

and I use Ruby on Rails at my current job.