93% of Paint Splatters Are Valid Perl Programs (2019)
mcmillen.dev
mcmillen.dev
I once wrote an entire CMS in Perl.
I still wake up, screaming, sometimes.
[UPDATE] It appears as if my flippant remark caused upset. I am sincerely sorry that this was the case. No offense was meant.
@peteretep, you have my sincere apology. It was a flippant remark. At the time I wrote it (mid-1990s), Perl was the best way to write portable server-side code. I took the time to learn it, and got fairly good at it. Since then, I have found my muse in other languages, but I sincerely did not want to give offense. I doubt that deleting the comment would help.
EDIT: verb tense changed
So, @peteretep, you have my sincere apology. It was a flippant remark. At the time I wrote it (mid-1990s), Perl was the best way to write portable CGI code. I took the time to learn it, and got fairly good at it.
Since then, I have found my muse in other languages, but I sincerely did not want to give offense. I doubt that deleting the comment would help.
Which all goes to show people clot up into tribes over any silly thing; and that its better not to wrap your identity too much around your tools.
Teh Internets tubes are a difficult landscape to navigate. It's almost impossible to write stuff without causing some offense.
Some folks take that as an admonition to never write anything fun. Others (like me) learn to apologize.
I'm fairly good at apologizing, as I need to do it so often. Each time, I learn to ameliorate my approach.
But I'll never stop posting in a humorous way.
It's cool.
I found it hilarious, but I'm a weird guy. I just can see where he is coming from where he meant it as a snide joke, not a personal attack.
I can't tell what's what, but man is it funny, to me it reads as deadpan humor by peterep with no insult intended, but obviously others read it differently.
Perspective is amazing and this has made my day from all the smiles and laughs XD
Ooh. I know this one!
Because they are human, and humans are important.
I care, though that won’t always stop me from writing stuff that may or may not cause offense.
But I care about others. I know that makes me apostate, in today’s rage-fueled insane asylum of society, but that’s how I roll.
This is rarely ever the intention, and I'm 100% certain wasn't the intention of the poster here.
The tools you use shouldn't be your identity. Chill.
I write software for Apple systems. It’s modern, high-quality code, but it’s for Apple.
There are people that have never met me, and have never had any prior interactions with me, that hate (and that’s not hyperbole) me; simply because of that one fact.
I’m constantly reading stuff about Apple, and people that use/like/support/develop for Apple, that is intentionally insulting and demeaning.
I should know better, and wish you nothing but the best. Perl is a language easy to learn, and difficult to master.
You have my sincere respect.
being on edge : tense, irritable
In other words-- GP recounted a memory of tense, anxious feelings about working with Perl.
OP was unwittingly made edgy by reading this and anthropomorphized Perl to express said feelings of tension and anxiety back to GP.
Seems plausible to me.
use strict;
Become more commonplace usage?
use v5.12 and later also automatically enabled strict. Perl 5.10 was released in late 2007, and 5.12 in 2010, so there was enough consensus over a decade ago to add it as the default for anyone opting into the bundle of all newer Perl features.
I could definitely picture programming in perl in the 90s before people commonly implemented use strict; being a complete hellscape though
I'm a -wall kinda guy.
There are people that have never met me, and have never had any prior interactions with me, that hate (and that’s not hyperbole) me; simply because of that one fact.
I'm fascinated to hear this. I quite dislike Apple quite a bit as a company/entity, for a variety of pragmatic and philosophical reasons, but the idea that that would extend to their employees (let alone those working on their platform) is wild to me.
There's lots of folks with axes to grind, that are excellent spinners, and lots of people that are maybe...a bit insecure, shall we say, that eat that spin, and turn it into venom.
The old "Ford vs. Chevy" spat can get like that. People kill each other over which football teams they like.
It's easy to see why folks don't like Apple; just as it's easy to see why folks don't like every other platform.
It's just that there's a lot of really unbalanced folks out there, and even more folks who like to set those people off.
The tools we use don’t define us. If you feel they do, please reevaluate. Perl (C, k8s, PHP, Python 2.x) are just tools. It’s like defining your identity by the brand of hammer you use.
Maybe they shouldn't but professionally they actually do. Try to get a job using a tool you love but doesn't have widespread adoption and it's hard. X years experience using tool Y? Better have got on board earlier. Adopt something early that really takes off while you write key libraries, docs, contribute (previous times o'reilly howto books) and it can make your career.
As a class we get awfully groupthink and boosterist of our tech choices. It's not good but it's completely understandable.
Hammers are completely fungible. Two similar languages used in similar places are not. See Perl and Python. Do you really have no preference seeing jobs that specify one or the other? If so that's great - I try to be thereabouts myself too. But it's clearly not the norm and it's not as irrational as you're making out here for all I dislike it.
edit: Trying to phrase this in a pithy fashion below.
It makes zero difference to you as a carpenter if every other carpenter on earth hates your chosen brand of hammer.
It makes a world of difference to you as a programmer if your language of choice is popular or not.
If you have only worked with PHP 4 for the past 20 years and have no plans to move off that language then yeah you will have a harder time. But getting defensive about it in an online discussion is a weird thing to do and honestly a red flag more than anything. The tools may define you as a professional (to some), but not as a person. Getting defensive is a personal thing. If you do it as a professional that's pretty bad. If you take professional discussions personally, in my experience you'll be very difficult as a coworker.
You lost me here. Even if someone thinks that Perl is an unholy abomination to work with, it doesn't follow that those with expertise in it are less skilled. If anything, it seems like it implies the opposite: expertise in it is harder. The counterpart would be something like claiming that Python is a toy language, simple to use but not suitable for Real Work. This claim is the exact opposite as the Perl complaint, and it seems like _that_ would imply negative things about expertise in the language.
I am very happy to use a nice, safe, memory-avoiding, strongly-typed language like Swift as my principal language, and on a GUI system.
I don't particularly care whether or not some rando on teh Internets tubes thinks I'm a "wuss."
He’s basically ribbing GP by purposefully misinterpreting the comment, so that what gave GP nightmares was his CMS rather than Perl itself.
That said I like perl (And lately, awk/sed) a lot more for things that python is just SLOW at. Plus for systems-level scripts, considering that most linux/BSD installs you encounter on anything at all will have perl installed, let's say you encounter a router or something like that you need to work on somehow internally -- Do you even have room for another interpreter on there? Better be at least initiated in using perl if you want to hack on that. I've also recently found it really useful for rescuing an ancient database file format from the 1980s which was super handy considering other tools choked on it.
I still tend to use it for that.
https://learnbyexample.github.io/learn_perl_oneliners/cover....
Perl is complex, but it is also rewarding when you use it for those tasks it's good at (like gluing code together and text processing).
Getting a hash passed by ref to a sub was a true syntactic challenge that I haven't experienced in a different language. Have to say though, using regex in Perl is chef's kiss
Also, the fact you can write a Perl script, run it everywhere and just not concern yourself with versioning and stuff is beautiful. If one of my .sh script gets over 40 lines, it becomes a .pl file, that's the rule.
Sadly it was not to be.
Moose[1], and more lightweight versions in Mouse[2] and Moo[3], have been around for ages by this point, almost 15 years in some cases. Unfortunately it was never made core (but something like them is supposed to be core in Perl 7), so it was always one of the "best practices" things that you had to be in contact with other people about Perl to know or read some Perl community stuff for, as it wasn't in the official docs.
> good data-science libs like pandas
To my knowledge Perl with PDL[4] was the go-to solution for bioinformatics people for a long time. I never did anything with it or Python's Pandas though, so I don't know how they compare (or more importantly, how they compared back when people actually used Perl for this stuff much).
> easy concurrency
While somewhat wonky, Perl's concurrency has been around in forms for a long time (the Coro[5] library was first released in 2001), if a bit harder to use until recently when like every other language the async/await stuff took over.
The parallelism story has always been pretty good, even if also slightly weird (the ithreads system used forks on Unix to emulate threads, but essentially it's an abstraction you can mostly ignore). Notably, the fact you actually had a threading system[6] with locks and semaphores that worked put it above Python and Ruby in that category.
> ergonomical web-frameworks to beat PHP
This took longer, but Mojolicious[7] seems to fit this category well, easily being used as simple as something like Sinatra to stuff close to Rails (IMO, and without ActiveRecord).
That's not to say Perl is perfect and should have always won out, just that I think the problems Perl experienced (and still experiences, and which recently "Perl 7" is trying to address) are more along the lines of the trade off that was made with regard to backwards compatibility and legacy support, and the effect that had long term on the community.
Perl's absolute adherence to backwards compatibility to the point that 20 year old scripts often run without issue on the latest release of the interpreter was always a boon to those that wanted stability and support. It's still used in some places (for example, where I work) as the primary scripting language for server based utilities that are expected to run in a Unix environment. The problem is that this is a trade-off that hurts more for new users and new technologies as time advances. It's great that our (literally) 20 year old scripts to clean up and do server management still function, but it's not as great that ergonomically Perl requirement to opt-in to many new features (so they don't cause old code with conflicting sub names to fail, as an example) means that there's been an ever increasing list of things you must enable in new Perl lest you deal with annoying old behavior makes it's use for new projects annoying at best. Having a good object system is only the first step, and you're right to call it out because as it was never moved to core, it never got as much use as it could have.
> Sadly it was not to be.
It is sad, because it's a nice language and I love using it most of the time, but I also feel like it wasted well over a decade of what should have been major advances because it was deemed less important than backwards compat and everybody just said "well there's a module for it, that's good enough" when it definitely wasn't sufficient. For chrissakes, Perl only got the most simplistic form of function signatures relatively recently, and what we got is far inferior to what was available and I used as a module a decade ago[8]. Perl's problems were always much deeper than not having some solution or equivalent to some nice tool or way of doing things in other languages. Those came, usually fairly quickly, it was all the other stuff that shed the users to the point that now it's hard to know if there's enough left to drive the change that's needed to keep it alive (in the sense that it's not just in bug maintenance mode).
Sorry for the rant. Obviously I'm sad about the current state of things... I've even mostly decided that I might as well do new projects in other languages, but it makes me sad that it feels like I'm doing so not because they're mostly superior at a level that's hard to compete with but because the Perl community (and this includes myself) couldn't get their shit together long enough to steal the good bits in a way that actually mattered to people looking to try a new language.
1: https://en.wikipedia.org/wiki/Moose_(Perl)#Examples
2: https://metacpan.org/pod/Mouse#SYNOPSIS
3: https://metacpan.org/pod/Moo#SYNOPSIS
5: https://metacpan.org/pod/Coro#SYNOPSIS
6: https://metacpan.org/pod/threads::shared
8: Function::Parameters, as in the bottom of this comment https://news.ycombinator.com/item?id=11633961, as the CPAN page does a crap job of showing it with types which is my favorite part.
I found all those frameworks, and hacking with inside-out objects to be too cumbersome and got in the way, so I just went back to blessed hashes. I liked to keep things simple.
> easy concurrency
I never got into Coro, but AnyEvent was wonderful to work with, and shaped my thinking in other languages... though you could get away with much more in Perl.
> ergonomical web-frameworks to beat PHP
I don't know how anything could be more ergonomic than Dancer (Perl's Sinatra equivalent).
> but I also feel like it wasted well over a decade of what should have been major advances
This, but thinking back at the time, I think what killed Perl was Google choosing Python. At the time, CPAN was the killer app and still had the lead vs all other higher-level languages. But then Google chose Perl instead of Python and as it was the hottest startup at the time where everyone wanted to work at, people started learning Python. Then the data scientists (except the geneticists) migrated over and NumPy, Pandas (and now TensorFlow) kept the momentum.
My understanding it that they're very equivalent in many respects, and popped up at about the same time. Often they're mentioned together as modern options, I didn't for brevity and I'm familiar with Mojo. I think Mojolivious tries to require only core modules with optional upgrade to non-core versions if supplied. Not sure how that compares with Dancer, been a long time since I looked at it.
> I think what killed Perl was Google choosing Python.
Possibly. I saw that as the natural extension of "python is the new teaching language" stuff that started happening in colleges at around the same time, and I think that had a lot to do with it as well. I'm the end it's all about users, and new users, as old users have a natural attrition through finding other things that interest them (whether a new language or non programming pursuits) or retirement or even death. Without new users eventually the community starves. And even if you get enough new users to replace the old ones, without enough new to keep a relevant position relative to all new users, the language will be relegated to a slow decay. That is, even if perl had enough new users to replace old ones, the total programmers grew manifold over the last decades, and without a slice of that pie you're stuck as a has-been language, and eventually people leave due to the natural pull of popular and vibrant language communities.
Master Foo once said to a visiting programmer: “There is more Unix-nature in one line of [Perl] script than there is in ten thousand lines of C.”
The programmer, who was very proud of his mastery of C, said: “How can this be? C is the language in which the very kernel of Unix is implemented!”
Master Foo replied: “That is so. Nevertheless, there is more Unix-nature in one line of [Perl] script than there is in ten thousand lines of C.”
The programmer grew distressed. “But through the C language we experience the enlightenment of the Patriarch Ritchie! We become as one with the operating system and the machine, reaping matchless performance!”
Master Foo replied: “All that you say is true. But there is still more Unix-nature in one line of [Perl] script than there is in ten thousand lines of C.”
The programmer scoffed at Master Foo and rose to depart. But Master Foo nodded to his student Nubi, who wrote a line of [Perl] script on a nearby whiteboard, and said: “Master programmer, consider this [Perl one-liner]. Implemented in pure C, would it not span ten thousand lines?”
The programmer muttered through his beard, contemplating what Nubi had written. Finally he agreed that it was so.
“And how many hours would you require to implement and debug that C program?” asked Nubi.
“Many,” admitted the visiting programmer. “But only a fool would spend the time to do that when so many more worthy tasks await him.”
“And who better understands the Unix-nature?” Master Foo asked. “Is it he who writes the ten thousand lines, or he who, perceiving the emptiness of the task, gains merit by not coding?”
Upon hearing this, the programmer was enlightened.
Adapted from: http://www.catb.org/~esr/writings/unix-koans/ten-thousand.ht...
I rewrote it in a couple of hours in 393 lines of perl and 6 OS installed (so stable for the life of the OS) libraries.
>Now you should be very careful that the outer bracket type is a round one, that is, a parenthesis. That's because you're assigning to an @array, so you need parentheses. If you wanted there not to be an @AoA, but rather just a reference to it, you could do something more like...
Still pretty tricky!
After reading Programming Perl from cover-to-cover, I suddenly felt like I had superpowers. I probably would have felt the same if I'd taken the time to learn bash scripting, but that's not how it worked out.
I tried to automate everything in my life with Perl. I'd walk to class in the morning and a script had already scraped and downloaded my lecture materials and sent a print job for me. It emailed me if it couldn't find notes for one of my scheduled classes. Others would manage my assignment submissions. I wrote scrapers to detect what events on campus might offer free food.
This was the first time I felt like I could write programs that actually had the power to improve my life. It changed my perspective on what I had the power to do as a programmer. I'll always be grateful to Perl for that.
I also learned how to quickly develop a thick programmer skin by asking questions on comp.lang.perl, but that's another topic altogether.
I missed the Perl boat by a few years, riding the Python2/Java wave at Uni instead, but I did encounter it a few times in legacy (read: 5-years old) code and could only marvel at its string manipulation capabilities.
I like to think of it as the extended bash that I never took the time to learn
Is there something about gaining scripting power that requires such poor usability? I resisted it for a while, but I just write all my workaday scripts in Python now, and I'm honestly a little baffled as to why I stuck with bash for so long.
It is possible to write lucid, maintainable Perl, something I have occasionally managed. It is easy to write ugly, unmaintainable Perl, and that I have too often done. If one looks at the Perl found on CPAN or in various good books on Perl, you can see how to do the former.
Some of them are not even paint:
https://www.mcmillen.dev/sigbovik/splatters/646f93a2d62ad819...
(You would of course want a “proper” scientific study to be more rigorous about these things.)
Edit: Also, the programs in the article appear to be mostly special characters?
I assume you mean the 4 listed in the "Errata" section?
The full list is here: https://www.mcmillen.dev/sigbovik/splatters.html
But if someone said "95% of paint splatters OCR as Welsh words" and it turned out that's because the OCR is looking for words, it doesn't say much interesting about Welsh like the title implies.
> Edit: Also, the programs in the article appear to be mostly special characters?
Look at the whole gallery though. Most of the OCR results are just a string of letters or a number. Those being valid is very mundane.
Edit: Actually if you want to make the analogy super precise, you could say "oh yeah those strings of letters are just proper nouns, which makes them valid Welsh, and of course numbers are valid Welsh"
I think it depends on the claim, right? IMO, a more analogous statement might be "95% of paint splatters OCR as grammatically-correct Welsh sentences." Unless the software is using something like GPT-3, that finding would be notable regardless of software tuning. Even if it tells us more about Welsh grammar than it does paint splatters.
> Look at the whole gallery though. Most of the OCR results are just a string of letters or a number. Those being valid is very mundane.
Yes, thanks, my mistake!
Well, I'd argue that "potato" is a grammatically correct sentence in most languages. (Imagine someone asked you what kind of soup you're holding.) And so is "7" and so is "Fififz".
And that's literally what's happening with Perl with the vast majority of these. It's treating a single word/number as a statement.
If you want to say something notable about grammar, you need to be generating sentences that are longer than a single noun.
combined with -T, it's given me a lot of peace of mind, which I miss when using PHP.
I'm seriously contemplating writing my own SetConst and GetConst functions. (Lately I've been finding a lot of sense in immutable variables.)
use const name => value;
Are consts not allowed to be set at runtime?
use constant name => value;
use constant true => (1==1);
use constant false=> (1==0);
Try Hack? It's not too bad and the typing can be added piecemeal.
Nobody is going to bother with your language when it's full out there of solutions as good as yours that people can install in one second.
I suspect this dates back to the Perl4 -> Perl5 transition, when they removed the requirement to call a subroutine using an ampersand, like &somesub could just be somesub.
Unlike, say Python, Perl doesn't require subroutines to be defined before they are called.
For reasons I don't fully understand, barewords preceded by '-' are still allowed, even with "use strict". Also allowed as hash keys like $foo{bar}=1 or (foo => "bar").
I believe the primary motivation for this is so you can pass argument lists to subs in a key => value format without having to quote the key. For example:
someSub( -opt1 => $opt1val, -opt2 => $opt2val );
> The "=>" operator (sometimes pronounced "fat comma") is a synonym for the comma except that it causes a word on its left to be interpreted as a string if it begins with a letter or underscore and is composed only of letters, digits and underscores. This includes operands that might otherwise be interpreted as operators, constants, single number v-strings or function calls. If in doubt about this behavior, the left operand can be quoted explicitly.
> Otherwise, the "=>" operator behaves exactly as the comma operator or list argument separator, according to context.
This - the fact, rather than any justification - is documented in `perldoc perlop`:
Unary "-" performs arithmetic negation if the operand is numeric, including any string that looks like a number. If the operand is an identifier, a string consisting of a minus sign concatenated with the identifier is returned. Otherwise, if the string starts with a plus or minus, a string starting with the opposite sign is returned. One effect of these rules is that -bareword is equivalent to the string "-bareword".
Typically Perlish, that means you can do stuff like the following because, I dunno, reasons?
$ perl -E ' use strict; use warnings; say - -foo' # or -(-foo)
+foo
Quoting `perldoc perlop` again:
The "=>" operator (sometimes pronounced "fat comma") is a synonym for the comma except that it causes a word on its left to be interpreted as a string if it begins with a letter or underscore and is composed only of letters, digits and underscores. This includes operands that might otherwise be interpreted as operators, constants, single number v-strings or function calls. If in doubt about this behavior, the left operand can be quoted explicitly.
Oldest one is 6 now, so not sure what that will be like going into their teen years. I feel like that's when it will be harder to get them offline but we will see
I think the problem is that the original tweet seems to think that "coding" is only a vocational skill, and it's definitely not. When you're that age, coding is just another form of expression and exploration. It can make something useful, but it's just puzzles and fingerpaint.
Anyways, I agree. It's another form of expression. Just like Math or Music.
But I do also agree, in part with her point, that kids should explore possibilities and be kids instead of having to worry about their vocation so young and being put on a track to become something.
Yeah that's a really tough line to track. At some point they're going to develop a sense of purpose and I have no idea when or where or what or how I can nurture it.
ANXIETY
"That lady" is Adrienne Porter Felt, who started as an engineer working on Google Chrome security a decade or so ago, and is now a Director of Engineering who manages a sizable chunk of the Chrome team.
I never worked with her directly, but she has a reputation as a good manager and I'm pretty sure she knows how programmers work.
If that's what all the kids are doing then it seems that's what a kids activity has become.
I used to play with neighborhood kids quite a bit when I was younger. Now my nephew plays Fortnite almost every day with 5ish different friends. He gets way more experience socializing than I ever did.
Ultimately (I think) kids need to learn to socialize with their peers. That means they need to participate with their peers doing what their peers are doing.
Ubiquitous socialising can be good but it cuts both ways, especially in an environment where you have no plausible deniability. In my opinion, people should be free to have their own thoughts, potentially get blasted for them, and move on with their lives the next day with minimal consequences. Certainly there should be no consequences coming from people that were not actually there, but just heard about it from some social media platform.
Granted if a kid were to just upload fun videos... have comments off ... not interact with the community, then maybe they'd be ok? But I think all of that is not super likely as my son's friends want to be youtube stars.
When I was 10 or so, I uploaded some dumb little skits with my friends to YouTube. The videos accumulated about 2k views total, and are all gone now, and no one will ever recognize me from them.
Poeple in general are extremely bad at generalizing situations to match their own experiences. People get old, see younger people playing differently (new toy, new technology, new societal framework), and fail to notice that the new kind of play is still mostly the same module/class of play they had, just with a different coat of paint.
E.G. In the 80s' we used to play with k7 recorders, recording singing, pretend news casts and comedy shows. Now kids are doing youtube and tiktok, and that is exactly the same module of play, it is not some brand new thing.
Having said that, I'd like to point out that, although for the kid it is basically the same kind of play, the fact that is more public (k7 had a limited reach, just neighborhood friends, but the internet has unlimited reach) means bigger stakes, so parents should be more vigilant, it is even more crucial now that parents get involved in the kids activities to keep an eye out.
I saw someone mention summer camp, some kind of structured mechanism to do it would help a lot i think.
Programming is great for self expression, exploring new ideas, amusing people. These are great goals for kids.
If you teach a kid to program, probably they're gonna put it to use in Minecraft or Roblox, not by writing accounting software. That's good! That's no more professional software development than kids building with Legos are doing engineering.
For us, it started at me - I had to make some big changes. When she was very little, I caught myself pushing her on a swing with my nose in my phone. Once I caught myself doing that, I broke the habit - now when we’re spending quality time together, my phone doesn’t exist.
This gives us space to do things outside, go on adventures and use our imaginations.
Devices are a huge part of her world so at other times, we either share one device and I let her quarterback or we spend an hour in bed on our own devices. I’m open about what I do on my device and she’s open about what she does on hers. In some cases, this creates an opportunity for me to teach and model. In other cases, it gives her an opportunity to teach me.
The last part has been especially important. As my girl has taught me things about her world, I’ve learned that some of my ideas about kids being kids are based on my own prejudice. I’ve caught myself expressing an overly rosy view of ‘when I was a kid’ and have tried to project that world upon her. Thankfully, my kid and I have the kind of relationship where she’s more than comfortable to tell me that I’m wrong.
93% of Paint Splatters Are Valid Perl Programs - https://news.ycombinator.com/item?id=19576425 - April 2019 (132 comments)
93% of Paint Splatters Are Valid Perl Programs - https://news.ycombinator.com/item?id=20834351 - Aug 2019 (7 comments)
Also this, recently:
The four noisy horsemen of Perl hate - https://news.ycombinator.com/item?id=27902296 - July 2021 (51 comments)
But any language that detects random string as a valid program is under suspicion of allowing more bugs than valid programs.
It may be easier to write in a language that isn't as strict, but it's going to take more effort to produce a *correct* result.
By the way, the first Raku Conference is happening in a few weeks, Aug. 6–8 https://conf.raku.org
on edit: actually how many country fences are valid brainfuck.
Spoken like someone possessing the supreme programmer's virtue of laziness.
i.e. a fun April 1st paper.
What I liked about Perl when I started using it was dictionaries, lists, regexes and CPAN. Coming from a statically typed language like C or Pascal it was freedom. Writing modules was weird, objects weird - so that meant messy code. And when they did implement something, it was always a sophisticated flavor of some rarely used programming concepts.