Skip: A programming language to skip the things you have already computed

by bpierre • 4 years ago

195 115 4 years ago

klohto• 4 years ago

Is it used anywhere? Github seams dead and the main contributor have left Facebook to start SkipLabs with DB related product SQLive.

julienv5• 4 years ago

That's what SQLive is written in. Hence the name SkipLabs!

da39a3ee• 4 years ago

anyone know what it was about skip that made it attractive to implement SQLive, put anything else in? What’s its pitch?

macintux• 4 years ago

Looks very interesting, but the last blog entry was over three years ago, and there are no mailing lists/IRC/Slack/any community forums to indicate signs of life. A shame.

sbussard• 4 years ago

Footer is fb not meta? Legit affiliation?

Copyright © 2022 Facebook Inc.

detaro• 4 years ago

automatically generated to current year and the repo hasn't been updated in 3 years.

jamesfinlayson• 4 years ago

Site doesn't load for me right now but looks like it: https://hub.packtpub.com/facebook-releases-skiplang-a-genera... I remember a colleague seeing this presented at a conference a few years ago but that was the last I'd heard of it.

tekknolagi• 4 years ago

It used to be a FB project, but no longer.

sidpatil• 4 years ago

http://web.archive.org/web/20220218003133/http://skiplang.co...

_zooted• 4 years ago

Homepage is skipping loading.

oshiar53-02• 4 years ago

To be fair, it has computed the page at least once at some point...

arrow7000• 4 years ago

Came here to say this

mcluck• 4 years ago

Looks interesting but here's what I don't understand. So many programming languages do not show example code on the homepage. That's like if Mazda didn't show cars on their homepage. Sure, you won't really know what it's like to drive it without getting your hands dirty but at least let me see the trim.

vjeux• 4 years ago

Great feedback! This is a big oversight.

spicybright2• 4 years ago

Do you have any example code that highlights the caching idea described on the main page?

floodle• 4 years ago

Even worse, the "tutorial" is a challenge of sorts, where you have to guess at the right answer given hints.

Mathnerd314• 4 years ago

The "documentation" is also tutorial-style.

notreallyserio• 4 years ago

Sure, that's a problem, but once you've figured out the answers you can just skip them next time.

bidirectional• 4 years ago

How is an interactive tutorial bad?

kazinator3• 4 years ago

In the spirit of testing: Mazda doesn't show cars on their home page. Here is mazda.com:

   Forbidden

   You don't have permission to access / on this server.

You have to add www. to mazda.com before you can see any damn cars.

And it's just one car I see; some fat SUV that looks like it's about to give birth to a pair of sporty twins.

LeoPanthera2• 4 years ago

In the USA, Mazda is at mazdausa.com, not mazda.com. There are eight cars on the front page.

gibolt• 4 years ago

Must be an NFT that you don't own

coldtea3• 4 years ago

You have to add www. to mazda.com before you can see any damn cars.

Who said mazda.com is the homepage, and not www.mazda.com?

0xTJ• 4 years ago

I don't get your point. Plenty of sites have `www` (though most will forward to it from the non-`www` version. A company's homepage isn't defined as their first-level subdomain. On top of that, many companies have different homepages. Some are for their products, some are for their company.

jaredsohn• 4 years ago

nissan.com doesn't show any cars. :)

ewittern• 4 years ago

In that regard, they are taking the language‘s name way too serious.

viktorcode• 4 years ago

Just click "learn more" button and you'll see step by step examples of the language features.

A quick take on the front page can't possibly explain much about a language, maybe just a sample of syntax? I don't know why it would've been useful.

xigoi• 4 years ago

Look at Nim's front page: https://nim-lang.org/ It features several code samples showcasing many important language features.

alpaca128• 4 years ago

I can only speak for myself, but this kind of abstract listing of features is absolutely meaningless to me. Why would I want to use it instead of others, how do the unique features make a difference at solving a frequent problem? After reading the page I know Skip does something with caching, but I still have no clue what difference it actually makes in a program.

In my opinion a programming language landing page without code example is like a game's Steam store page without screenshots - it's cool to know it's a "fast-paced racing game" but that could apply to everything between Tux Racer and Gran Turismo.

TypeScript's website shows right at the top in a few lines how it makes an easy mistake in JS impossible. Zig's site has a nice example that gives you an idea of some added features, how it differs from C but still feels similar. Rust's website had an editable and runnable code snippet showing iterators and pattern matching (although for some reason they later replaced it with the phrase "blazingly fast" in the redesign).

Mawr• 4 years ago

Just click "learn more" button and you'll see step by step examples of the language features.

I got bored after about 5 pages, everything presented there is the same as in any other language. I want to see what differentiates this language from others, not what's the same.

That is what a landing page of any new language must accomplish - sell the reader on the power of your unique features as fast as possible.

mshockwave• 4 years ago

what's worse, even if you click on "more info" it goes to a tutorial that teaches you from zero. But I'm more eager to know killer features in this language.

pharmakom• 4 years ago

Maybe a feature list tells you more than a code list might show? Often syntax is the less interesting thing.

toomanydoubts3• 4 years ago

I'm not sure. If we're talking about turing complete languages, aside from performace, you can get all kinds of features in all kinds of languages. It's actually a matter of how expressive, safe, idiomatic and well designed they are for everyday usage.

When discussing a language which is defined by caching values, we sure do want to see how that looks like in practice.

julienv5• 4 years ago

It's still very much used at SkipLabs. In fact, it's the language that SQLive is written in: http://sqlive.io

Unfortunately, we use a proprietary fork of SKIP, so that's why the OSS project looks dead.

We might open-source our version some day, but no plans for now.

jerye• 4 years ago

This sounds interesting... Though it doesn't seem like that domain is working.

julienv5• 4 years ago

Fixed! thanks!

interpenetrate3• 4 years ago

If done right, SQLive is the sort of thing that could entirely change the way many applications are written. Solving the problem of real time collaboration at the database level has been a dream of mine for years. Unfortunately I'm doubtful that existing RDBMSs will be able to integrate such a thing, seeing how shoddy implementations of materialized views have been.

I really hope you are successful in tackling this problem and that you can share your work.

julienv5• 4 years ago

Thanks! You can try the version in the site for testing if you want! If you want to build something, let us know!

FridgeSeal• 4 years ago

Check out the work Materialize is doing as well - similar vein.

netfortius• 4 years ago

... "proprietary fork" and "might open-source"... let us know when you change your stand

vrotaru• 4 years ago

Well Julien Verlaguet is the original author of SkipLang, so "proprietary fork" does not deserver the square quotes, IMO

unnouinceput• 4 years ago

It's so dead that even the Facebook copyright at the bottom was not updated

dapids2• 4 years ago

It literally says 2022

phonethrowaway• 4 years ago

it doesn't say meta, duh

brundolf4• 4 years ago

It mentions on the home page:

Caching with Reactive Invalidation

But a Cmd+F for "cach" in the docs doesn't turn anything up. Anybody know where it goes into more detail on that?

snorkel• 4 years ago

http://skiplang.com/blog/2017/01/04/how-memoization-works.ht...

brundolf4• 4 years ago

Cool, thanks

rco87862• 4 years ago

A whole homepage about a programming language, and nowhere on the page does it show the programming language. :-/

rowls66• 4 years ago

Why is seeing the syntax of a programming language so interesting? I think that syntax is one of the least interesting aspects of a programming language.

gigatexal• 4 years ago

It’s actually the opposite: one will be writing that syntax when using the language if it’s a chore to write then nobody will want to use the language. Also it’s easier to understand what a language is doing looking at code than a technical document.

spicybright2• 4 years ago

I can't find any code that highlights the main feature of the language. Only example programs like hello world, arithmetic, etc. which are fairly useless.

stirfish• 4 years ago

It's nice to be able to see what a new language is able to express easily. Have you ever browsed Rosetta Code? I find syntax interesting because it shapes how a language is used.

https://rosettacode.org/wiki/Rosetta_Code

postalrat• 4 years ago

Syntax interesting I is a aspects think so the language. Is language the of why seeing programming of of programming a syntax one interesting? Least that

zem• 4 years ago

it's the user interface of a language. you could as well say that screenshots are the least interesting thing about a gui app and what it can do is more important, but people want to see the screenshots anyway!

hota_mazi• 4 years ago

This is like saying that you prefer to read the description of a new car instead of seeing the car itself.

davidgtl• 4 years ago

It's not about syntax, it's about quickly seeing language choices and verbosity of various constructs. There is a lot more information in a good example rather than a front page feature list.

moron4hire2• 4 years ago

Well, one particular thing we'd see if there were a meaningful example on the home page is that Skip doesn't use a keyword to declare new symbols, opting instead declaration being conflated with first-assignment.

http://skiplang.com/docs/functions.html#variables

Misspelled variable assignments will silently succeed, rather than error because the symbol doesn't exist. I think that's a massive mistake and it speaks to a relative lack of large project development on behalf of the Skip designers.

So, I guess, in a backwards way, it's better for them to hide their syntax away from the home page, because I wouldn't have spent as much time on the site if I had seen that right away. "Time on site" metric improved, but to no meaningful result.

Also, in the year 2022, the site does not have a TLS certificate.

rco87862• 4 years ago

It's like viewing the floorplan of a house vs a written description of the floorplan. The latter is verbose, long, wordy and ultimately cannot paint a picture as accurate as just showing the floorplan.

oshiar53-02• 4 years ago

I assume this uses thunks not unlike some lazily-evaluated languages?

dang• 4 years ago

Skip – A programming language to skip the things you have already computed - https://news.ycombinator.com/item?id=18077612 - Sept 2018 (100 comments)

hayley-patton2• 4 years ago

Thanks to Skip's tracking of side effects the garbage collector only has to scan memory reachable from the root of a computation

But any typical copying GC only has to scan memory reachable from the root(s) of a mutator. I suspect there are more interesting optimisations, but I am unable to discern them from this statement.

julienv5• 4 years ago

Well, imagine your type-system was able to tell you exactly what roots are mutable in what scope. Now you can run a garbage collection on a single scope. That's not how it works, but in pseudo code:

collect { // Some code return ...; }

Well, if you don't have a type-system that tells you what is mutable, you have to either scan all the mutable roots, or maintain a write-barrier to know what could have captured data in the scope.

The problem with write-barriers, which is the go-to solution for generational GCs. Is that it is a conservative approach. Meaning, it will promote garbage in the cases where the pointer that lived outside of the scope is dead.

Let's take an example, again, in pseudo code:

myObject = { field1: []};

collect { myObject.field1 = [A, B, C]; myObject = null; }

If you use a write-barrier, what is going to happen is that [A, B, C] is going to be promoted, because the barrier is going to track the object myObject, and it doesn't realize that it's dead.

However, instead, imagine you have a type-system that tells you that in that scope, the only thing that can be mutated is myObject!

Well, now you can run that in a loop, without accumulating garbage!

Makes sense?

hayley-patton2• 4 years ago

So, myObject and the list are allocated in different generations/scopes/etc, and we'd typically trip the write barrier, if we kept the program as is. Right. I don't see how it follows after, that we can avoid the barrier based on just mutability information. It is true that we can only "capture" data with the involvement of mutable objects, but I am struggling to figure out what more we can derive from mutability information. It feels like I'm missing some inference step between "the only thing that can be mutated is myObject" and "now you can run that in a loop, without accumulating garbage" and I can't think of one.

I would be more tempted as a compiler writer to delete all the code, since nothing is ever read outside the loop; but such analysis again doesn't depend on mutability type information per se, rather on what mutations occur in the program.

fennecs• 4 years ago

It’s not like memoisation and thunks are new features. What has this language added in this space I can’t do in say Haskell?

written by a bunch of html/js developers

Ah so because they have rediscovered something it’s meant to be revolutionary?

KnobbleMcKnees• 4 years ago

My first thought was wondering why this couldn't be partially implemented (i.e. the important bits like memoisation of idempotent, state-bound functions or contexts) in any other programming language.

Then I wondered if this was already implemented for some functional languages. Then I read your comment.

It was quite the ride.

syrusakbary• 4 years ago

It seems the last blogpost was from Aug 2018 [1] and last commit in the repo from June 2021 [2], so it doesn't look very maintained at the moment of writing this comment (Feb 2022)

[1] http://skiplang.com/blog

[2] https://github.com/skiplang/skip

farzher• 4 years ago

adopting a new language is such a huge task. doesn't seem like a big enough feature for a whole entire new language. glad people are trying new things though

xialvjun• 4 years ago

Viewing https://news.ycombinator.com/item?id=18077612 , Skip seems to be an interesting and promising programming language. I hope it will continue to be developed.

xwdv• 4 years ago

Imagine if we had a shared distributed database and could even skip any computation that has ever been computed.

continuational• 4 years ago

But then you have to wonder - what is the storage requirements for storing the results of every computation, ever?

Salgat• 4 years ago

Reminds me how you can do the same thing with a simple "@cache" at the top of a python function.

brundolf4• 4 years ago

That doesn't work if one of the arguments gets mutated

llIIllIIllIIl• 4 years ago

Different arguments would provide a different result, wouldn’t it?

bbarn• 4 years ago

"In Skip the = operator introduces a new variable, but unlike other languages the = operator alone can not modify an existing variable."

Using != to assign a value feels really strange to me. I know some languages use := but in those languages typically it's because = is an equality operator. I'm sure there are some exceptions but != just so universally means "not equal" to me.

sgillen• 4 years ago

I too would prefer := or even =!

asddubs• 4 years ago

or <= which vhdl does (it differentiates between variables and signals so it uses both := and <=)

janjones2• 4 years ago

But it's not just `!=`, it is `!variable = value`, see [1]. That does not look like "not equal" operator.

[1] http://skiplang.com/docs/lvalues.html

shrimpx2• 4 years ago

Although !x mean "not x" pretty universally.

mpweiher• 4 years ago

Objective-S uses |= for one-way dataflow constraints, which you could also call "permanent assignment".

http://objective.st/

fjfaase• 4 years ago

It does make sense if you realize that it is also used for fields and that a.!b.c = d is different than a.b.!c = d. The first makes a copy of the object at a.b before assigning a new value to its field c.

pygy_3• 4 years ago

This looks fascinating, I wish I had found it sooner.

Am I correct that the cache invalidation logic is akin to pull streams, unlike computations in projects like S.js/Surplus, Solid.js or Sinuous, that are push-based?

shrimpx2• 4 years ago

I guess the vision is to have SQLive hooked up to Skip hooked up to React. Then your website is a pure function of your database, and clicks in the UI just mutate the DB, and all the views updates seamlessly.

otikik• 4 years ago

This part caught my eye:

http://skiplang.com/docs/lvalues.html

I think some of the phrasing there ("the variable x is going to be updated to point to a new copy of the object pointed to by x where field1 has the value 0") would really benefit from some graphics. Nothing fancy, something similar to what the Rust book does for ownership would help.

https://doc.rust-lang.org/book/ch04-01-what-is-ownership.htm...

amelius• 4 years ago

For how long are values cached? And how is garbage collection of the cache handled?

pygy_3• 4 years ago

For context, this is the brainchild of Julien Verlaguet, one of the creators of the Hack language used at FB.

Skip was meant to gradually replace Hack, but the FB codebase turned out to be too tangled to gradually convert to the reactive paradigm, and a big rewrite was not possible, so FB pulled the plug, then Julien kept on pushing the project independently.

comboy• 4 years ago

It would be nice to put right there on top info what's the difference between it and using already popular functional language with memoization.

Guessing if something has side effects instead of making it explicit makes me already want to run away when I think about debugging.

noduerme• 4 years ago

Whoa so - I had to kinda dig to get to the Lvalues section of the docs, but ya really bury the lede there! I love the concept of x.!o = 3 and !x.o = {copy of x, with o=whatever}. I didn't read far enough to understand whether this makes deep clones with sub-objects. (I'm guessing...not?) Still, that itself is very innovative.

thom• 4 years ago

I've been tempted to learn Rust just for Differential Dataflow, less sure I'd learn something even more esoteric for some similar functionality, but a very interesting space to see progress in.

andrewfromx• 4 years ago

I like the name. In 4th grade I presented the argument to my PE teacher that skipping was better exercise than running and he let me skip around the track while everyone else ran.

pshirshov• 4 years ago

A quick question: does it support HKTs and implicits/typeclasses? Is there a decent macro system?

rvcdbn• 4 years ago

Interested to learn more about how memory management works - any details on this?

Skip: A programming language to skip the things you have already computed

Related Stories