No more "minimal early Lisps", pulleezz

If there's one thing that irritates me about the pg-led Lisp renaissance it's the fascination with the Ur-Lisp. Rewriting that Ur-Lisp has become a pastime. Here are just some of the reasons I think that's dumb:

Lisp is a bad model if you want some kind of axioms of computing

We already have the lambda calculus for that, and you can build a real language with that (see Haskell). If you want to write some minimal thingy in C, consider writing a Haskell, not a Lisp.

That you can write a Lisp evaluator in Lisp was interesting in 1959 (and maybe 1960), and let's face it, that Ur-Lisp was one broken language.

Car, cdr, wtf?

Contents of the address part of register, contents of the decrement part of register, what the fuck? What are they doing in a language, even a toy, written in 2010? Please use data structures.

Where are the closures?

If you're looking for a challenge, as opposed to redoing something that's been done ad meam nauseam for half a century fachrissakes, try to find a minimal and explainable way to do closures. Bonus points for efficient flat environments. Languages without closures are so 1959.

Where are the macros?

A Lisp without macros is meaningless.

Where's the control flow?

A Lisp without some kind of continuations and condition system is useless.

Where are the semantics?

If you read the RnRS attentively, you'll see that Lisp has evolved a deep and subtle set of semantic concepts, none of which feature in the Ur-Lisp.

What's the purpose?

Surely you're not learning much by repeating JMC's flawed Ur-Lisp from 1959. If you want to learn something, implement a language with closures and macros. If you want to learn more, make it a compiler. If you want to blow your head, implement hygienic macros or a higher-order module system or a static type checker. That's today's standard.

Look, Lisp is such a great language, but if anything we have to push it harder, not continuously go back to 1959.

Update: More fully-featured, modern Lisps, pulleezz

Dalvik DEX

Good binary formats rock. Reading the ELF spec was quite eye-opening for me.

Dalvik's DEX format is another nice one.

Interestingly, Dalvik is the brainchild of Dan Bornstein, who was at Kaleida Labs (RIP), where he worked on ScriptX, one of the many avant-garde codes that fell victim to the WWW-induced ice age of GUI innovation.

The 2010 Linux Storage and Filesystem Summit

As usual, Jonathan Corbet does an admirable job of informing us on happenings in kernel land in his summaries of The 2010 Linux Storage and Filesystem Summit, day 1 and day 2.

On testing:

It was suggested that the real test should be "put the new code on the Google cluster and see if the Internet breaks."

On Google:

There are two fundamental types of workload at Google. "Shared" workloads work like classic mainframe batch jobs, contending for resources while the system tries to isolate them from each other. "Dedicated workloads" are the ones which actually make money for Google - indexing, searching, and such - and are very sensitive to performance degradation. ...

The workloads exhibit a lot of big, sequential writes and smaller, random reads. Disk I/O latencies matter a lot for dedicated workloads; 15ms latencies can cause phone calls to the development group.

Meta: Blogging is difficult

The Axis of Eval is my second blog, after the venerable plans within plans within plans. (Yeah, there were others.)

Both of these blogs have gone through the same stages:

1. A few days of writing in solitude with a couple of friends.
2. Chris Neukirchen mentions the blog on Anarchaia or Trivium.
3. Al3x twitters about it.
4. Attack of the unwashed HN masses (just kidding).

It's hard to go from 1 to N. The rush of thousands of readers is great, but it's confusing (being able to write to thousands of peers in real-time isn't really accounted for by our DNA and social structure). For me, it's kinda hard to write when I know that a post I write will show up in the inboxes of many people most of which I don't know.

I'd like to keep a certain entertaining and informative niveau, and in the best case, I'd also like to improve it. In the early days, aggressive diss-posts and funny flames flow freely, because the audience is small and trusted. And those posts are entertaining. But in a more public setting, I have a bit of a bad feeling writing them, because I feel they may harm people I write about, when all I'm intending is to vent about some ideas I think are bad or ridiculous, or would like to tell a stupid joke.

So, what I want to say is that there are some difficulties to blogging that are seldom written about, and I'm still trying to figure out the boundaries of this strange new thingy, and where to draw the line between fact and fiction in blogging.

--Manuel

No Paranoia Rule

Good rules are few and far between in the programming scene.

The first time I heard of what I now call the No Paranoia Rule was in the following comment by Luke Gorrie on LtU:

"Oh my, what if Luke installed an exception handler that ROT13 encoded every string on the heap, then how would Jane debug her programs?"

I hear that early on, people opposed subroutines for similar reasons. And of course, macros are frequently criticised for their potential of wreaking havoc.

But that's where the No Paranoia Rule comes into play. Stop being paranoid, and don't discount language features for their potentially devastating effect.

As Luke further states,

This [being paranoid] is not the way to illumination.

(Clemens A. Szyperski coined the term No Paranoia Rule.)

You know you're reading LtU when...

I believe the preferable solution is typeclasses as record types, instances as records, scoped implicit parameters for propagating them around, associated types as record members, and a backtracking logic solver for instantiation. But that just solves the immediate problem; on top of that I'd like a dependent type system with staging to enable type dependency on values (e.g. arrays-with-length), and a higher-order logic solver so simple logic-programming idioms like let zip(as,bs)=f(x) in ... can be expressed and translated via instantiation into efficient runtime code. And a total (non-partial) function subset of the language for reasoning with proofs as programs. — Tim Sweeney

Three Principles of Lisp

I've been thinking about what it is that still sets Lisp apart from all other dynamic languages. I've come to three core principles, that define what Lisp means for me.

Liberal use of syntactic abstraction

Lispers don't fret over when to build a domain-specific language (DSL) or not. When it makes sense, they build one, when it doesn't, they don't. Thanks to the trivial syntax, and the long experience with tools like pattern matching and quasiquotation, DSLs in Lisp are created rapidly.

In other languages, the creation of new language constructs or new languages requires countless hours of busy work. In Lisp, DSLs can be ready to use after a couple of lines entered at the REPL. In Lisp, creating DSLs is a no-brainer, and I think that's one of the cornerstones of Lisp.

Lisp isn't doctrinaire about creating DSLs - in fact it is standard advice to only use macros when plain functions won't do it anymore. But because creating DSLs in Lisp so easy, they are created whenever they make sense. Which, as it turns out, is a lot of times. I would go so far as to say that the liberal use syntactic abstraction makes Lisp a boilerplate-free language, but that's another topic.

There's no such thing as a constant mentality

Lisp thrives on interactivity. Newer languages like Haskell are encroaching on the natural habitat of Lisp (and surpass it some areas), but none of them can match Lisp when it comes to no holds barred, nitty gritty interactivity. Even defconstant isn't necessarily constant, and for interactive systems (Emacs!) that's exactly what you want. In an interactive system, you can't tolerate constancy - this would be like using Lego blocks that are glued together.

I think a fundamental insight for understanding Lisp's superiority in the interactive domain is that Lisp is an asynchronous language, in a very peculiar sense: any Lisp expression may be entered at any time. In many cases, you could cut up a Lisp source file at top-level expression boundaries, rearrange them, and the resulting source file would still have exactly the same effects as the original file when loaded into a Lisp. (Just as one example, in Common Lisp it's possible to create subclasses of classes that don't exist yet.)

This is simply a natural consequence of making the read-eval-print loop (REPL) the cornerstone of Lisp semantics. In C, the dominant paradigm is main(), in Lisp it's (repl). And the top-level is tricky, some say hopeless. In Lisp, it is fundamental to always be able to define a function FOO that calls an as-yet undefined other top-level function BAR - a fundamental idiom of interactivity:

(defun foo ()
  (bar))

Toolchain approach

Lispers don't need to create a new language to try out new language design ideas. Many times, new languages can be defined as macros. If they don't, and an interpreter or compiler is needed, it can still stand on the shoulders of Lisp. In the Lisp world, new languages are built buy combining large, battle-tested building blocks, and polishing or updating them when needed, instead of starting over from toothpicks and double-sided duct tape. A large Lisp like Common Lisp is like a toolchain of decades-old tools that have proven their worth, and have been codified in standards, folklore, and implementations.

Now of course, that's also true of other languages. But in other languages, starting a new language is a from-scratch affair. And on the long way of parsing, analyzing, and interpreting or compiling language constructs, the creator of the new language invariably introduces a lot of things that are different from other languages, so every new non-Lisp language is often fundamentally different from other languages, in subtle areas such as control flow or scoping, which requires decades of fixing.

In Lisp, new tools are tried out as separate functions, macros, DSLs, subsystems, or, in the extreme case, code walkers or complete new implementations. The rest of the language stays the same. Even a new Lisp implementation will often share a lot of the design space with other Lisps. And that's the reason Lisp had proper lexical scoping for decades, while it's just become a fixture in new languages, and that's also why Scheme now has hygienic macros and phase separation, and other languages will have them in decades - because in Lisp, all of these new constructs can be developed in the Lisp fabric, without rebooting the process every time. Which turns out as a nice metaphor:

Others reboot, Lisp keeps running.

So true

Great response to Ask HN: Do you use more than one programming language?

Not only do I use multiple languages professionally, I don't know some of them. – anamax

Concurrency's Shysters

Excellent article by Bryan Cantrill, Concurrency’s Shysters:

[C]oncurrency is still being used to instill panic in the uninformed. This time, it is chip-level multiprocessing (CMP) instead of SMP that promises to be the End of Days — and the shysters have taken a new guise in the form of transactional memory. The proponents of this new magic tonic are in some ways darker than their forebears: it is no longer enough to warn of Judgement Day — they must also conjure up notions of Original Sin to motivate their perverted salvation. “The heart of the problem is, perhaps, that no one really knows how to organize and maintain large systems that rely on locking” admonished Nir Shavit recently in CACM. (Which gives rise to the natural follow-up question: is the Solaris kernel not large, does it not rely on locking or do we not know how to organize and maintain it? Or is that we do not exist at all?)

I know much too little about the subject to have an informed opinion, but I think countering hype in all its forms it important.

Happenings in GCC-land

Ian Lance Taylor is working on split stacks. This will permit the stacks of gccgo goroutines, which are mapped to native threads, to start small and grow as needed. This will of course be usable from C as well, and should be welcome news for the millions-of-threads!!!1! faction.

Tom Tromey is working on making GCC an incremental compiler: GCC will run as a server and maintain a model of the user's program. When a translation unit is recompiled, GCC will re-compile the minimum necessary. One of the goals is to make GCC a backend for IDEs that do stuff like autocompletion based on the program model (Interview, paper from GCC Summit).

People from Intel and others are working on transactional memory. Sections of code can be marked as atomic, and their temporary changes will be saved in thread-local storage. The semantics will make it appear as if transactions were protected by a single global lock.

It's a great time to be writing a Lisp->C compiler! :)

Efficient Method Dispatch in PCL

Efficient Method Dispatch in PCL by Gregor J. Kiczales and Luis H. Rodriguez Jr. is an all-out awesome yet compact paper about generic functions in their portable CLOS implementation.

Some salient points:

They add a level of indirection between objects and their classes, called wrappers. An object points to a wrapper and that points to the class of the object. Inside each wrapper they store a random value for use by the memoization tables of generic functions.

Each generic function has a small memoization table that maps the hashed random value of a wrapper to the program counter (PC) of the corresponding method defined for the wrapper. The memoization tables use a simple random % memotablesize computation, and linear scanning after that.

How they handle class redefinition is swell: they simply set the wrapper's random value to zero, thereby invalidating the wrapper. They always leave the first entry of generic functions' memoization tables empty, and because an invalidated wrapper's random value will now hash to that empty entry, they only need to check for an invalid wrapper after a memoization miss occurs. Genius.

Generic functions run through a kind of state machine, depending on how they're used. (Similar to how polymorphic inline caches switch between mono-, poly-, and mega-morphic states, but with additional states tuned to CLOS semantics.) For example, if a generic function is only ever used with two different classes it can use a simple IF to test for these two classes. Here's another kicker: if a generic function's state machine detects that it's used as a slot accessor, it can directly store the slot offset instead of the method's PC in the memoization table. AWESOME.

Piet

I think the use of color in syntax is very interesting. I first heard of this idea in ColorForth.


Sorting in ColorForth.

Piet is awesome and takes this to a whole new level! Here are some sample programs:


Prints "Piet".


A random number generator.

Proof that parsing is evil

total    blank   lines w/   nb, nc    semi- preproc. file
lines    lines   comments   lines   colons  direct.
--------+--------+--------+--------+--------+--------+----
1339      182      242     1072      666       25 lua-5.1.4/src/lparser.c

As DJB sez, don't parse.

This is getting silly

Guess what's a goal for C# 5.0?

Tadaaaaaaaaaaaaaa....

Compiler as a service. Which is code for eval. Anders Hejlsberg shows it off in the great The Future of C# at 59:30.

Unbeknownst until just now to your correspondent, the JVM actually has a similar facility, but as Earl says, it's over-engineered to the brink of spontaneous self-combustion. (Love that quote.)

(I don't want to diss Java and the JVM too much, btw. I think it's a great platform for when you need to get something done. It's just not inspiring.)

CaaS is great on so many levels! Lispers will be able to claim another first on a vital technology. No future languages will be REPL-less.

My prediction for C# 6.0: Hygienic macros, quasisyntax, and phase separation. Seriously.

On Understanding Types, Data Abstraction, and Polymorphism

Cardelli and Wegner's On Understanding Types, Data Abstraction, and Polymorphism from 1985 is a treat. No wonder it has more than 2000 citations on Google Scholar.

It starts off with a simple functional language, and then adds universal and existential quantification. Universal quantification (∀) brings parametric polymorphism ("length returns the length of any list, whatever the type of its contents are"). Existential quantification (∃) lets you hide the details of a data type ("there exists some data type that implements a stack, but I'm not telling you more"). All of this is described in terms of a set-based description of types, which clears up a lot of issues.

It seems that this paper is the granddaddy of papers about polymorphism. It's expository style is unmatched.

P.S. I discovered this paper while skimming Adrian Moors' thesis Type Constructor Polymorphism for Scala: Theory and Practice. Its final chapter gives a great overview of where type systems are going.

What's phase separation and when do you want it? (Part 1)

One thing I've learned, now that I'm working on my third attempt at implementing an Acceptable Lisp™ (after an abandoned project years ago, and the JavaScript-based, also abandoned, but at least somewhat interesting and remotely complete CyberLisp) is that there's probably no way to understand Lisp, unless you implement one. (So I don't expect you to understand any of the following wisecracks, unless and until you've written your own Lisp, or spent some quality time thinking this whale of an issue through. It's like the masters of old said: you can only explain something the listener already knows. Or as Perlis said, you can't communicate complexity, only an awareness of it.)

My tip: if you want some Serious Insight™, write a compiler and follow the footsteps of modern Lisp compilation managers such as PLT Scheme's or XCVB. What's the difference between those and the rest of the pack?

Phase separation.

That probably doesn't ring a bell with you unless a) you've followed Scheme recently, or b) are writing your own Scheme, or c) are simply an all-around wizard. All of which are good for you!

Let me explain... Lisp, because it has macros that are written in a Turing-complete language (Lisp), is fundamentally a multi-phase language. We could also call it multi-time. Why? Because there's your ordinary runtime runtime, but before that happens, there's also a compile-time runtime. The right-hand sides of macro definitions run in the compile-time runtime. (I'll just call the runtime runtime runtime and the compile-time runtime compile-time from now on. (Yeah I had to do that thing with the three "runtime"s in a row there.))

This is another concept, like hygiene, that is only now (or just recently) becoming understood by the peoples. Where by the peoples I mean your correspondent. (The concept has been debated at length on the wonderful r6rs-discuss list (sorry, I'm too lazy to dig up the threads). It's been discussed almost more than case sensitivity, which should give you an indication of how novel and controversial the topic is.) This is deep stuff. At first. And later, too.

Compile-Time

One fundamental feature of Lisp is that you get to extend the compile-time with your own code, macros. (Now, Tom Lord would say that's totally the wrong way to do it, but hey, that's the way it's been done the last couple decades, so I'm covering that.)

There's a bit of sneakiness in how macro calls look like function calls in Lisp when they're a totally different thing, and there are languages that make a syntactic difference between the two, but Lisp doesn't, and for the better.

The first fundamental thing to understand is that macros don't exist at runtime.

When you do a (defmacro foo () ...) in a Lisp with phase separation, there will be no object called FOO in the runtime. But where is it? It's in the compile-time.

Yeaaah.

I know, it's a seriously weird concept, but it's absolutely fundamental to grasp this: macros are compile-time entities, they don't exist at runtime.

So when and where is this compile-time? The when is easy: when you compile the file. (Of course it's not that easy, but let's leave it at that for now.)

The where is more tricky. A Lisp source file contains two kinds of stuff that are fundamentally different: macro definitions exist at compile-time, while the other stuff such as global variables and functions (obviously) exist at runtime. (That a source file describes multiple runtimes, the runtime runtime and the compile-time runtime, is one of those subtle, small things that add up to make Lisp the powerhouse that it is.)

That's where XCVB-and-SBCL (and PLT Scheme, and probably Chez, and probably others) come in with a shattering device: the CFASL.

CFASLs

Let's take a look at non-C, plain FASLs first.

A FASL (fast-load file) is the output that a Lisp compiler produces from a .lisp input file. If your compiler compiles to bytecode, the FASL contains bytecode. If your compiler compiles to machine code, the FASL contains machine code. If your compiler compiles to JavaScript, the FASL contains JavaScript. FASLs are fundamentally implementation-specific, but always contain some form of executable code which can be run on a (real or virtual) machine.

What's in a FASL? Say we have a Lisp file containing the following expressions:

(defun foo ()
  (print "hello world!"))
(foo)

Then, the FASL will contain executable code that a) defines a new function called FOO, and b) calls that function, printing "hello world!" to the console. Simple.

Now what's in the FASL for the following Lisp file? Hint: not so simple.

(defmacro bar ()
  `(print "hello meta-world!"))
(bar)

The FASL for that file will (gasp!) only contain executable code for (print "hello meta-world!"). The reason is that we're defining the macro BAR, and the (macro) call to BAR in the third line gets expanded at compile-time to the PRINT expression.

And that's where CFASLs come in. (The C stands for compiler, or compile-time.) The CFASL of the above Lisp file will contain (defmacro bar () `(print "hello meta-world!")).

See? The FASL contains the runtime expressions of the Lisp file, and the CFASL contains the compile-time expressions of the Lisp file, and never the twain shall meet.

Why, oh why?

Now, this is where it gets tricky. >:->

We have to distinguish the two ways in which an Acceptable Lisp™ needs to work: interactive REPL mode, and static file-compilation mode.

Let's start with the file-compilation mode, because it's simpler, and fachrissakes, let's not discuss module dependencies for now.

In file-compilation mode, the Lisp's job is dead simple: slurp in the source file, process it, and produce a FASL from the runtime expressions and a CFASL file from the compile-time expressions.

If we want to execute the resulting program, all we have to do is load the FASL. Loading means simply to execute it, and since it contains the runtime expressions of the input file, it will do whatever we told it to do. Note that we're not using the CFASL! Since the CFASL contains compile-time expressions, there's no need to even touch the CFASL when we want to run our program.

Now, for something a bit more challenging, the good ol' REPL mode.

So, we're at the REPL, feeling free, taking a sip, and sneaking a sneaky (defmacro quux () `(print "hello")) in there.

WTF? What's the Lisp supposed to do? After all, a macro definition is not a runtime expression!

Now, older Lisps, which are not so big on following the latest trends in the scene, or newer Lisps that just don't care, will get this wrong. They'll intermingle runtime and compile-time in the REPL, leading to all kinds of undesirable effects, as Matthew Flatt will be pleased to point out to you. This means you'll have the compile-time entity in your runtime, which is simply sick. (In this post's model; Tom Lord and John Shutt will tell you something totally different.) One of the bad results of that is that code that works at the REPL won't necessarily work when put into a file. But don't take my blahger's word for it, take Professor Flatt's peer-reviewed one.

So what's an Acceptable Lisp™ to do, at the REPL, when the luser enters a DEFMACRO? An Acceptable Lisp™ has to keep the runtime runtime and the compile-time runtime strictly separated. This means it will need one environment for the runtime, that contains the global variables and functions, and another, completely separate environment for the compile-time, that contains macro definitions.

Let's try to wrap this up...

Wrap-up

We've seen that a Lisp file contains code for two different times: the runtime runtime and the compile-time runtime. If we want to make our users' lives easier, we as Lisp implementors have to keep these times separate.

In Part II of this post, I'll write about how PLT Scheme handles module dependencies in such a phase-separated model.

Now I gotsta hack.

semantic, adj.

In English, mental refers both to things involving the mind in general, as well as crazy in particular.

I propose to use the word semantic the same way for PLs.

If a language has wack semantics, you'd say ``Woah, that language is totally semantic!''

I, for one, welcome our new app-inventing grandchildren of Interlisp overlords

Outlining! 2D forms!

App Inventor for Android is based on OpenBlocks, which looks very much like Scratch, but uses Kawa instead of Squeak.

All of this stuff seems to use a dataflow/event-driven layer instead of a "common" programming language. I wonder if such 2D forms would also be useful as a syntax for an ordinary PL... Need to investigate more...

Optimism

Nice and thought-provoking comment from one markt on LtU:

One of the things that stands out to me about APL, and CS in general is the pure, unfettered optimism concerning the human condition. What other field of human endeavor seeks to provide tools to enhance human productivety and intelligence, to augment the facutlties of the mind? The implicit assumption that this can be done is a leap of hope in the purist form.

"Tools for thinking" I mean who would think of such a thing? Seems like this goes unstated a lot of the time, I imagine most of us here take it for granted. But it really is an amazing proposition (imo).

(no longer a) PSA

(Sorry for this stupid post, I'll get back to the topic at another time.)

Using multiple kinds of parentheses in Lisp considered silly

Complaining about parentheses in Lisp is like complaining about the heat in the kitchen—as Jamie Oliver said, the cocktail of hot oil, sharp knives and cocaine is fucking lethal. Wait, what am I trying to say? Anyway, I'm not going there.

I'm stealing a picture from that post though, to show something that bothers me:



Yeah, that WTF? is spot on, but for a different reason.

Note how a couple of unshapely ]'s have snuck their way into the midst of that lovely column of gorgeous )'s.

Now why is that WTF?-worthy? Well, one of the Big Benefits of using just two characters for all punctuation is the editing convenience you get from that. After you've moved code around in Lisp, you just keep your finger on the delete key until the superfluous parens are gone for good, or, else you just keep your finger on the )-key until the right amount of parens are where they should be.

This is completely broken when you use, for some irrational reason, a second or – gasp! – third kind of parenthesis. Doing that brings you back to the laughable world of lots of irrational and silly punctuation. You have to manually match up a [ with a ]. Think about that! It defeats one of the Big Benefits of using S-expressions, the only sane, rational syntax there is.

Silly.

(Of course, a lot of the time you use Emacs's sexp-aware killing and stuff, and then you can get away with multiple kinds of parens. But still, that doesn't work in the general case.)

C# 4.0: The industrial response to Lisp?

I just skimmed the C# 4.0 specification, and I'm quite fascinated by it. The spec itself is extremely well-written and readable, as far as language specs go. Anyone who's ever tried to read ECMA-262 or the JLS will agree.

But more importantly, the features of the language really raise the bar for all dynlangs out there. With dynamic, C# effectively usurps untyped programming, while maintaining static types elsewhere. With LINQ's expression trees, they seem to tackle the same problem as Lisp does with S-expressions. And MSR's work on evolving generics, a lot of which seems to finds its way into C#, is highly interesting.

That C# qua language is way ahead of Java is clear. What's more worrying is that all dynlangs, including the few acceptable ones, like Common Lisp and Factor, are starting to look horribly dated against C#. Seriously.

Note the following quote by grandmaster Tim Sweeney:

In Bracha's work here, and in Microsoft's work on C# 3.0, I sense an undercurrent dragging the language model toward the LISP/Smalltalk "ideal" of metadata-intensive, introspective, dynamic, loosely-typeable programming programming. ... If you go this route, one day you'll realize you evolved the ultimate hacker language, and it became a godawful mess for writing real programs. (My emphasis)

Evolving the ultimate hacker language, and making it a godawful mess for writing real programs, is of course the topic of this blog. Note also the following quote on C#'s type system, again by Sweeney:

These [Variance and Generalized Constraints for C# Generics] extensions stretch the C language family to an impressive new local optima

Now, I must say that I haven't ever used C#, so I don't know how it stacks up ITRW. I'm assuming, based on long observation, that Microsoft with high likelihood fucked up the pragmatics completely, and that programming in C#, as opposed to reading about it, is deeply depressing.

But what I'm saying is that the Perl-Python-Ruby-PHP-Tcl-Lisp-language has to catch up, maybe for the first time since 1959!

Manifesto on JAR's Next Language

I just had a good laugh reading this, so I thought I'd share:

Reasons to design a programming language:

• All current languages are terrible.
• It's fun, easy, and satisfying.
• You get to choose a silly name for it.
• Languages are games, and people like games.
• Everyone else is doing it.
• There's a chance it might be useful for something.
• By doing it once again, perhaps some people will learn something new.