Hacker News Reader

Comment by deckar01 | Hacker News Reader

I assumed this was title related to standardized machine learning practices. Expanding "ML" to "Meta Language" would make this title clearer.

View on HN

The expansion of the acronym is more an etymology than the language's actual name. Calling it "Standard Meta-Language" would probably cause more confusion for most people. Similar to how nobody ever refers to Fortran as "Formula Translation", or APL as "A Programming Language", or COBOL as "Common Business-Oriented Language", or BASIC as "Beginner's All-Purpose Symbolic Instruction Language^H^H^H^H^H^H^H^HCode".

(edit: basic, not basil)

BASIC = Beginner's All-Purpose Symbolic Instruction Code (not Language).

Those abbreviations do not collide with a more common abbreviation in the same domain. Standard as a prefix in this title does not disambiguate its meaning. Anything to clarify it is a programming language would be nice.

The name "Standard ML", with a capital S, is not ambiguous - it's (still, just about) unambiguously the name of a programming language.

You have to call the language something. For decades it was mostly just referred to as ML. In recent years people (including me) have started calling it Standard ML explicitly to avoid confusing people who now expect ML to mean machine learning. It doesn't always work, as we see.

Expanding to "meta-language" sadly is no good either, since nobody has ever called it that and most practitioners wouldn't even recognise it.

SML is a possibility that's used often enough, but there's a semantic difficulty there - the difficulty is that Standard ML is the standard version of a language called ML; it isn't quite a language called SML.

> You have to call the language something. For decades it was mostly just referred to as ML. In recent years people (including me) have started calling it Standard ML explicitly to avoid confusing people who now expect ML to mean machine learning.

IIRC, both SML and Standard ML were used to distinguish it within the ML family of languages before machine learning became popular.

> For decades it was mostly just referred to as ML. In recent years people (including me) have started calling it Standard ML explicitly to avoid confusing people who now expect ML to mean machine learning.

Aren't they 2 separate languages? They have 2 separate Wiki articles:

https://en.wikipedia.org/wiki/Standard_ML

https://en.wikipedia.org/wiki/ML_(programming_language)

Or maybe the difference is more like Lisp vs Common Lisp.

If you're worried about the collision, I'd suggest petitioning the machine learning people to stop using the acronym. ML the programming language had been around for decades before it was common to use ML to mean "machine learning."

> a more common abbreviation

Google "standard ml" returns 1,090,000,000 results. At least the first page refers only to the language.

Google "ml machine learning" returns 183,000,000 results.

Google "ml programming" returns a page with some links referring to the language Standard ML and others to the language ML, none to machine learning.

I didn't even know there were people referring to machine learning as ML.

Standard ML outdates the common usage of ML as "Machine Learning" by a few decades though.

I'd like to do the exact opposite. Many a times have I hoped for a good article about a good language, but got presented with a memefied statistics article.

At least on HN I still know what NLP stands for…

> At least on HN I still know what NLP stands for…

Neuro-Linguistic Programming?

I agree that the abbreviation ML is too ambiguous to be useful in a title without added context. This title is particularly misleading due to its lack of context.

https://en.wikipedia.org/wiki/ML#Computing_and_mathematics

> use the original title, unless it is misleading

https://news.ycombinator.com/newsguidelines.html

We are agreeing that ML as an abbreviation on HN can be misleading.

Is there a "standard" for machine learning that's regularly capitalized?

Title case.

Title case is inherently misleading and, therefore, should not be used.

ML is a entire major language family that itself has roots in lisp. I've starred the "important" ones you'd probably know.

In the "core" ML family: original ML, Standard ML (SML), Concurrent ML (an extension of SML), Polly ML (yes, it's a different language that SML is implemented on top of).

Then there was the "lazy" family: Idris, Lazy ML, Miranda, Haskell, Purescript, and Elm (though it's not actually lazy).

Then you have the "pragmatic" family: Caml, Ocaml, Coq, F#, f-star, Alpaca, and Scala (these last two are somewhat less related here).

You also have the rough category of ML without a GC: ATS and Rust

There's a few others, but I'm not so familiar with them.

How are any of these coming from Lisp? I was under the impression that ML was a completely separate camp from Lisp.

ML was implemented in Stanford Lisp on a DEC10. The layout of the language is very lispy. Make parens around functions non-optional and add them around let, and a couple other constructs makes the code almost indistinguishable from lisp. First-class functions everywhere, linked lists, and garbage collection were also very-much tied to lisp at the time. Initially being implemented on top of Lisp is probably a key reason these were present.

Idris is eagerly evaluated.

You are correct (I'm sorry, I don't know much more than what I read in a cursory inspection some time ago). I should have called it the "lazy and/or completely immutable" category.

As are PureScript and Elm. Maybe "lazy" is just no longer a good name for the branch as a whole, given that it keeps growing new eager leaves.

As someone familiar with the name already, expanding it would be very confusing for me.

As someone familiar with Standard ML, you probably aren't the target audience of a tour of the language. The goal of this article is obviously to introduce new users to the language. Tricking users into clicking the link by being vague is not constructive way to recruit new users.

If anything, shortening to "SML".

Nobody says "Standard Meta Language", though.

I find it very irritating how people try to make others give up established words and phrases. There will always be ambiguity.

See also "crypto" or the recent spat here on HN about an old SQLite page.

I thought the M stood for module?

Nah, It's Meta (source: https://en.wikipedia.org/wiki/Standard_ML).

SideNote: I have a friend though, who used it for almost 15 years and thought the M stood for Math before we spook about this.

>Expanding "ML" to "Meta Language" would make this title clearer.

No, I don't think it would make it clearer, at all.

Is there any such thing as "standardized" machine learning at this point or even in the foreseeable future? The target is moving too fast.

Reply on news.ycombinator.com