Lambda the Ultimate

I really like the phrase "the LTU Radar". Can we get a graphic of the radar? What other languages are on the LTU Radar?

According to the Departments page, the other languages are Javascript, Python, and Ruby, in addition to Scala.

I suspect the radar has a rather slow sweep time, though, and most recently passed over a cluster of the languages formerly known as scripting languages. ;-P

Yes, the "formal" definition of the LtU radar has to be the languages for which we decided to have a dedicated "spotlight" department.

Thinking back on the history of LtU for a bit, I remember that the first language in this category was Python. At the time not only was Python becoming prominent, but also many cool PL related projects were done in and around Python (logical programmin in Python, partial evaluation for Python etc.) so it was a natural department to have. The adoption of Javascript as a mainstream language, and the interesting discussions about the future of the language merited dedicating a department. As always, the fact that we had an editor committed to the language (excuse the pun), helped put it on the LtU radar.

Ruby and Scala were added by me in the hope they will get more attention here. Ruby never managed to do that, but after waiting quite awhile it seems Scala is starting to get the attention here that it deserves. I guess we are not that much ahead of the curve with Scala as we should strive to be...

Two other languages deserve a mention in this trip down memory lane. Erlang was a frequent topic of discussion, without having a dedicated department, way before every hacker knew what it meant. Luke was the guy responsible. Finally, for quite a stretch of time Frank was the one who made sure LtU was Haskell centric. I think this explains to a large extent the (wrong) impression that LtU is an FP site.

Why is the impression that LtU is an FP site wrong? The name, "Lambda the Ultimate" is definitely FP-biased. I definitely don't think LtU is an FP site, but I definitely would understand why it would be seen as such by a new reader.

Why is the impression that LtU is an FP site wrong? The name, "Lambda the Ultimate" is definitely FP-biased.

Is it? The view of lambda as virtually synonymous with FP is a programmer's-eye view which sees only the tip of an enormous iceberg, that began forming decades before any high-level programming languages existed.

Glancing at the titles of the eponymous Steele and Sussmann papers gives some hint as to the bigger picture: Lambda was, chronologically, described as "The Ultimate Imperative", "The Ultimate Declarative", "The Ultimate GOTO", and "The Ultimate Opcode".

This was about more than a functional programming language: it was about techniques for implementing programming languages, including imperative ones, and for understanding programming languages, both informally as well as formally, by providing semantics for languages. That is why it is important to LtU.

That series of papers was just one part of a much larger field of study surrounding lambda, which encompasses mathematics, mathematical logic, type theory, programming languages, and programming language semantics. The Lambda papers were preceded by work such as Landin's on applying lambda to Algol 60 in the early '60s, and his invention of the SECD abstract machine and thus formal operational semantics for programming languages - again, not just functional ones.

For a more complete history, see the History of Lambda-Calculus and Combinatory logic, which covers a span of almost 90 years, and gives some idea of the diversity and breadth of the subject.

Surely important. But the name of the site IS LtU, actually, many people might not be familiar with Steele's paper series, what are they going to take away from the name? FP and definitely lambda calculus is one model computation of which there are many. So...many conversations here are not heavily related to the lambda calculus, but it would take a new reader sometime to figure that out.

But the name of the site IS LtU, actually, many people might not be familiar with Steele's paper series,

The first real link in the body of the Getting Started page is to the Lambda the Ultimate series of papers, so hopefully a new reader with a serious interest in programming languages would soon figure that connection out.

Just a couple of links later on the above page is a discussion of category theory from 2004. That has since become a significant topic on LtU - e.g. the Theory page shows a strong category theory bias. Lambda, the Ultimate internal language of Cartesian closed categories! Again, this is a topic that an uninitiated programmer may tend to conflate with FP, seeing talk of monads, arrows, and formal type theories. The fact that Haskell is the most popular language that seems most interested in exploiting category theory feeds into this.

So yes, I agree there are various reasons it's an easy mistake to make. It's still a wrong impression, though.

FP and definitely lambda calculus is one model [of] computation of which there are many.

But again, there's a distinction between available models of computation, and approaches to semantics of languages. Someone who's only thinking in terms of the former has not yet achieved LtU-nature, but that's what the archive containing thousands of links is for.

Anton is of course right about the Lambda papers. But as far as getting the impression wrong, the issue is much simpler:

One would hope that the reader reads the second line of the logo which state "the programming languages weblog". If that's not enough of a clue, how about visiting the FAQ page which states "Mostly this site deals with issues directly related to programming languages, and programming language research. However, we allow ourselves moderate forays to bordering issues like programmability and language in general...This weblog is dedicated to the study of general properties of programming languages..." Or check the first few posts from the archives.

Still not enough? How about checking which departments exist here? I think the OOP, history, implementation, logic/declarative departments are a good clue as to the scope of the site. The fact that the only languages with dedicated departments are Javascript, Python, Ruby and Scala might also help newcomers orient themselves.

Not enough you say? Then how about doing the simplest thing and going over the stories on the home page? At the moment, for example, only one story is about FP, while there are three stories (about COBOL, BitC and Scala) that clearly go beyond classic FP languages, not to mention items about concurrency and security.

And by the way, as a matter of historical record: The name was meant to be "Lambda: The Ultimate Weblog" NOT "the 'Lambda the Ultimate' weblog".

But as far as getting the impression wrong, the issue is much simpler

That's true, and worth noting. What I was trying to account for was that even after (presumably) reading some or all of those descriptions and hints, some people apparently still reach the "FP site" conclusion. One of the points I wanted to make for the record is that PL semantics has a mathematical bias, and that is often mistaken for FP by people who are looking at it from a programmer's perspective.

PL semantics definitely has a mathematical bias, but many of us focus more on PL pragmatics (usability), which is an entirely different beast and very much still PL.

In my opinion, LtU is analogous to a SIGPLAN conference, with lots of ICFP and POPL and bits of OOPSLA and PLDI mixed in. Which is of course completely OK, whether LtU is an FP site or not really depends on the community that decides to participate here.

What LtU is really depends on the community that decides to participate here.

That's exactly right. Which means false impressions can easily become self-fulfilling prophecies.

But as one member of the community, I at least am interested in way more than FP, and indeed like you I think pragmatics are of central importance.

Maybe someone can write a programming language for analyzing LtU posts, and someone else can write one for generating radar images. ;-)

Here at my company we have a "technology radar" onto which I try to push a "functional programming" ufo - with moderate success.