MANOOL is Not an Object-Oriented Language!”

The Programming Language MANOOL


MANOOL is a fairly readable homoiconic language with primarily value (non-referential) semantics that balances the programmer's productivity with scalability…

What is the purpose of MANOOL?

MANOOL is a general-purpose language suitable for diverse programming tasks from several problem domains. However, it has substantial bias toward scripting and hopefully represents an evolutionary step over existing scripting languages. Thus, MANOOL should be compared to and can serve the same purpose as Python, Ruby, PHP, Scheme, JavaScript, or Lua, which implies properties typical for such languages, e.g.: short edit-compile-test development cycle (at least for code bases under 1 MLOCs) and run-time type checking.

What real-world problem does your project solve?

MANOOL is intentionally designed in such way that it represents a specific point on a continuum; on one side of this continuum are high-level languages designed with the programmer's convenience and productivity in mind, and on the other side are languages designed with execution speed and/or solution scalability in mind (whereas, as a matter of fact, software quality and reliability may correlate with either of those extremes, depending on the situation). Based on my past programming experience, I argue that no existing mainstream language addresses both goals at once in a balanced way. As a result, programs are either more expensive in development than they should be or do not scale with workload as needed.

Think, for instance, about the number of registered users and their contributions on an in-house social-network Web-site. Working as a server infrastructure administrator, on multiple occasions I saw serious scalability issues popping up suddenly after a year of production use due to flaws in backend software architecture, including the choice of programming language and/or its implementation.

As a more elaborate illustration, using a rigid static type system and gratuitous sharing mutable state in the program (widespread in traditional OOP) is far from how many people think about problem solving in terms of basic (immutable) mathematical objects, and thus this places undue cognitive load on developers for simple backend scripting. On the other hand, implementing backend algorithms in an average exploratory-programming language (especially in a straightforward and/or idiomatic way) often leads to poor run-time performance scalability w.r.t. Web-site workload growth.

OK, but what warrants a whole new language in case of MANOOL?

Starting off with some relatively unpopular language makes little sense for me as a language designer, since I might miss in this case some improvement opportunities (whatever it means) while getting almost nothing in return. But why not just extend an existing mainstream language to suite the above stated goals instead of creating one from scratch?

Achieving competing (and even incompatible) goals is hard and may lead to overly complex and difficult to adopt language designs. MANOOL leverages two principles in order to deal with this problem, which are currently not observed among mainstream languages:

Both of those principles require things to work slightly differently on the very basic level, which suggests that introducing a whole new language is more appropriate than trying to extend an existing one.

Why should I learn MANOOL?

It depends on who is asking. One possible reason is that playing around with MANOOL means joy and fun that existing mainstream languages can hardly offer to you. E.g., in brief:

What does it offer to potential project maintainers and contributors?

MANOOL is a personal, solo-developer project with severely limited resources. Thus, to be viable, it almost inevitably has to use a straightforward, streamlined, and modular implementation, which is based on simple algorithms and data structures (from the compiler theory standpoint). Let's take, for instance, the implementation size — the MANOOL translator is written in under 10 KLOCs, whereas the most widespread Python interpreter builds upon at least 100 KLOCs.

This does not necessarily mean that the MANOOL implementation is cheap or otherwise low-grade but rather that extra development effort can be committed to ensuring high implementation quality and reliability. This also implies lower project entry requirements, encouraging more people to participate in the development. Besides, such compact code bases are more suitable for educational purposes (than larger ones, which are often full of legacy stuff).

Give me a complete example of what code in MANOOL may look like

A “Hello World” program might look like

((extern "") in Out.WriteLine["Hello, world!"])

(using the 2nd version of the syntax, see below), and in the following sample program a recursive factorial function is defined and invoked:

( (extern "") in
: let rec (Fact = (proc (N) as
    : if N == 0 then 1 else N * Fact[N - 1]))
    Out.WriteLine["Factorial of 10 = " Fact[10]])

What's next? Do you have a roadmap for MANOOL?

Sure, here it is (as of September 2021):

  1. Complete a JIT compiler for MANOOL to achieve run-time performance only marginally slower than that of the most sophisticated dynamic-language engines on the market (such as V8 and LuaJIT) but only at a fraction of their complexity — this is doable as per my experiments.

  2. Replace {/} in the syntax by (/) in the 2nd version of the language (as shown in this writeup). The idea is to appeal more to at least one established language community (Lisp/Scheme), albeit at the cost of extra complexity (including a more complicated LL(2) parser).

  3. Complete and polish the MANOOL language Specification and the Tutorial.

  4. And ultimately, build a MANOOL ecosystem (libraries, tools, success stories) and a user community — any help is welcome!