<\/span><\/h2>\n\n\n\nThere are a few things about APL that are truly unique from what you’ve come across. Let’s take them, one at a time.<\/p>\n\n\n\n
1. Array-oriented processing:<\/strong> In APL, all objects are arrays. Whether it’s a single number, a list, a table, or even a very complex block of data with lots of layers, they’re all the same. So if you add 5 to something, it’s just added to each value contained in it. No loop needed.<\/p>\n\n\n\n2. Short, symbolic syntax:<\/strong> Unlike languages such as FORTRAN or SUM, APL’s symbols are short words. In other languages one character can be used for what takes a full line of characters. At first it looks frightening, but once you have grown accustomed to it you don’t mind anymore.<\/p>\n\n\n\n3. Built-in math functions:<\/strong> Some of the things included are sorting, reversing, finding sums, etc., and matrix work. No need to import a library or to write them yourself.<\/p>\n\n\n\n4. Interactive environment.<\/strong> You enter text, click enter and receive an answer immediately. Perfect for experimentation and learning along the way.<\/p>\n\n\n\n5. Implicit looping (rank and scalar extension):<\/strong> APL is technical in that respect, but simply means that APL will loop over your data for you. What you say, it does.What you say, it does it without saying anything.<\/p>\n\n\n\n6. Tacit programming:<\/strong> This is a slightly more complicated one. Functions can be written without ever naming the data that they act upon, thus maintaining a very compact code. It can be omitted at this stage for beginners.<\/p>\n\n\n\n<\/span>APL Programming Language Symbols Explained<\/strong><\/span><\/h2>\n\n\n\nTo be honest, the first thing that everyone notices about APL are the symbols. They’re also the one reason why people fall in love or run away. However, there’s a reason to have them. Iverson intended APL to be a notation for mathematics, and each of its symbols represents a single clear concept. After becoming familiar with the meaning of a symbol you can read it quicker than a long string of words.<\/p>\n\n\n\n
These are some of the most frequently used ones:<\/p>\n\n\n\nSymbol <\/td> Name <\/td> What it does <\/td><\/tr> \u2373 <\/td> Iota <\/td> Makes a list of numbers. \u23735 gives you 1 2 3 4 5 <\/td><\/tr> \u2374 <\/td> Rho <\/td> Reshapes data or tells you its shape. 2 3\u2374\u23736 makes a 2 by 3 table <\/td><\/tr> +\/ <\/td> Reduce (sum) <\/td> Adds up everything in a list. +\/1 2 3 gives 6 <\/td><\/tr> \u233d <\/td> Reverse <\/td> Flips a list backwards <\/td><\/tr> \u234b <\/td> Grade up <\/td> Gives the order you\u2019d need to sort a list from smallest to largest <\/td><\/tr> \u2218.\u00d7 <\/td> Outer product <\/td> Multiplies every item in one list by every item in another, like a times table <\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\nWhat can be confusing to the beginning programmer is the fact that the same symbol may be used for two purposes. It depends on whether or not there is something on its left. If a function has a single argument (right-hand side), then it is monadic. It is dyadic in its form when it takes two (one on each side).<\/p>\n\n\n\n
A shape can be described by a word by itself, like \u2374, which will tell you the shape of something. However, add a number to its left, and it will change the structure of your data to become that number. Two meanings, one symbol. It may sound crazy, but that’s how subtracting with – works too: 8-5 subtracts and -5 flips the sign.<\/p>\n\n\n\n
Yet another quirk: APL is reading from right to left, not left to right. So 2\u00d73+4 means \u201cadd 3 and 4 first, then multiply by 2.\u201d It’ll stump almost anyone when they start with it, so don’t worry if it seems like it’s backwards.<\/p>\n\n\n\n
<\/span>How to type APL symbols<\/strong><\/span><\/h3>\n\n\n\nThis is the one that everyone is afraid of, but it’s not so hard to do:<\/p>\n\n\n\n
\nTryAPL:<\/strong> Can use a touch screen as a keyboard (no memory required)<\/li>\n\n\n\nDyalog IDE:<\/strong> allows you to type symbols with the prefix key, and displays a keyboard layout.<\/li>\n\n\n\nKeyboard layouts:<\/strong> you can get an APL layout for your computer and switch to it when required<\/li>\n<\/ul>\n\n\n\n<\/span>APL Programming Language Examples<\/strong><\/span><\/h2>\n\n\n\nWords and phrases can only take you so far with regard to reading. The kicker is when you’re running something. Here are five small examples you can use in TryAPL to play with: I’ve included a brief python version next to each one, for comparison.<\/p>\n\n\n\n
<\/span>Example 1: Making a sequence<\/strong><\/span><\/h3>\n\n\n\n\u237310<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\nResult:<\/strong> 1 2 3 4 5 6 7 8 9 10<\/p>\n\n\n\nThat\u2019s it. One symbol and one number. You would write a line like this in Python: list(range(1,11)). APL is not hard, its shorter. <\/p>\n\n\n\n
<\/span>Example 2: Sum and average<\/strong><\/span><\/h3>\n\n\n\ndata \u2190 4 8 15 16 23 42Result:<\/strong> 108 for the sum, and 18 for the average.<\/p>\n\n\n\nThe +\/ is used to add all together. The average is the total divided by the number of items. Quick note: just note that the length of a list is given by \u2374data but the results is a list containing one element, but this is okay here. This would be written as sum(data) \/ len(data) in Python. <\/p>\n\n\n\n
<\/span>Example 3: Turning a list into a table<\/strong><\/span><\/h3>\n\n\n\n3 4\u2374\u237312<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\nResult: <\/strong><\/p>\n\n\n\n1\u00a0 2\u00a0 3\u00a0 4In this case, \u237312 is the numbers 1 through 12, and 3 4\u2374 is the 3 rows and 4 columns. In Python, it would require a nested loop or using NumPy. <\/p>\n\n\n\n
<\/span>Example 4: Multiplication table<\/strong><\/span><\/h3>\n\n\n\n(\u23735)\u2218.\u00d7\u23735<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\nResult:<\/strong> <\/p>\n\n\n\n1\u00a0 2\u00a0 3\u00a0 4\u00a0 5The one that always elicits the “wait, really?” from novices. The outer product is the product of each element of the first list times each element of the second list. That’s two loops in Python and it takes a bit of code. <\/p>\n\n\n\n
<\/span>Example 5: Sorting data<\/strong><\/span><\/h3>\n\n\n\nscores \u2190 72 95 60 88Result:<\/strong> 60 72 88 95<\/p>\n\n\n\nRemember \u234b from earlier? It returns an order to sort the list. Then we select items from scores in that order. In Python, it’s just sorted(scores); it’s actually a little cleaner. Not all tasks are shorter in APL. <\/p>\n\n\n\n
<\/span>What Is APL Programming Language Used For?<\/strong><\/span><\/h2>\n\n\n\nAPL is not really a language that you would use to develop a website or mobile app. It is best suited for any situation where there are a lot of numbers, complex mathematics, and a need for a quick response. These are the top locations of occurrence.<\/p>\n\n\n\n
1. Finance and insurance:<\/strong> It’s here that APL has remained steadfast. It is applied by banks, trading companies and insurance companies in the fields of risk modelling, pricing and actuarial analysis. These positions include massive tables of data and formulas which are frequently modified, and APL allows analysts to quickly write and modify these tables. <\/p>\n\n\n\n2. Data analysis and statistics:<\/strong> The default for APL data is an array, so data-related operations, such as finding averages, sorting, grouping, matrix work, seem natural. You’ll appreciate it if you’ve ever had to deal with a long list of loops or a messy spreadsheet.<\/p>\n\n\n\n3. Scientific and mathematical computing:<\/strong> APL was originally created for math, and so is appropriate for simulations, modelling, and research. Scientists can type an idea nearly as they’d type it on paper.<\/p>\n\n\n\n4. Machine learning prototyping:<\/strong> Neural networks are basically array math, so they can be written in a surprisingly few lines of APL. Not a typical production ML, but some people draw with it to quickly jot down ideas.<\/p>\n\n\n\n<\/span>Real-world use<\/strong><\/span><\/h3>\n\n\n\nThe most popular modern version is by a company called Dyalog whose customers include finance, insurance and research groups, it says. Some of these companies have been operating APL systems for years and years, and they still continue to operate them because the code is still working and they can change it very quickly.<\/p>\n\n\n\n
So no, it’s not an attempt to compete with Python. It’s more of a specialised tool. The APL programming language can still write a few lines that other languages need to write a page.<\/p>\n\n\n\n
<\/span>Advantages and Limitations of APL Programming Language<\/strong><\/span><\/h2>\n\n\n\nAPL is like any tool\u2014it’s wonderful for some things, and it’s aggravating for others. Let’s see both sides before entering into something.<\/p>\n\n\n\n
<\/span>Advantages<\/strong><\/span><\/h3>\n\n\n\n1. You write less code:<\/strong> This is the mega one. An APL expression can often be a single line of code in another language.In other languages, a loop and a few lines of code may be one short APL expression. The fewer lines of code, the fewer spots for bugs.<\/p>\n\n\n\n2. Fast to build and test ideas:<\/strong> It’s interactive so you can experiment, watch the outcome and adjust it immediately. It’s a time-saver for analysts who need to dive into the data without delay.<\/p>\n\n\n\n3. Powerful array operations:<\/strong> They have all the tools of the trade\u2014sums, sorting, reshaping, matrix math\u2014right there. To perform simple numerical calculations, no additional libraries are required.<\/p>\n\n\n\n4. It changes how you think:<\/strong> If you don’t use APL at work, it will help you think more about data as a whole chunk than one item at a time. Many claim that it also enhanced their Python or R writing skills.<\/p>\n\n\n\n<\/span>Limitations<\/strong><\/span><\/h3>\n\n\n\n1. The learning curve is real:<\/strong> It takes a bit of a while to get back to being normal with the symbols, with reading from right to left, with the whole array mindset. You will experience some ups and downs for a few weeks.<\/p>\n\n\n\n2. Symbols can be hard to read: <\/strong>Once you know them, short code is lovely. However, after returning to your own code after a month or even other peoples’ after a month is over, it can feel like a puzzle. Detailed feedback is very much appreciated.<\/p>\n\n\n\n3. Smaller community:<\/strong> There are many less tutorials, forum answers and pre-made libraries than there are for Python. There is support for those who are stuck, but it may require digging.<\/p>\n\n\n\n4. Fewer job openings:<\/strong> Yes, there are real jobs that call for APL usage, primarily in finance, but not quite as many as Python or Java or SQL.<\/p>\n\n\n\nYou can try APL right in your browser on <\/em>