Do You Know Of Any ~ATH Implementations Or Dialects Besides Yours And Mine?

I am so bad at updating. (and a side note)

I am still intending to update this.

but I am not good at time management.

I have like 2 hours of free time after school, (because of certain inefficiencies on my part) so that's my stupid excuse.

I am working on writing some responses to some requests for clarification, and then I will do the second part of the converter TO binary.

the print command can have the end of a loop in.

The contents of the print command can be executed.

so if you say:

import blah A; import bleh B; ~ATH(A){ ~ATH(B){ print heh } ~ATH(NULL){ print giant frogs alert; B.DIE(); } }  

it should print

"heh } ~ATH(NULL){ print giant frogs alert giant frogs alert"

this of course, is not particularly useful as far as I can tell, except possibly for quines, and possibly obfuscation. but really, is it possibly to write anything in ~ATH that isn't obfuscated?

speaking of not obfuscated, I have started working on adding a feature for user defined functions. I have it pretty much worked out how it will work, but I am not sure how I want to make the user DEFINE the functions. so I haven't been doing NOTHING with regards to this.

User defined functions can be used now! Also PRINT2 added!

PRINT2 prints a string object as a string. If you give it something that is not a string object it will probably crash, or output the empty string.

Now, to import user defined functions use 

importf filename as FUNCTIONNAME;

Why didn't I put this up earlier?

because I have other projects and because sometimes I am busy/lazy.

OK!

So to make a function to import, you pretty much make a ~ATH file like normal, except the object passed into the function is put into the variable ARGS, and when you want to pass the output object of the function out, you have it as an argument to the DIE method.

It can either be when saying THIS.DIE(RETURNOBJ); or ANYOTHERVARIABLE.DIE(RETURNOBJ);

The object your killing doesn't have to be the one tied to the life of the function or anything.

Oh I should clarify, the THIS variable/object will refer to the current function, not the file that called it.

It CAN be returned, and will be dead when you exit the function.

PROGRESS!

FUNCTIONS WORK.

LESS SPAGHETTI CODE.

although this starts to make me wonder if maybe this strays too far from what ~ATH is like in comic, 

but w/e.

Welcome.

If its not obvious, this blog will teach techniques in a version of ~ATH Specifically, drocta ~ATH. And yes, I'm serious. I believe that I probably am the author of the longest ( possibly up to third longest) ~ATH program. Specifically, bubble sort. Further, I believe I am the author of the first ~ATH interpreter. (I am the drocta of drocta ~ATH) Link here: http://www.mspaforums.com/showthread.php?50314-ATH-interpreter This blog is intended to serve as a tutorial on writing things in drocta ~ATH. First I will go over the syntax. Then how to make conditional like things. Then how to make finite loops. Then perhaps storing "numbers". Then copying "numbers" Then comparing "numbers" Then lists In such a way I intend to work our way up to you understanding how to write bubble sort in drocta ~ATH. After that, we might even implement brainf*** in it. Or a universal Turing machine. Of course, when I say "numbers", the quotes are there for a reason. Numbers are not built into drocta ~ATH. One has to build them. Now how do I tag things...? Ah, that's how. In case it wasn't clear, the interpreter for drocta ~ATH does not attempt the impossible. It cannot trigger the apocalypse. It's just a python script that interprets ~ATH scripts. Don't worry if you don't already know how to program. I intend to make the posts fairly accessible.

I made a parser

I made a parser for ~ATH in addition to the interpreter. I might make the interpreter use the parser at some point in the future, or I might not. If I made the interpreter use the parser, the code for the interpreter would probably be a little cleaner, and possibly a little faster.

The parser is available from my github.

To use it, call tokenize on the text, and then read_all_from on the result of tokenize.

The output will be a list of lists

https://github.com/drocta/TILDE-ATH-Parser

(haven't updated lately because of other unrelated projects, and also other reasons that aren't necessary to describe)

Syntax explanation

~ATH is a programming language. A programming language is a language that can be used to write programs in, that can afterwards be run. ~ATH appears in Homestuck, and drocta ~ATH is my attempt at making it a real thing.

drocta ~ATH is limited to what is physically possible of course, because it actually exists. It is also limited to fit well with the comic, and how much of it I have created so far.

As such, so far all the programs that can be written in drocta ~ATH are text only, do not yet accept input (input in progress), and tend to be stupidly long.

It is however capable of computing anything a computer can given enough time and memory.

~ATH is based on objects, and the lives thereof. 

For example, in homestuck, there are programs that relate things to the lifespan of a universe, or a person, etc.

Of course, in order to make an interpreter possible, we have to limit ourselves to virtual objects. Ones that do not actually exist.

References to these objects are stored in things called variables. You probably know what these are already, but if you don't, think of a variable as a box that can have a thing in it. Each box has a name. You can say "do something with whatever is in the box called "apple".

In ~ATH, every object is either "alive" or "dead".  Note that it is the object that is alive or dead, not the variable that refers to the object.

Now onto the actual syntax!:

Currently(as of build 7), drocta ~ATH has 5 different commands: import, ~ATH(){}, .DIE();, print, and BIFURCATE.

First we will go over the import statement. The import statement has the purpose of creating a new variable and a new object. Unlike other operations that can create a new object, the object created by import is initially unrelated to all of the other objects.

The syntax of the import command is:

import anything other than semicolons here VARIABLENAME;

The things between import and the last space before the semicolon are ignored.  

In future versions, the import statement might also do additional things, such as using things from other files. But for now it just initializes variables.

EXAMPLES:

If you wanted to make a new variable called BANANA, and you wanted the reader to know that BANANA is a fruit, you would say:

import FRUIT BANANA;

or you could say

import YELLOW FRUIT BANANA;

The things that go before the variable name don't actually matter, you can say

import ghsdgh hgsdkhg hgksdhg hgskdjg BANANA;

if you wanted.

Now for the next command: The eponymous ~ATH loop!

The syntax of it is as follows:

~ATH(VARNAME){ Some other code goes here }

What this does, is each time the code execution gets to the ~, it checks what the variable in the parentheses is, and then checks the object the variable points to. If the object is alive, it continues. If the object is dead, it skips to after the }. When the Code execution reaches the }, (that is, if it did not skip to after it), it will jump back to the corresponding ~. Code execution will keep going around in this loop until the object pointed to by the variable is not alive. In later versions, there may be an additional requirement that the } be followed by EXECUTE(code here); Where code here can be replaced with ~ATH code, NULL, or possibly a file name. But currently, this is not the case. Now onto the command BIFURCATE!: Bufurcate actually has 2 forms, the standard bifurcate, And the reverse bifurcate. The first of the two has syntax as follows:

BIFURCATE VARNAME1[VARNAME2,VARNAME3];

What this does, is it takes the object pointed to by the first variable, and determines two objects, which are stored in the other two variables. The two objects determined will always be the same for a particular object.  If you for example say:

BIFURCATE V1[V2,V3]; V2.DIE(); BIFURCATE V1[V4,V5]; 

then V4 and V2 will point to the same object, which will be dead.

The other form of the BIFURCATE command has syntax as follows:

BIFURCATE [VAR1,VAR2]VAR3;

This is pretty much the inverse operation. That is, it undoes the other one. If you say for example:

BIFURCATE A[B,C]; BIFURCATE [B,C]D; 

then A and D will point to the same object. To be clear, undoing the other is not the only time you can use it. It will take any two objects, and determine an object from those two. If there is already an object for the combination of those two, then that object is the resultant object. If none has been created yet (and the two arent split from something in that order), it will create a new object, which it will put in VAR3.

To be clear, if you say:

BIFURCATE A[B,C]; BIFURCATE [C,B]D; 

 A and D will NOT point to the same object, unless you create A such that B and C are the same. That is, if you say:

BIFURCATE [Y,Y]A; BIFURCATE A[B,C]; BIFURCATE [C,B]D; 

 A and D WILL point to the same object, and B and C will both point to the same object as Y (and as each other).

The Final command is print.

print is pretty simple, but its syntax might change somewhat.

Currently the syntax is:

print the text you want to output;

Note the semicolon. the semicolon is important.

So hello world would be

print Hello, World.;

SYNTAX EXPLANATION COMPLETE! PLEASE SUGGEST CLARIFICATIONS IF NECESSARY.

MESSAGE END.

re: "computationalalchemist answered: You could use Prolog-style lists for split."

"computationalalchemist answered: You could use Prolog-style lists for split."

I'm not totally sure what you mean, but my best guess as to what you meant is probably pretty similar to how I making it.

I implemented it and I think it works, I need to double check though.

With what I have now when I use 

split [THIS,THIS,THIS,THIS,THIS,THIS,NULL]COUNTER; ~ATH(COUNTER){ print "bananas"; BIFURCATE COUNTER[JUNK,COUNTER]; }EXECUTE(print "ok";); print "whee!";

it yields 

bananas bananas bananas bananas bananas bananas ok whee!

Which seems to make sense to me, and also, fits with how lisp lists work, and apparently also prolog lists.

also where it says split it will also accept bifurcate. they are actually treated as the same command.

import statements aren't fully implemented yet though.

I think I will put this version on github pretty soon.

Thank you for the advice.

User input, and a basic calculator.

Most programs people use have some form of user input. A calculator isn't much use if it always uses the same numbers after all!

~ATH of course accepts user input and output as shown in the file Roxy sent Jane.

Also I just found out you can put more than one read more line in one post.

The input command has the syntax:

INPUT VARNAME;

What this does is when program execution meets this line, the program pauses execution, allowing the user to input text. When the user hits enter, program execution will continue and the variable VARNAME will be made to point to an object corresponding to the text the user entered.

This object is such that the left half is the object that corresponds to the first character. If there is no character after that, the right half will be the NULL object. Otherwise the right half will be the object corresponding to the input without the first character. If they hit enter without inputting any characters the object will just be the NULL object.

Now that we have that all explained, we can start to make programs that actually take user input!

As you might have guessed from the title, the thing we will be making is a very basic calculator. All it does is add two numbers, like in the last example.

But in this, it will get the numbers from the user!

One simple way to do this is to use the length of the input text as the number: The way we define what we call numbers just so happens (heh) to be such that if we interpret the object for the input string as a number, the number will be the same as the length of the input!

This isn't the greatest solution, but it is easier than other methods. We will use this method first and then move on to other methods that are harder to write, but will be nicer when using the end program.

HERE WE GO:

ok, so like I said, much of it is pretty much the same as that previous program, so we might as well just include said here:

SOME CODE TO GET A AND B HERE import bluh BLAH; BIFURCATE [BLAH,A]ATEMP; BIFURCATE [BLAH,B]BTEMP; BIFURCATE ATEMP[JUNK,ATEMP]; BIFURCATE BTEMP[JUNK,BTEMP]; BIFURCATE [BLAH,NULL]C; BIFURCATE C[JUNK,C]; ~ATH(ATEMP){ BIFURCATE ATEMP[JUNK,ATEMP]; BIFURCATE [BLAH,C]C; } ~ATH(BTEMP){ BIFURCATE BTEMP[JUNK,BTEMP]; BIFURCATE [BLAH,C]C; } BIFURCATE [BLAH,C]CTEMP; BIFURCATE CTEMP[JUNK,CTEMP]; ~ATH(CTEMP){ BIFURCATE CTEMP[JUNK,CTEMP]; print some text; } print DONE!; 

so pretty much what we need to do it put the code to get A and B where that goes(at the beginning), as well as stuff to tell the user how to print stuff.

Like I said, the objects from the input commands can be interpreted as numbers.

so this becomes:

print INPUT SOMETHING WITH THE NUMBER OF CHARACTERS AS THE FIRST NUMBER YOU WANT TO ADD; INPUT A; print INPUT SOMETHING WITH THE NUMBER OF CHARACTERS AS THE SECOND NUMBER YOU WANT TO ADD; IMPORT B; import bluh BLAH; BIFURCATE [BLAH,A]ATEMP; BIFURCATE [BLAH,B]BTEMP; BIFURCATE ATEMP[JUNK,ATEMP]; BIFURCATE BTEMP[JUNK,BTEMP]; BIFURCATE [BLAH,NULL]C; BIFURCATE C[JUNK,C]; ~ATH(ATEMP){ BIFURCATE ATEMP[JUNK,ATEMP]; BIFURCATE [BLAH,C]C; } ~ATH(BTEMP){ BIFURCATE BTEMP[JUNK,BTEMP]; BIFURCATE [BLAH,C]C; } BIFURCATE [BLAH,C]CTEMP; BIFURCATE CTEMP[JUNK,CTEMP]; ~ATH(CTEMP){ BIFURCATE CTEMP[JUNK,CTEMP]; print some text; } print DONE!; 

So yeah. That should work. I still need to test this, but I am pretty dang sure that this works.(have to go do homework now) In the next post I will explain how to make it so that the user can type in the number as an actual number!

unary to binary (back again, an objects journey) (part 1)

ok, so last time I didn't actually get to converting from unary to binary again.

So I guess I will do that here. (HAHA I JUST GOT IT TO DIVIDE WITH REMAINDER. APPLYING THAT TO THE CONVERSION IN PART 2)

so to do this we have to find the largest power of two not greater than the number,  then the largest power of two not greater then what is remaining of the number, and repeat until we reach zero.

OR we can

find the largest power of two not greater than the number, then for each power of two less than the number, if it is not greater than it, use a 1 and subtract it from the remaining number.

and then for both of these, of course, reverse the result at the end so it is in the right order. 

I think so far the second one of these sounds better.

but come to think of it, repeated integer division by two might work well.

that is, instead of repeated doubling, and then checking which one is larger, it might be faster to do repeated halving. (with a remained of either zero or one)

this could be then used to make a integer base 2 logarithm, with a remainder thing.

and instead of checking each power of two less than it, we could just find the integer base 2 log, and the base 2 log of the "remainder" and the base 2 log of THAT remainder and so on.

that seems like its the best way to me so far, but how do we take the integer logarithm AND get the log remainder thing?

getting the logarithm is fairly simple, just keep dividing by the number until you reach 1.

but how do you get the remainder part? and also how do you do the integer division in the first place with these number like things that we have created?

well lets answer the first of those questions first.

It seems like the remainder of the logarithm is probably related to the remainders of the divisions that make it up.

so lets look at taking the integer base 2 logarithm of 9:

9 4 (rem 1) 2 1

that took 3 halvings, and there was one remainder of one.

what if we try on 10 or something though?

10 5 2 (rem 1) 1

that also took 3 steps and one remainder, but the remainder was on a different step.

now lets try 11

11

5 (rem 1) 2 (rem 1) 1 (rem 0)

now wait, theres a pattern here, which might allow for converting it even faster.

the remainders of the divisions are the remainders of the logs expressed in binary!

we COULD turn them back into unary, so we could take the log of it again,

but what we are doing it converting it into binary anyway! (so that would be a silly step)

so if we look at what we were doing with the repeated division again, but looking at the numbers in binary, it will be pretty evident how to convert it in a better way.:

lets try 9 again (which is 1001)

1001

100 (rem 1) 10 (rem 0) 1 (rem 0) 0 (rem 1)

whats that? the remainders are the number we want in binary?

and now that I think of that it was kind of obvious?

yes. I for some reason did not think of that earlier.

so yeah, thats what we will do, we will repeatedly integer divide the number by 2 until we reach zero, and the remainders will be the number in binary.

but wait, we haven't even written out how to divide the number by two in the first place!

well turns out its not actually that hard to do.

in a loop (that keeps looping provided the number is not zero) subtract one, and if its still not zero, subtract one again. if it is zero after the first subtracting, the remainder of that division is 1. (only the last loop will cause the remainder to be 1) count the number of times where 2 was subtracted. 

so like for 5 that would be

5 3 1 1 2 0 2 (remainder 1)

so that gives the correct answer (2 with remainder zero)

ok, now to implement this integer division by 2 in drocta ~ATH (wow! already? that was fast! /sarcasm)

//given that NUM is the variable that has the name we are halving BIFURCATE [NUM,NUM]2NUM; BIFURCATE 2NUM[NUMCPY,JUNK]; BIFURCATE [BLAH,BLAH]2BLAH; BIFURCATE 2NULL[SUBCOUNT,REMAINDER]; ~ATH(NUMCPY){ BIFURCATE 2BLAH[UNEVEN,JUNK]; BIFURCATE NUMCPY[JUNK,NUMCPY]; BIFURCATE [NUMCPY,NUMCPY]2NUMCPY; BIFURCATE 2NUMCPY[NUMCPYCPY,JUNK]; ~ATH(NUMCPYCPY){ BIFURCATE [BLAH,SUBCOUNT]SUBCOUNT; BIFURCATE NUMCPY[JUNK,NUMCPY]; BIFURCATE 2NULL[UNEVEN,NUMCPYCPY]; } ~ATH(UNEVEN){ BIFURCATE [BLAH,REMAINDER]REMAINDER; BIFURCATE 2NULL[UNEVEN,JUNK]; } } BIFURCATE [NUMCPY,NUMCPY]2NUMCPY; BIFURCATE 2NUMCPY[NUMDIV2,JUNK]; 

ok, that should divide the number by 2, put the result in NUMDIV2, and put the remainder in REMAINDER.

this takes linear time based on the size of the unary number. (provided I didn't make a mistake)

so now it is time to make sure it works, hold on a second while I check it. (I mean, im not going to post this until after I check it, so it doesn't really make sense to tell you to wait, because you just keep reading, but as I am writing this, I am about to test it.)

ok, I tested it. and it didn't seem to work, but then I realized that I made a mistake in making the test. (I checked the wrong variable)

but yeah, turns out that works...

ok, so now we need to repeatedly divide the number by two to get the remainders.

and we need to store these remainders in a list or something.

I couldn't help but notice that out of some of the best things you could do, no link to the download in the sidebar? You may want to do that...

uh, yes. that would be good to do.

I put a link in the about section, but I don't actually know how to put links in the sidebar outside of that.

do you think you could explain that for me?(I am not particularly experienced with tumblr)

EDIT:Nevermind, my friend from school told me where the feature is.

@ karkathateseveryonesposts

Anon in that ask was me, not rping as you say but what I meant from computer-exploding code is I was unsure if ~ATH was supposed to crash once running a code like Sollux's.  Also, there couldn't be a syntax error as the program crashes if I press enter with no code on it yet.

Ah! My apologies for my misunderstanding, I interpreted the reference to computer exploding scripts combined with your handle to [ justification for why I misinterpreted snipped for brevity].

You don’t type the script in when you run the program. You need to first save the script as a seperate file, and then when you run the interpreter, type in the filename.

The last bit of the interpreter says like

filename=raw_input() filelink=open(filename,'r') script=filelink.read(-1) result_obj=evalScript(script,NULL_obj) raw_input("press enter to close")

I probably should have made it more clearly ask for a file name instead of just having a prompt with no explanation.

(also, wow, this reminds me that this code was written in python 2.7 and not python 3. Maybe I should update it to use python 3? There’s also many other things about it that ought to be improved but it isn’t high on my priorities..)

My ~ATH keeps crashing whenever I run a non-computer exploding code, or if I press enter. Why is this?

Based on the username which recently followed this blog at the same time as I received this ask (which I won't say what it was because you asked this on anon, and haven't received permission), I'd guess that you may be running a different version of ~ATH than the one I document here :P

Unfortunately, I don't have a specification for the version you have.

If you have such a specification, or the executable for it, I'd be happy to take a look if you could send it to me, haha.

If on the off chance you are using the version I document here, then presumably you have a syntax error or something in your code?