The Imperium programming language - IPL

[Mechatrommer]

if some one has time, skills and interest in improving a language, why not not get a pre-existing C compiler

That's precisely what I did with my-c
So, it can be done!

can you publish for discussion? its much better than discussing uthopia. keep re-stating something secret will not do any good, similar to uthopia story..

[Mechatrommer]

Assembly language already support it, its the job of assembler to make best use of any available machine specific codes for it.

no! assembly is 1 to 1 map to machine code, no need to translate in any fancier way. assembler is just that, 1 to 1 mapping.. duh! how you can make such conclusion?

It's a language too, assembler can do anything (and more) that C can do.

and havent you got the idea yet? its a 1 to 1 mapping from brainf*ck grammar to at least something human can cope... so instead of typing 0x0001F 10E, we can type MOV REG1 for example.

[Sherlock Holmes]

Consider:

Unlike Java, which has a special operator >>> for logical shifting apart from the usual >> and <<, C and C++ have only arithmetic shifting with some areas left undefined and implementation-defined.

Most CPU architectures include instructions for arithmetic shift, logical shift and rotates, yet C that "designed for hardware" language doesn't support most of these variants, don't you get it? it was never designed for the applications you keep claiming it was designed for, it wasn't even designed if I can be blunt.

[Sherlock Holmes]

Assembly language already support it, its the job of assembler to make best use of any available machine specific codes for it.

no! assembly is 1 to 1 map to machine code, no need to translate in any fancier way. assembler is just that, 1 to 1 mapping.. duh! how you can make such conclusion?

It's a language too, assembler can do anything (and more) that C can do.

and havent you got the idea yet? its a 1 to 1 mapping from brainf*ck grammar to at least something human can cope... so instead of typing 0x0001F 10E, we can type MOV REG1 for example.

Yes I know that, but it is a language isn't it? are you claiming it's not a language? has no syntax rules? are you saying that?

[DC1MC]

Tja, I've tried to suggest something way easy to implement IMHO (that is: doable) than a whole new language, but I guess that slowly, slowly this thread will either goto (sic) irrelevance or deeper into d1ck measuring contest with more and more impractical stuff.
Out of everything I think DiTBho got the best outcome, at least they (see, political correct addressing mode  ) got something that is usable and makes their (again ) life easier.

The whole syntax boogaloo of C++ is because they've tried to cram as way much and outside of C paradigm stuff into it, but still it become one of the most used and popular languages, just because one doesn't need to throw away everything what they know and are familiar with, as well as the existing code base just to get some extra stuff.

But by all means, let me know when your improved language generates code, I will give it a fair try.

 Cheers,
 DC1MC

[Mechatrommer]

Assembly language already support it, its the job of assembler to make best use of any available machine specific codes for it.

no! assembly is 1 to 1 map to machine code, no need to translate in any fancier way. assembler is just that, 1 to 1 mapping.. duh! how you can make such conclusion?

It's a language too, assembler can do anything (and more) that C can do.

and havent you got the idea yet? its a 1 to 1 mapping from brainf*ck grammar to at least something human can cope... so instead of typing 0x0001F 10E, we can type MOV REG1 for example.

Yes I know that, but it is a language isn't it? are you claiming it's not a language? has no syntax rules? are you saying that?

there is difference between human language vs machine language... the purpose of the former is to fill every gap way down to the later. the way i see whenever anyone want to create a new language, its the human (friendly) language, a higher level one... but if what you are after is brainf*ck language, then by all means go for it, i've provided a link for you earlier to win the contest... just make it clear to us..

[Sherlock Holmes]

Assembly language already support it, its the job of assembler to make best use of any available machine specific codes for it.

no! assembly is 1 to 1 map to machine code, no need to translate in any fancier way. assembler is just that, 1 to 1 mapping.. duh! how you can make such conclusion?

It's a language too, assembler can do anything (and more) that C can do.

and havent you got the idea yet? its a 1 to 1 mapping from brainf*ck grammar to at least something human can cope... so instead of typing 0x0001F 10E, we can type MOV REG1 for example.

Yes I know that, but it is a language isn't it? are you claiming it's not a language? has no syntax rules? are you saying that?

there is difference between human language vs machine language... the purpose of the former is to fill every gap way down to the later. the way i see whenever anyone want to create a new language, its the human (friendly) language, a higher level one... but if what you are after is brainf*ck language, then by all means go for it, i've provided a link for you earlier to win the contest... just make it clear to us..

So assembler is a language then, good we agree on that. Now can the C language do anything that assembly language can't? You want to talk about languages, lets talk...

[Mechatrommer]

...snip..
So assembler is a language then, good we agree on that. Now can the C language do anything that assembly language can't? You want to talk about languages, lets talk...

in the end... your brainf*ck language will get eventually be translated into machine codes anyway... that turing complete machine?... like all other programming languages in existance. you want bit type and string support? go for it! it will get introduced into bloatness of machine codes anyway... like most of the rest (Basic, Pascal etc)... and sorry, i'm not further discussing since its clear what you re after now (brainf*ck language however powerful it is) and my first reply has become more and more apparent (getting worse actually). it has no practical use for me... if you've made it clear since your first post, i would've not make the first post, not even a glimpse. but the ship had sail, this thread will be a curse for me appearing whenever i hit "Show new replies to your posts". tldr.... only one thing i ask from you.... please make it short! pleaseeee! (ps: i have another 1 or 2 never ending cursed threads going on that keep posting no practical knowledge to me)... cheers.

[Sherlock Holmes]

...snip..
So assembler is a language then, good we agree on that. Now can the C language do anything that assembly language can't? You want to talk about languages, lets talk...

in the end... your brainf*ck language will get eventually be translated into machine codes anyway... that turing complete machine?... like all other programming languages in existance. you want bit type and string support? go for it! it will get introduced into bloatness of machine codes anyway... like most of the rest (Basic, Pascal etc)... and sorry, i'm not further discussing since its clear what you re after now (brainf*ck language however powerful it is) and my first reply has become more and more apparent (getting worse actually). it has no practical use for me... if you've made it clear since your first post, i would've not make the first post, not even a glimpse. but the ship had sail, this thread will be a curse for me appearing whenever i hit "Show new replies to your posts". tldr.... only one thing i ask from you.... please make it short! pleaseeee!

Just let me know when you'd like to continue actually discussing programming languages, I'll be here.

[Mechatrommer]

i'll be here! i thought you are the one asking for discussion? instead of posting proposal thats becoming more and more resembling Brainfuck or just talk like...



at least you can relieve big stress out of people...

[Sherlock Holmes]

i'll be here! i thought you are the one asking for discussion? instead of posting proposal thats becoming more and more resembling Brainfuck or just talk like...



at least you can relieve big stress out of people...

I asked for suggestions about things, features people would like to see in a new language if we were designing a new language and we wanted that language to help with development on MCUs.

A few people in the thread have understood that and politely contributed, but some others have not and prefer to object to the sheer effrontery in even hinting that anything better than C is possible.

[Mechatrommer]

..but some others have not and prefer to object to the sheer effrontery in even hinting that anything better than C is possible.

maybe for your own good? certainly anything is not impossible, we just open to that, but we dont know how.. since you think you can do it, just prove it in short possible time, not suggesting far more worse example from your wonderland delusion.

remember. when you are done, you are going to do a great length documenting it to make people understand. and thats where peer reviews/discussion will actually take place... and know how many takers and nayers you are going to get.. that number will means something... at this stage.. we just a bunch of blocking stone to you (we tried not to be though), some are just drawing their honest opinions, even if you dont like it.

[Sherlock Holmes]

..but some others have not and prefer to object to the sheer effrontery in even hinting that anything better than C is possible.

maybe for your own good? certainly anything is not impossible, we just open to that, but we dont know how.. since you think you can do it, just prove it in short possible time, not suggesting far more worse example from your wonderland delusion.

remember. when you are done, you are going to do a great length documenting it to make people understand. and thats where peer reviews/discussion will actually take place... and know how many takers and nayers you are going to get.. that number will means something... at this stage.. we just a bunch of blocking stone to you (we tried not to be though), some are just drawing their honest opinions, even if you dont like it.

"Do" what? I'm seeking to discuss the subject not bicker, I'll progress this at my own speed, I've got nothing to prove to you or anyone else, I'm not seeking to discuss me, my abilities, how fast I work, or any of that stuff, you want to talk about me all the time, I'm done interacting with you, have a nice day.

[MIS42N]

Or tell the compiler #operator async disable. Since the legacy code didn't use async as a reserved word, this works.

My background was programming in multiple languages cobol fortran c basic various job control languages and a pile of assemblers ibm 360, icl 1900, pdp 8, pdp 11, 8086, 6502, z80, pic 8 bit, avr 8 bit and a few more that I've forgotten. Language is pretty much irrelevant, like human languages. Can express a concept in any of them. For fun I've written programs using subleq, just to satisfy this thought. It doesn't take a lot to come up with a set of macros that look like instructions in a higher language, then use them.

So I am much more interested in what the proposed language can do rather than how it is expressed. Good documentation offsets most of the obfuscation present in any language, and more importantly, what the programmer was trying to achieve (which often is not obvious from the code regardless of how clear the language). Let's move on from this syntax discussion. Let the OP have a language with no reserved words, we can live with that anomaly even if we don't want it. Finish the entree, move on to the main meal.

[tggzzz]

My background was programming in multiple languages cobol fortran c basic various job control languages and a pile of assemblers ibm 360, icl 1900, pdp 8, pdp 11, 8086, 6502, z80, pic 8 bit, avr 8 bit and a few more that I've forgotten. Language is pretty much irrelevant, like human languages. Can express a concept in any of them. For fun I've written programs using subleq, just to satisfy this thought. It doesn't take a lot to come up with a set of macros that look like instructions in a higher language, then use them.

Ditto, and yes.

Plus the first language I used professionally was such a macro language cross-compiled on a PDP-11 for a 6800 target.

So I am much more interested in what the proposed language can do rather than how it is expressed.
...
Let's move on from this syntax discussion.

Exactly, but the OP has repeatedly ignored those points in favour of discussing grammars, parse trees and similar compiler constructs.

That's fine, but won't get any further than people wanting to build their own processor.

[MIS42N]

Interpreted languages and LLVM are OK to a point. But when dealing with time constrained hardware problems with unpredictable inputs it just doesn't work. A few years ago I wrote a top octave generator using an Arduino nano (16MHz clock, ATMega328 CPU). It simultaneously produced 12 square waves on 12 pins with frequencies from 4435Hz to 8368Hz with transitions to the nearest 500ns. It would be interesting to see what processor could achieve that with an interpreter. Or even with a compiler. The object code was huge but in the main generated by assembler macros so the source was not so big.

I use interpreted language to test modules. Quick development, plenty of ways to generate input and test output. But then hand translate to the target language for compilation. I have done the reverse, hand translate from IBM assembler to C because nobody else could support an IBM assembler program. Another case of it doen't matter what language, can express an intent in any (as long as there's not timing constraints).

[DiTBho]

A few years ago I wrote a top octave generator using an Arduino nano (16MHz clock, ATMega328 CPU). It simultaneously produced 12 square waves on 12 pins with frequencies from 4435Hz to 8368Hz with transitions to the nearest 500ns. It would be interesting to see what processor could achieve that with an interpreter. Or even with a compiler. The object code was huge but in the main generated by assembler macros

So, you wrote it in assembly   

Some mpus have hw helpers.
Special programmable circuits that only need glue code.

my-c does not optimize the generated assembly code
It's a problem if you need performance.
But it's not its purpose.

[Sherlock Holmes]

Or tell the compiler #operator async disable. Since the legacy code didn't use async as a reserved word, this works.

My background was programming in multiple languages cobol fortran c basic various job control languages and a pile of assemblers ibm 360, icl 1900, pdp 8, pdp 11, 8086, 6502, z80, pic 8 bit, avr 8 bit and a few more that I've forgotten. Language is pretty much irrelevant, like human languages. Can express a concept in any of them. For fun I've written programs using subleq, just to satisfy this thought. It doesn't take a lot to come up with a set of macros that look like instructions in a higher language, then use them.

So I am much more interested in what the proposed language can do rather than how it is expressed. Good documentation offsets most of the obfuscation present in any language, and more importantly, what the programmer was trying to achieve (which often is not obvious from the code regardless of how clear the language). Let's move on from this syntax discussion. Let the OP have a language with no reserved words, we can live with that anomaly even if we don't want it. Finish the entree, move on to the main meal.

I'm not proposing a language, go on, check all the posts for yourself. You can't talk about language as distinct from "how it is expressed", that IS LANGUAGE. We can't even discuss languages without using a language, anyway, I'm tired, good night.

[Mechatrommer]

Most CPU architectures include instructions for arithmetic shift, logical shift and rotates, yet C that "designed for hardware" language doesn't support most of these variants, don't you get it? it was never designed for the applications you keep claiming it was designed for, it wasn't even designed if I can be blunt.

As already touched upon, by some in this thread.  You may have to be careful, if you support, too low a programming level, in a programming language.  It could end up being counter-productive.

Mr Ritchie probably has figured it out decades ago, so by realizing this... there is escape plan to wash C's hands off... assembly-program-in-c so its impossible in C, only whenever the assembly/machine code make it impossible (unavailable) including THAT arithmetic/logical shift... this is also probably why the C got the bashing of having so many different standards across platforms... want to invent a unified language for all at the very low level? good luck!

[DiTBho]

And don't forget tr-mem!
It's something rare at the moment ( POWER10 had it, how many users do use GNU/Linux POWER10? ), but I think it will be the future in multi-core

Special atomic memory model meeded

[Sherlock Holmes]

Most CPU architectures include instructions for arithmetic shift, logical shift and rotates, yet C that "designed for hardware" language doesn't support most of these variants, don't you get it? it was never designed for the applications you keep claiming it was designed for, it wasn't even designed if I can be blunt.

As already touched upon, by some in this thread.  You may have to be careful, if you support, too low a programming level, in a programming language.  It could end up being counter-productive.

Let's say, a very low level instruction set, is included as part of the high level language, but it is TOO low a level.

So it might include (just an example, other instructions could have similar, unintended consequences, this instruction has already been mentioned in this thread).
A subtract (assembly like) instruction as part of the high level language, complete with a functionally defined borrow flag, over-flow detection and maybe some other bits and pieces, such as a zero (all zeros) flag.

One or more actual CPUs the new programming language targets, may have this very instruction, so that is fine.

But any other supported architectures (CPUs), may have significantly different subtract instructions.  They might use a carry, rather than a borrow flag, and its definitions could vary.  There might not be an overflow flag, or the way it works could vary.  The zero flag, may be set under different conditions, e.g. index registers may or may not, affect the zero flag, etc.

So, in order to implement the new programming language, and have consistent results, between differing architectures.  What would have been a single subtract (machine code) instruction, in one of the supported architectures.  Has to be replaced, by lots of instructions, just to emulate the functionality, of a different architecture.

So, for the original/primary supported archetecture, the following program, might be replaced with a single (fast) instruction.

a = a - 2;
Becomes, after compiling:
Subtract-with-borrow 2

But on other archetectures, with completely differing subtract machine code instructions and flag setups.  It might look like this:
a = a - 2;
becomes, after compiling:
clear-carry-flag
Subtract-with-carry 2;
Instructions, including branches to invert the carry flag
Instructions to calculate the overflow status
Other instructions to emulate the original subtract instruction, shown above, in the first example
Etc.

So instead of being a fast, single instruction (such as SUBB), it becomes a horrible mess of 10 or 20 instructions to emulate a differing architecture.

So, the C language (I suspect, but I'm NOT sure of the actual reasons), by NOT going too low a level on purpose, avoids having to do lots of messing around and slowing things, because of all the extra instructions needed to fix the compatibility issues, between architectures.

As long as some method of putting in custom assembly code and/or intrinsics, is included in a new language.  It should not be necessary, to go too overboard, with low level functionality.  Because the programmers, can just put in assembly code or similar, as and where necessary.

I.e. C seems to stick to things, which is relatively common, on just about all architectures, rather than risk taking too many assumptions and/or having to do (potentially slow) fixes, to recreate the compatibility.

You raise some good points. But your example using a = a - 2; is something done routinely, even in C. You'd need to show a better example of an abstract idea that's "low level" and that C doesn't support, that you consider would be too low level.

C does support embedded assembly language though, can one get any lower level than that?

[Sherlock Holmes]

A few years ago I wrote a top octave generator using an Arduino nano (16MHz clock, ATMega328 CPU). It simultaneously produced 12 square waves on 12 pins with frequencies from 4435Hz to 8368Hz with transitions to the nearest 500ns. It would be interesting to see what processor could achieve that with an interpreter. Or even with a compiler. The object code was huge but in the main generated by assembler macros

So, you wrote it in assembly   

Some mpus have hw helpers.
Special programmable circuits that only need glue code.

my-c does not optimize the generated assembly code
It's a problem if you need performance.
But it's not its purpose.

What is "my-c"? as written it sounds like "mice" or maybe "Mike".

[Mechatrommer]

C does support embedded assembly language though, can one go any lower?

what kind of nonsense question is that? the plain answer for beginner is no. if yes, please elaborate. you may prove i'm a beginner as well! or maybe we are using different dictionary here? if you keep posting nonsensical thing like this and before, dont be surprised this thread will become amusement center, which from experience, will eventually get locked down, you dont want it to happen do you?

[DiTBho]

What is "my-c"? as written it sounds like "mice" or maybe "Mike".

It's a C compiler I initially wrote to support a MIPS5+ prototype.
Later, I modified it for internal company needs.

[Mechatrommer]

I didn't explain the a = a - 2 section very well....

you keep rephrasing that same basic stuffs, even if you explained it in pseudo/general idea earlier, anybody with basic/intermediate knowledge should understand very well the point you want to make.. you forget he's an expert an experienced programmer, he already know that yet insist on continuing... let him do his stuffs. you cant win!