1. I don't like the intro page of the documentation.

• Posted by _tom (admin) Jun 05, 2013
• 3260 views

Here is a suggested change to the first page of the documentation:

Accelerate your programming.

Euphoria programs run faster than conventional interpreted languages; Euphoria programs can be even faster after compiling. The power and simplicity of Euphoria makes rapid program development a reality.

Accommodate you needs.

Euphoria gives you flexibility and power. Euphoria is a general purpose, versatile, multi-gui, portable, and multi-platform. Develop programs quickly using the interpreter. Compile programs and produce independent applications for distribution. Write at a high level and yet interface to C. Unique simplified data handling requires only atoms/sequences. The standard library provides plenty of utility. Euphoria speed does not compromise subscript checking, uninitialized variable checking, garbage collection, and numerous other run-time checks.

Accomplish more.

Euphoria is small and friendly: easy to read, easy to learn, and easy to use. Atoms/sequences are higher level and easier to use than conventional data-types. Write code the way you want with simple English words and freeform layout. Euphoria is responsive and gives useful error messages when you need them. Euphoria is supported with GUI libraries, tracing, testing, documentation tools, editing choices, sample programs, and a helpful forum. Euphoria is free and open source and actively being developed.

Yes, it is true.

Euphoria users are understandably enthusiastic. With little effort you can confirm, for yourself, that Euphoria is an exceptional programming language. To learn more about Euphoria just try it out--download one 12MB file, install quickly on Windows, Unix (Linux, BSD, ...), or OS X, and get a ready to use programming environment with documentation and sample programs.

• Benchmarking to show that Euphoria programs execute faster than Python, Perl, Ruby, and PHP.
• Examples to illustrate the flexibility of Euphoria.
• Easy-to-learn and easy-to-program illustrated; a language that is simple, small, consistent, and predictable.

I welcome your ideas

Tom

3. Re: I don't like the intro page of the documentation.

• Posted by DerekParnell (admin) Jun 05, 2013
• 3182 views

Great job. I like the feel of it. It is active, straight-forward and enthusiastic.

However, because the data-type situation is very different to most anything else that existing programmers have dealt with, you might want to elaborate a bit more ...

5. Re: I don't like the intro page of the documentation.

• Posted by jaygade Jun 06, 2013
• 3179 views

I like that. I disagree with the language being "statically-typed" as Derek said it was.

Objects are dynamically typed. Atoms and sequences represent values and lists or strings, respectively. I think that useless nailed it with some minor changes:

Euphoria combines the flexibility and rapid-development of dynamically-typed languages with the reliability and accuracy of statically-typed languages.

There are three main^* built-in types: atoms for representing single values such as characters, integers, or real numbers; sequences for strings, lists of values, and even lists of lists; and dynamically-typed objects which can represent any value or list. Users requiring stricter type-checking can define their own types ensuring that bugs get caught earlier in development.

7. Re: I don't like the intro page of the documentation.

• Posted by mattlewis (admin) Jun 06, 2013
• 3142 views

procedure FOO(integer a, sequence b) 
   ---  whatever --- 
end procedure 
 
FOO( "abc", 3) 

We do this in some cases. Right now, I think it's only for inlined routines. When I run your code on 4.0.0 (or later), I get:

$ eui405 bug.ex 
/home/matt/eu/test/bug.ex:4 
<0146>:: Type Check Error when inlining literal 
FOO( "abc", 3) 
             ^ 

Matt

9. Re: I don't like the intro page of the documentation.

• Posted by _tom (admin) Jun 06, 2013
• 3108 views

A one-of-a-kind language.

Euphoria has a unique data-type design. Three, easy to learn, data-types achieve what in a conventional language requires many, harder to learn, concepts. An atom is any single datum (number, integer, character). A sequence is any collective data (atom, sequence, object, nested sequence, string, list, array, stack, tree, ... ). An object is dynamically either an atom or sequence. Specify integer to gain performance. Create your own user-types to ensure strict type-checking to catch bugs. Euphoria is a remarkable blend of static and dynamic typing.

11. Re: I don't like the intro page of the documentation.

• Posted by useless_ Jun 06, 2013
• 3077 views

There's too much punctuation. It's not correct to drop in commas after every modifier or phrase.

And it isn't unique, but i don't remember all the languages with sequences, etc. Python has sequences, another language about the time RDS built Eu had sequences. I don't remember enough.

useless

13. Re: I don't like the intro page of the documentation.

• Posted by jaygade Jun 06, 2013
• 3053 views

Right, but I think that you could argue that the datatype design as a whole is unique, even though it contains features from other languages.

Lisp has atoms and lists and of course lists can contain atoms or other lists or even functions (which are just lists or atoms themselves), etc., but it's not a procedural language. I don't know if you can define user-types with boundary checking and error conditions, although apparently understanding Lisp is one of those things that elevates you above mere mortals.

Isn't using Java's Vectors and C#'s Lists way more complicated as well? Can they easily contain arbitrary mixed data like sequences do?

lua has sequences too, but their type system is more complicated than Euphoria's I think.

I don't know much about Icon, although a quick scan of the Wikipedia page shows it has arbitrary lists.

Maybe "unique" is a little strong.

15. Re: I don't like the intro page of the documentation.

• Posted by DerekParnell (admin) Jun 06, 2013
• 2997 views

Oops ... I'm sorry to have started this tweaking, Tom. Anyhow, with Kat's thoughts in mind, how's this tweak ...

A one-of-a-kind language.

Euphoria has a simple data-type design. Its three built-in data-types achieve what in a conventional language requires many, harder to learn concepts.
An object can dynamically hold either an atom or a sequence. An atom can hold a single number (floating point or integer). A sequence is a variable length array of objects. Together, these can be used to build more complex types (eg. string, list, stack, tree, ... ). You can also use integer as a special form of atom to gain performance. Additionally you can define your own user-types. Euphoria is a remarkable blend of static and dynamic typing.

17. Re: I don't like the intro page of the documentation.

• Posted by _tom (admin) Jun 06, 2013
• 3038 views

Kat has a sense of what is needed for the first page of the documentation.

I expected some comments (grunts from trolls) about my blatant use of the letter "a" to start each paragraph--that's my attempt a PR style writing.

I can find nearly identical descriptions of all languages. They are all fast, easy to learn, powerful, easy to read, ..., and more of the same. There is never a "page two" that justifies any claims.

"built-in data-types"

• that clarifies the following reference to "user-types"

"array of objects"

• "array" is an implementation of a mathematical matrix (as in Fortran)
• "objects" can unfortunately be confused with OOP

"An object can dynamically hold either an atom or a sequence."

• makes sense to an existing Euphoria user, but how do I explain this to someone new to Euphoria?

single datum vs compound data

• I am searching for a nice distinction between atom and sequence
• any ideas?

Ancients (the really old ones) had a math system based on one, two, and many.

Therefore two (the largest number available) is reason describe Euphoria as having two data-types: atom and sequence. An object is not a distinctive data-type, only an "either or" of the two fundamental types. This makes the description of Euphoria smaller and simpler.

An "integer" is just an appendix; if integers did not exist then Euphoria programs would still run. Integers for efficiency make sense to experts but are just noise to a novice. This suggests:

"Use the integer data-type (a special case of the atom) when speed and low memory use are important."

A "list" data-type appears in many languages. A shopping list is natural so that any programming language better be able to make one. Lisp is lists (wonderfully elegant) but not everyone is willing to put up its eccentricity. Lists have been added to many languages but that just increases the number of compound data-types available for that language. Euphoria uses the sequence (aka list) as a fundamental design choice; as best as I can tell sequences make Euphoria unique.

A "primitive" data-type appears in many languages. Typically that means integer, float, boolean, and character, but sometimes ending up with a dozen variations. Euphoria uses the atom and just the atom; as best as I can tell that also makes Euphoria unique.

So an atom-sequence design is unique and special to Euphoria. (Adding object, integer, and user-types results in a practical language.)

Therefore I conclude that: "Euphoria has a unique data-type design."

The atom-sequence is the basis of clean unified design. Operators always work the same ( + always adds, not sometimes concatenates). Sequences always behave the same (same indexing, slicing, ..., for all sequences). In a conventional language things are not so tidy. So again I conclude that: "Euphoria has a unique data-type design."

Remaining features could be described as "industry standard." Loops, conditionals, assignment, and the rest are logical adaptations of features found in all languages.

So, can I stick with "unique" and "one-of-a-kind?"

After I do several re-writes, I will appreciate Kat's final word on fixing my grammar.

19. Re: I don't like the intro page of the documentation.

• Posted by useless_ Jun 06, 2013
• 3039 views

I think most programmers would understand that, but you could say "user defined types" or "you name the data type and what is allowed in it".

I agree with what you say about "array", to me a sequence is 2D, and cannot be 3D, 4D etc.. But it can be nested, it can be reinterpreted.
Can't help with the "object", except to say it's untyped and the real root of Eu's datums.

It's an untyped variable which can hold any other type. A wonderful feature is it can be tested whenever you want to see what type is in it. It's untypedness slows Eu down somewhat, which is why there are atoms and sequences.

RDS explained it nicely, but i'll paraphrase since we have a weasel mascot now: an atom is one of anything that fits in a computer's bite, a sequence is lunch.

If you want to present that much arguement to justify it, even to newbies. If someone is learning a new spoken human language, i'd give them other simpler data before i gave them an 800 page dictionary and an 800 page cultural reference.

useless

21. Re: I don't like the intro page of the documentation.

• Posted by DerekParnell (admin) Jun 06, 2013
• 3035 views

Yeah, I was going to say this too, but I re-read Jay's post again, and he is correct though might benefit from rewording. Jay is saying that assigning integer values to an 'atom' variable takes up no more space than assigning it to an 'integer' variable.

23. Re: I don't like the intro page of the documentation.

• Posted by DerekParnell (admin) Jun 06, 2013
• 2924 views

I'm not disagreeing, Kat.

25. Re: I don't like the intro page of the documentation.

• Posted by mattlewis (admin) Jun 07, 2013
• 3054 views

Yes, this was the main thing I was thinking of, but I think there are other cases where a calculation result would be demoted if stored into an integer variable. For instance:

atom x 
integer y 
object z = 2.5 
x = z - 0.5 
y = z - 0.5 
? x & y & z 

You can perhaps see this happening most easily by translating. The key part is this:

 
    /** bug.ex:7	y = z - 0.5*/ 
    _1y_170 = NewDouble(DBL_PTR(_1z_171)->dbl - DBL_PTR(_9)->dbl); 
    if (!IS_ATOM_INT(_1y_170)) { 
        _1 = (object)(DBL_PTR(_1y_170)->dbl); 
        DeRefDS(_1y_170); 
        _1y_170 = _1; 
    } 

Yes, except that regexes are sequences to begin with, so the only difference is the extra bit of metadata for regex cleanup.

If you store an atom with a delete routine as an integer, the actual atom created will only be stored in a temp, and it will go away because the integer does not hold onto a reference. Simple demo:

procedure its_gone( object o ) 
	printf(1, "cleaning up after %d\n", o ) 
end procedure 
 
integer foo = delete_routine( 1, routine_id("its_gone" ) ) 
puts(1, "after foo...\n") 
 
atom bar = delete_routine( 2, routine_id( "its_gone" ) ) 
puts(1, "after bar...\n" ) 
bar = 0 
puts(1, "bar is gone...\n") 

Matt

27. Re: I don't like the intro page of the documentation.

• Posted by mattlewis (admin) Jun 07, 2013
• 2901 views

None of this is 4.1. It's all in 4.0.

Yes, they are. We try fairly hard to keep things as integers, because they're a lot faster. It's probably a reasonable assumption to make as a normal euphoria coder, but once you start messing with the back end or the translator, you have to be careful about weird edge cases.

Matt

29. Re: I don't like the intro page of the documentation.

• Posted by useless_ Jun 07, 2013
• 2857 views

How about a mouseover or alt text there to say it's a mongoose and why?

Then, if you want to,

Hyperlinks for in-depth reading.
The following isn't showing in creole without the curly braces, consider this my bug notification of the problem:

[quote _tom] 
;Accommodate you needs.  
[/quote] 

_tom said...

Euphoria gives you flexibility and power.

Euphoria is a general purpose, versatile, multi-gui, multi-platform, and portable. Write at a high level and yet interface to C

DLLs if you want

I'd hyperlink each of GUI libraries, tracing, testing, documentation tools, editing choices, sample programs, and a helpful forum. To separate pages, even, for some indepth non-distracting reading of each topic, with examples. The wiki should be used too.

I am not sure how this is laid out on the page, but i suggest two columns here, your multiply in one and a foreach in the other, then below those with your "while" in one, and an equally good "case" demo in the other. I'd then mention the interesting way you can jump into a while loop to emulate "repeat-until" loops, and the controlled "goto" to get you out of loops of any depth or complexity. Some mention should be made of tasks.e multitasking.

zip or rar (and link to current stable version, with the docs)

Looking really good, Tom.

useless

31. Re: I don't like the intro page of the documentation.

• Posted by jaygade Jun 07, 2013
• 2838 views

Create a page on the Wiki for all of this stuff, and discuss it here (with links).