Re: Is Phix the new de facto standard for Eu programmers?
- Posted by katsmeow Jan 30, 2021
- 2178 views
I am going to snip text here and there (in the interest of post size) and i respectfully hope i do not mess up the intended messages...
But it's all about not taking big steps to improve programming with OE, it's all about NOT making more features possible?
I literally never said this. And this summary of my statement couldn't be more untrue. I want to see a lot more things come to Euphoria.
And then you limited the expansion of OE to the usual stuff in almost all languages....
Things I'd like to see in Euphoria include threads, classes, and a built in compiler, all of which Phix now has.
I am not against all that, i'd like to see a bit of it too, i was very happy to see Matt's OOEU, and i did play with the early Bach versions. If you remember, around that time frame, i was experimenting with swapping code in the form of DLLs in and out of running programs by using load/unload, and found a huge memory leak in the process. Using DLLs had other problems, and no one was in favor of the feature. I also recently acknowledged this...
But I have entirely different ideas on how they should behave or be implemented within Euphoria itself
as one of the perfectly valid reasons OE stopped growing. I later edited my post away.
I think you've misinterpreted my "dialect" analogy. Spoken languages grow and evolve over time..... And programming languages are really no different than that...
But OE isn't growing, changing, evolving, splintering, except for Phix. And i applaud Phix, but i cannot help but fear it will be constrained to one incremental step of growth.
I have great aspirations .....
Like i deleted the DLL swapping code years ago, and most of irc.e years before that, a couple days ago i deleted all the globally sharing code. It's big step stuff OE could do, regardless of what features it has in common with other languages. I don't want to compare OE/Phix to other current languages, but there's things i could do with computers 30+ years ago that i cannot do in any language on any modern computer!
I do plenty of "computering" elsewhere but unfortunately I find myself in a continuous Catch-22 paradox: I cannot use Euphoria for my work projects because it is not modern enough for those purposes,
I meant "computering" with and within OE itself. At least catching up to OOEU.
Today, at least, fire is used to make car engines run, steam turbines make electricity. But OE cannot really do any more to munge data today than was done in a Vic20 or C64 or ZX80 or Apple 40 years ago.
I completely disagree with this statement. Firstly, because data munging hasn't really changed in the entire history of electronic computing.
Ok, how many modern languages allow you to stop the program, fiddle around in the variables, edit some source code, and restart at any point in the program (with caveats) as if it was never stopped? How many allow you to swap source code while running, like i did with DLLs? How many allow the program to build a list of the variables and then look at the values? How many languages come ready to communicate to the world (at the table-top level) the physical computer? I did all that and more in the early 1980's, and cannot with a modern language, so i must stand by what i said.
Aside from other branches of computer science like quantum computing and machine learning, all the data we ever process is still just some series of bytes and bits and I think Euphoria is still very well suited to that type of work.....
Quantum computing is a physical process inherently multi-threaded, but OE cannot handle threads. And OE cannot know what it does and does not know, so machine learning is out. What i did with global vars had one huge gotcha: it was very convoluted and buggy to get the entire program (or series of programs) to use new proceedures added during runtime (this was also a problem with swapping DLLS).
And second, because I have recently implemented several new tools for processing "modern" forms of data, such as my JSON, HTML, and template parsers.
You do not think form manipulation and parsing was done 40+ years ago? Did you forget i had added a SGML parser to a strtok version many years ago? Perhaps you remember the overt fighting it caused, over the use of "strings", "tokens" (words), thereby causing the destruction of the basic concept of Eu itself. In 1990 i began converting a dictionary to be an XML table, and it was years later before i got online with win95B.
I can agree with you, Greg, that having modern features in OE would be helpful, despite those who say OE should be as simplistic as RobC intended Euphoria to be in version 1.0 some 20+ years ago. I have said OE needs bigger steps beyond other languages. What you and i have in common, in terms of modernising OE, is that the current users of OE have no need of any new features. Do you realise that when i put out strtok that no mainstream language had that library, and they added it quickly, but it was years later OE officially added the functionality to the language?
I respect your contributions and efforts Greg, and ditto to Pete and Phix, and Matt-Derek-Jiri-DavidC-etc, but while playing catch-up to other languages, can you also please leapfrog them? And by "leapfrog" i mean also add the functions (by the same or different means) enjoyed by programmers 40 years ago.
To wrap this up, it's fine with me if Phix is a standard, but "standard" implies "static", and OE is already static.
Kat
