1. new Eu version, no built-in socks yet
/me notes that
http://www.rapideuphoria.com/refman.htm
doesn't have any sock/ports commands like
http://www.rebol.com/manual/netports.html
does. 
Kat
2. Re: new Eu version, no built-in socks yet
What's an internet server got to do Euphoria ???
3. Re: new Eu version, no built-in socks yet
What's an internet server got to do with Euphoria ???
4. Re: new Eu version, no built-in socks yet
-----Original Message-----
From: Bernie Ryan <bwryan at PCOM.NET>
To: EUPHORIA at LISTSERV.MUOHIO.EDU <EUPHORIA at LISTSERV.MUOHIO.EDU>
Date: Friday, November 26 1999 07:02
Subject: Re: new Eu version, no built-in socks yet
|What's an internet server got to do with Euphoria ???
|
Nothing. But what's your question got to do with the desire for socket
support in Eu?
cheers,
Derek Parnell
dparnell @ vic.bigpond.net.au
Melbourne, Australia
5. Re: new Eu version, no built-in socks yet
----- Original Message -----
From: Bernie Ryan <bwryan at PCOM.NET>
To: <EUPHORIA at LISTSERV.MUOHIO.EDU>
Sent: Thursday, November 25, 1999 1:42 PM
Subject: Re: new Eu version, no built-in socks yet
> What's an internet server got to do with Euphoria ???
Not much, why do you ask? Altho i guess if they were supported, one could
write a server, or talk to one, or talk to other puters on an intranet
like a linux Eu app on a linux box connected to the internet, and a win98
box running windoze apps ), linking various Eu apps on the same puter,, or
other such things.
Kat
6. Re: new Eu version, no built-in socks yet
On Thu, 25 Nov 1999 15:09:30 -0600, Kat <KSMiTH at PELL.NET> wrote:
>> What's an internet server got to do with Euphoria ???
>
>Not much, why do you ask? Altho i guess if they were supported, one could
>write a server, or talk to one, or talk to other puters on an intranet
> like a linux Eu app on a linux box connected to the internet, and a win98
>box running windoze apps ), linking various Eu apps on the same puter,, or
>other such things.
>
>Kat
Kat, all these things are already supported (albeit in a beta version),
with srvsckip.ew! There are several examples of socket apps out there, not
the least of which is the internet server by Hawke'. Greg Phillips and
Jesus Conseguera have written some interesting demos, and I've got a beta
version of a pop3 mail server. (Regrets if I've left anyone out).
As for Linux, the API calls will be slightly different, but it shouldn't be
too hard to translate with the proper documentation.
Since both the Windows and 'nix platforms already support sockets through
the standard API's, what would be gained by creating built-ins to do this
in the interpreter?
Brian Jackson
7. Re: new Eu version, no built-in socks yet
>Not much, why do you ask? Altho i guess if they were supported, one could
>write a server, or talk to one, or talk to other puters on an intranet
> like a linux Eu app on a linux box connected to the internet, and a win98
>box running windoze apps ), linking various Eu apps on the same puter,, or
>other such things.
>
>Kat
This sounds like an internet server to me not a language like Euphoria.
This is from the reference that you cited:
Besides REBOL's high level support for protocols, such as HTTP, FTP and
SMTP, REBOL also provides support for low-level port access. Ports can be
used to access files, the console and for networks. This chapter will
concentrate on networking.
Ports are the key to implementing REBOL Internet servers, such as web
servers or chat servers. Schemes control how ports act. Networking is
typically based on the tcp scheme.
8. Re: new Eu version, no built-in socks yet
----- Original Message -----
From: Bernie Ryan <bwryan at PCOM.NET>
To: <EUPHORIA at LISTSERV.MUOHIO.EDU>
Sent: Thursday, November 25, 1999 4:07 PM
Subject: Re: new Eu version, no built-in socks yet
> >Not much, why do you ask? Altho i guess if they were supported, one could
> >write a server, or talk to one, or talk to other puters on an intranet
> > like a linux Eu app on a linux box connected to the internet, and a
win98
> >box running windoze apps ), linking various Eu apps on the same puter,,
or
> >other such things.
> >
> >Kat
>
>
> This sounds like an internet server to me not a language like Euphoria.
>
> This is from the reference that you cited:
>
> Besides REBOL's high level support for protocols, such as HTTP, FTP and
> SMTP, REBOL also provides support for low-level port access. Ports can be
> used to access files, the console and for networks. This chapter will
> concentrate on networking.
> Ports are the key to implementing REBOL Internet servers, such as web
> servers or chat servers. Schemes control how ports act. Networking is
> typically based on the tcp scheme.
Ok, i don't feel like fighting, i'll let it sound like whatever it does to
you. I'd just be repeating what i said at the top of the page anyways. If
you want a common interface between several Eu sessions on one or many
puters, and have it easily transported from Windoze to Linux to Whatever OS,
ports are the way to go. HTTP, FTP, POP, and SMTP, etc are merely
handshaking protocols anyhow.
Kat
9. Re: new Eu version, no built-in socks yet
----- Original Message -----
From: Brian Jackson <bjackson at 2FARGON.HYPERMART.NET>
To: <EUPHORIA at LISTSERV.MUOHIO.EDU>
Sent: Thursday, November 25, 1999 4:10 PM
Subject: Re: new Eu version, no built-in socks yet
> On Thu, 25 Nov 1999 15:09:30 -0600, Kat <KSMiTH at PELL.NET> wrote:
>
> >> What's an internet server got to do with Euphoria ???
> >
> >Not much, why do you ask? Altho i guess if they were supported, one could
> >write a server, or talk to one, or talk to other puters on an intranet
> > like a linux Eu app on a linux box connected to the internet, and a
win98
> >box running windoze apps ), linking various Eu apps on the same puter,,
or
> >other such things.
> >
> >Kat
>
> Kat, all these things are already supported (albeit in a beta version),
> with srvsckip.ew! There are several examples of socket apps out there, not
> the least of which is the internet server by Hawke'. Greg Phillips and
> Jesus Conseguera have written some interesting demos, and I've got a beta
> version of a pop3 mail server. (Regrets if I've left anyone out).
I haven't seen srvsckip.ew, i did a search on the Eu archive page and the
recent contribution page, and it wasn't found. So far Jesus Conseguera's
include file hasn't worked.
> As for Linux, the API calls will be slightly different, but it shouldn't
be
> too hard to translate with the proper documentation.
>
> Since both the Windows and 'nix platforms already support sockets through
> the standard API's, what would be gained by creating built-ins to do this
> in the interpreter?
Speed and common interface.
Kat
10. Re: new Eu version, no built-in socks yet
Kat wrote:
> /me notes that
> http://www.rapideuphoria.com/refman.htm
> doesn't have any sock/ports commands like
> http://www.rebol.com/manual/netports.html
> does.
>
> Kat
Hmm... but that brings along one problem... (or so I think)
Implementing TCP support has its caveats... and the biggest one being
that if the support isn't secure, hackers/crackers/lamers can exploit
whatever programs, or language, for that matter, that uses the
implementations. Bad, bad thing.
And I can already see that the above paragragh is malformed, but to
figure it out, gotta think like me...
All in all, I generally disagree with TCP/port/socks/anything support
_into_ Euphoria. Outside of it offers more control, I would suppose,
especially with those whom release their source code. (See what happend
to Linux? Thats probably my best example)
w00 :)
--"LEVIATHAN"
11. Re: new Eu version, no built-in socks yet
- Posted by Pete Eberlein <xseal at HARBORSIDE.COM>
Nov 25, 1999
-
Last edited Nov 26, 1999
On Thu, 25 Nov 1999 16:48:10 -0600, Kat <KSMiTH at PELL.NET> wrote:
>Ok, i don't feel like fighting, i'll let it sound like whatever it does to
>you. I'd just be repeating what i said at the top of the page anyways. If
>you want a common interface between several Eu sessions on one or many
>puters, and have it easily transported from Windoze to Linux to Whatever
OS,
>ports are the way to go. HTTP, FTP, POP, and SMTP, etc are merely
>handshaking protocols anyhow.
>
>Kat
I wrote a common interface between Eu sessions for Linux a while back in a
little library called RDC (remote dynamic calls.) It is somewhat similar
to Java's RMI (remote method invocation.) Basically you start one program
as the server and register the functions and procedures that can be
called. The client program just gets the routine_ids of the server and
calls them in a fashion similar to call_func and call_proc.
If anyone is interested, I could start the RDC messageboard server program
on my computer again. Only one person connected to it back then... I guess
not too many people have net connections working under Linux.
The server is limited to serving one connection at a time. Unfortunately
the server is also not at all secure. It can be crashed as easily as
sending the wrong number of arguments to a routine. (Rob, error handling
routines, please?)
Pete
12. Re: new Eu version, no built-in socks yet
- Posted by monty <boot_me at GEOCITIES.COM>
Nov 25, 1999
-
Last edited Nov 26, 1999
Hi Pete,
Is there any idea of what happens when multiple people try to access
the server? Was wondering if there is something that says
"server busy" or something like that. Maybe you could have it open another
port if the first is occupied...
Monty
PS
Is there a liink to the current Linux Neil? ...not sure which is current.
On Thu, 25 Nov 1999, you wrote:
> I wrote a common interface between Eu sessions for Linux a while back in a
> little library called RDC (remote dynamic calls.) It is somewhat similar
> to Java's RMI (remote method invocation.) Basically you start one program
> as the server and register the functions and procedures that can be
> called. The client program just gets the routine_ids of the server and
> calls them in a fashion similar to call_func and call_proc.
>
> If anyone is interested, I could start the RDC messageboard server program
> on my computer again. Only one person connected to it back then... I guess
> not too many people have net connections working under Linux.
>
> The server is limited to serving one connection at a time. Unfortunately
> the server is also not at all secure. It can be crashed as easily as
> sending the wrong number of arguments to a routine. (Rob, error handling
> routines, please?)
>
> Pete
13. Re: new Eu version, no built-in socks yet
- Posted by Everett Williams <rett at GVTC.COM>
Nov 25, 1999
-
Last edited Nov 26, 1999
Kat wrote:
>> Since both the Windows and 'nix platforms already support sockets through
>> the standard API's, what would be gained by creating built-ins to do this
>> in the interpreter?
>
>Speed and common interface.
>
>Kat
Somebody on this list...can't remember just who at this juncture... is always
howling on about paradigms (might even have been me). As a paradigm, the
Internet isn't likely to go away soon and Internet functionality through such
things as socks support sure makes the job a lot easier. I wouldn't mind a
call interface to a paradigm that makes this discussion possible, but I would
want the calls to be the same on all platforms. That might mean that
somebody comes up with a little lib like VEL or some free package that runs
on all the platforms that Eu does that would allow all versions of Eu to make
the same calls. Socks or some hook to IPv6 is the only form of IO that is
absolutely going to be found on every machine in the galaxy. Local IO is not
particularly portable and therefore would not be a good candidate for
embedding in a portable Eu, but this one sure looks like an opportunity
for everything from multi-player games to mail programs...to whatever
internet functionality that one can dream of. What fun. If I were going
to embed any IO, this would be it.
Everett L.(Rett) Williams
rett at gvtc.com
14. Re: new Eu version, no built-in socks yet
----- Original Message -----
From: Everett Williams <rett at GVTC.COM>
To: <EUPHORIA at LISTSERV.MUOHIO.EDU>; <KSMiTH at PELL.NET>
Cc: Everett Williams <rett at GVTC.COM>
Sent: Thursday, November 25, 1999 9:27 PM
Subject: Re: new Eu version, no built-in socks yet
> Kat wrote:
>
> >> Since both the Windows and 'nix platforms already support sockets
through
> >> the standard API's, what would be gained by creating built-ins to do
this
> >> in the interpreter?
> >
> >Speed and common interface.
> >
> >Kat
>
> Somebody on this list...can't remember just who at this juncture... is
always
> howling on about paradigms (might even have been me). As a paradigm, the
> Internet isn't likely to go away soon and Internet functionality through
such
> things as socks support sure makes the job a lot easier. I wouldn't mind a
> call interface to a paradigm that makes this discussion possible, but I
would
> want the calls to be the same on all platforms. That might mean that
> somebody comes up with a little lib like VEL or some free package that
runs
> on all the platforms that Eu does that would allow all versions of Eu to
make
> the same calls. Socks or some hook to IPv6 is the only form of IO that is
absolutely going to be found on every machine in the galaxy. Local IO is not
> particularly portable and therefore would not be a good candidate for
> embedding in a portable Eu, but this one sure looks like an opportunity
> for everything from multi-player games to mail programs...to whatever
> internet functionality that one can dream of. What fun. If I were going
> to embed any IO, this would be it.
Eu's lack of this caused a couple people to move from Eu to Rebol. The thing
is, i really need Eu for working code, and only a small part of Rebol's
abilities for ease of connections. Like i've said here before, rather than
making a lot of difficult calls to a big include file to do port i/o, a few
extensions to the existing file commands will do fine. Then an include file
can be made for the different protocols, and interfacing the ports to VEL,
etc. for a graphic html web browser or whatever.
Kat
15. Re: new Eu version, no built-in socks yet
- Posted by Everett Williams <rett at GVTC.COM>
Nov 26, 1999
-
Last edited Nov 27, 1999
Kat wrote:
>Eu's lack of this caused a couple people to move from Eu to Rebol. The thing
>is, i really need Eu for working code, and only a small part of Rebol's
>abilities for ease of connections. Like i've said here before, rather than
>making a lot of difficult calls to a big include file to do port i/o, a few
>extensions to the existing file commands will do fine. Then an include file
>can be made for the different protocols, and interfacing the ports to VEL,
>etc. for a graphic html web browser or whatever.
>
>Kat
I think I'll hire you next time I need someone to pick their way through a
minefield. You seem to have a talent for stepping on all the live ones...and
in a much more endearing manner than myself. If Euphoria had an
easily extendable call interface, maybe using templates or zing! zing!
STRUCTURES, then it might be quite easy to do in REBOL what REBOL
does best and in Euphoria what Euphoria does best. Wouldn't that be a
novel idea for reducing the demands on Rob for changes to Euphoria. Give
us that flexible, extendable call interface along with name spaces and
structures and the rest can be accomplished whereever is best. No more
peeking and poking, no more routine_id and call_c_func(). Just a single,
flexible call interface and some data naming and structuring and Euphoria
can then just coast along on it's beauty forever.
Everett L.(Rett) Williams
rett at gvtc.com
16. Re: new Eu version, no built-in socks yet
- Posted by Robert Craig <rds at ATTCANADA.NET>
Nov 26, 1999
-
Last edited Nov 27, 1999
Kat writes:
> Eu's lack of this caused a couple people to move from
> Eu to Rebol.
I guess they don't mind that the Rebol interpreter is
44 times slower than Euphoria (factorial benchmark).
> The thing is, i really need Eu for working code,
> and only a small part of Rebol's abilities for ease of connections.
I certainly agree that Euphoria should have better
Internet capabilities. Another one for the wish list...
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
17. Re: new Eu version, no built-in socks yet
Robert Craig wrote:
>Kat writes:
>> Eu's lack of this caused a couple people to move from
>> Eu to Rebol.
>
>I guess they don't mind that the Rebol interpreter is
>44 times slower than Euphoria (factorial benchmark).
With an Internet that loafs along at something less than 56k baud for
most of us, Rebol's speed is irrelevant if it is only used for the Internet
part of what I am doing.
>> The thing is, i really need Eu for working code,
>> and only a small part of Rebol's abilities for ease of connections.
>
>I certainly agree that Euphoria should have better
>Internet capabilities. Another one for the wish list...
>
>Regards,
> Rob Craig
> Rapid Deployment Software
> http://www.RapidEuphoria.com
In reply to Brian Jackson's wish list....
Brian Jackson writes: > Does Euphoria support OLE/COM/ActiveX > (or whatever
it's called THIS week)?
Euphoria lets you access the WIN32 API at a low level. I guess in theory that
lets you do almost anything, but there's currently no high-level support for
OLE/COM/ActiveX
> If not, are there plans to incorporate it into
> EXW.EXE anytime soon?
I wasn't planning anything, but if enough people lobby for it and/or explain
how it should be done, I'm willing to consider it.
Regards, Rob Craig Rapid Deployment Software
http://www.RapidEuphoria.com
As much as I would like to fight against it and use CORBA instead, it
appears that Billy Boy has triumphed again with COM even showing up
in Linux. As for OLE, it seems a bit premature when we don't even have
a generalized call interface or any form of Eu modularity to be looking at
support for a sophisticated and completely Windows based form of
inter-process communication like OLE. But, it would be useful!
On, Active/X, I draw the line in the sand. that particular piece of
functionality
is responsible for 90 percent of the security problems of IE. Why would we
want it for Euphoria? In addition, Active/X is a completely Windows centric
format. Ugh! Why not Euphoria as CGI, or Euphoria as JavaScript
replacement, or Euphoria as JavaBean generator instead.
Everett L.(Rett) Williams
rett at gvtc.com
18. Re: new Eu version, no built-in socks yet
----- Original Message -----
From: Everett Williams <rett at GVTC.COM>
To: <EUPHORIA at LISTSERV.MUOHIO.EDU>
Sent: Saturday, November 27, 1999 11:48 AM
Subject: Re: new Eu version, no built-in socks yet
>
> On, Active/X, I draw the line in the sand. that particular piece of
functionality
> is responsible for 90 percent of the security problems of IE. Why would we
> want it for Euphoria? In addition, Active/X is a completely Windows
centric
> format. Ugh! Why not Euphoria as CGI, or Euphoria as JavaScript
> replacement, or Euphoria as JavaBean generator instead.
I don't allow ActiveX to run on my puters, and every day i see Java that
doesn't run right.
Kat
19. Re: new Eu version, no built-in socks yet
- Posted by Everett Williams <rett at GVTC.COM>
Nov 27, 1999
-
Last edited Nov 28, 1999
Kat wrote:
>I don't allow ActiveX to run on my puters, and every day i see Java that
>doesn't run right.
>
>Kat
I'm no fan of Java as a language, but the bytecodes and the sandbox are
worthy paradigms that should be the way the world heads. JavaBeans and
Enterprise JavaBeans along with JINI which is java-centric are the only
anodyne to what MS is pitching to the WWW and they appear to be much
more stable than straight Java code. Our worthy language author keeps
touting the speed of Euphoria against VB, Perl, Python, and Rebol. That
comparison will hold a lot more water when the functionality of Euphoria
approaches that of those languages. Granted, once one gets to Socks or
IPv8, one has to factor out the world wide wait when doing comparisons,
but that is not an impossible task. Also, three of those cost less than
the registered version of Euphoria. That is not a dig against Euphoria. That
is a note that says that one cannot compare apples with oranges. Euphoria
can play with those guys and beat them either with internal functionality
or easy access to external functionality. My vote would be for a little of
both. Direct, internal, portable Internet access would give the language a
huge boost. Most other functionality can be had with portable libraries or
masking routines shipped with product or easily available from other sources.
The one other item that I see holding that particular scenario back is the
lack of a powerful string handling facility either internally or as an embedded
set of functions. I repeat what I have said before. String data is not numeric
data except at the grossest level. It is metadata with a grammar of it's
own. The fact that I can add 256 to a[1] when a = "BCD" where B, C, and
D are of a set whose values are limited to 0-255 in the input and output
streams indicates a profound conflict between the Euphoria sequence and
my understanding of text string data. Their contiguity between the quotes
implies the contiguity of their logical boundaries. That contiguity is not
found.
Text is a bit-stream whose subsets correspond to our view of their
representation. The rules for the flow of that bit-stream are entirely external
to Euphoria. I would not recommend that they be internalized...that is a
thing to be accomplished with the algorithmic beauty of Euphoria. All I am
asking for is a recognition within the language of the intrinsic nature of
strings, and some primitives that will speed it's handling. It is a clear subset
of the sequence in it's most general terms, but like so many other things
that are subsets of the sequence, they are not defined nor definable
subsets without huge overhead. Typing will not help here, because of the
overhead, and because it cannot be invoked globally without impeding the
natural flow of the language. Always having to invoke the low-order eight
bits of a word when we are really not supposed to know that it IS a word
is a most frustrating contradiction.
Most structures can be defined, at least logically, as subsets of the
Euphoria sequence, but with no easy translation to external formats and
no paradox free naming conventions, that fact is of little use or comfort.
Everett L.(Rett) Williams
rett at gvtc.com
20. Re: new Eu version, no built-in socks yet
- Posted by Bernie Ryan <bwryan at PCOM.NET>
Nov 27, 1999
-
Last edited Nov 28, 1999
Kat:
If you really want to make Euphoria internet-aware you can download the
inetsdk it's only 46 megs.
( ftp://ftp.clinet.fi/.3/ftp.microsoft.com/developr/MSDN/inetsdk/ )
Then you can write the win32 wrappers to the WININET API protocols
like gopher, FTP, and HTTP without needing any windows sockets code
or TCP/IP management code. There is an article about the API with a CPP
example in August 1997 C/C++ Users journal
21. Re: new Eu version, no built-in socks yet
- Posted by Robert Craig <rds at ATTCANADA.NET>
Nov 27, 1999
-
Last edited Nov 28, 1999
Everett Williams writes:
> Also, three of those cost less than the registered version
> of Euphoria.
When you consider the labor-intensive nature of
programming, and the many hours that it demands,
I can't imagine many professional programmers,
or even novices, who would not be willing to pay
$39 (or $25) to improve their productivity by even a
few percent. Would any software company
switch to the second best tool for a project,
just to save $39? It costs them more than that *per hour*
for each programmer (especially with overhead costs
added in).
> the lack of a powerful string handling facility
What exactly, are you trying to do that can't be easily
done already with a few lines of code?
> All I am asking for is a recognition within the language of
> the intrinsic nature of strings, and some primitives that will
> speed it's handling.
Have you ever written a Euphoria program in which the
lack of speed of string operations was a problem?
I have used languages, such as APL and Basic where
strings and numbers are divided into two separate spaces,
and you must constantly convert back and forth between them.
When I started using C several years ago I realized that
this artificial separation is completely unnecessary and
counter-productive. I am very comfortable with the fact
that strings are not a special object in Euphoria, requiring
special built-in routines, and special conversion functions.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
22. Re: new Eu version, no built-in socks yet
Mr. Williams,
Your views on Euphoria, and programming in general are well thought and
insightful.
To the casual observer, however, it would seem that you spend more time
discussing
the virtues of other programming languages than talking about Euphoria itself.
Perhaps, instead of talking about what Euphoria should have built into it, you
should
sit down and write some libraries and routines that do such things?
Of course, some things, such as Java-like bytecodes, the sandbox, etc. have to
be
built in to the interpreter itself. In which case, why not dust off the old C
compiler and write your own language?
Suggesting changes is one thing.
Suggesting them time and time again becomes preaching.
Respectfully,
Greg Phillips.
Everett Williams wrote:
> Kat wrote:
>
> >I don't allow ActiveX to run on my puters, and every day i see Java that
> >doesn't run right.
> >
> >Kat
>
> I'm no fan of Java as a language, but the bytecodes and the sandbox are
> worthy paradigms that should be the way the world heads. JavaBeans and
> Enterprise JavaBeans along with JINI which is java-centric are the only
> anodyne to what MS is pitching to the WWW and they appear to be much
> more stable than straight Java code. Our worthy language author keeps
> touting the speed of Euphoria against VB, Perl, Python, and Rebol. That
> comparison will hold a lot more water when the functionality of Euphoria
> approaches that of those languages. Granted, once one gets to Socks or
> IPv8, one has to factor out the world wide wait when doing comparisons,
> but that is not an impossible task. Also, three of those cost less than
> the registered version of Euphoria. That is not a dig against Euphoria. That
> is a note that says that one cannot compare apples with oranges. Euphoria
> can play with those guys and beat them either with internal functionality
> or easy access to external functionality. My vote would be for a little of
> both. Direct, internal, portable Internet access would give the language a
> huge boost. Most other functionality can be had with portable libraries or
> masking routines shipped with product or easily available from other sources.
>
> The one other item that I see holding that particular scenario back is the
> lack of a powerful string handling facility either internally or as an
> embedded
> set of functions. I repeat what I have said before. String data is not numeric
> data except at the grossest level. It is metadata with a grammar of it's
> own. The fact that I can add 256 to a[1] when a = "BCD" where B, C, and
> D are of a set whose values are limited to 0-255 in the input and output
> streams indicates a profound conflict between the Euphoria sequence and
> my understanding of text string data. Their contiguity between the quotes
> implies the contiguity of their logical boundaries. That contiguity is not
> found.
> Text is a bit-stream whose subsets correspond to our view of their
> representation. The rules for the flow of that bit-stream are entirely
> external
> to Euphoria. I would not recommend that they be internalized...that is a
> thing to be accomplished with the algorithmic beauty of Euphoria. All I am
> asking for is a recognition within the language of the intrinsic nature of
> strings, and some primitives that will speed it's handling. It is a clear
> subset
> of the sequence in it's most general terms, but like so many other things
> that are subsets of the sequence, they are not defined nor definable
> subsets without huge overhead. Typing will not help here, because of the
> overhead, and because it cannot be invoked globally without impeding the
> natural flow of the language. Always having to invoke the low-order eight
> bits of a word when we are really not supposed to know that it IS a word
> is a most frustrating contradiction.
>
> Most structures can be defined, at least logically, as subsets of the
> Euphoria sequence, but with no easy translation to external formats and
> no paradox free naming conventions, that fact is of little use or comfort.
>
> Everett L.(Rett) Williams
> rett at gvtc.com
23. Re: new Eu version, no built-in socks yet
> Everett Williams wrote:
>
> > The one other item that I see holding that particular scenario back is the
> > lack of a powerful string handling facility either internally or as an
> > embedded
> > set of functions. I repeat what I have said before. String data is not
> > numeric
> > data except at the grossest level. It is metadata with a grammar of it's
> > own. The fact that I can add 256 to a[1] when a = "BCD" where B, C, and
> > D are of a set whose values are limited to 0-255 in the input and output
> > streams indicates a profound conflict between the Euphoria sequence and
> > my understanding of text string data. Their contiguity between the quotes
> > implies the contiguity of their logical boundaries. That contiguity is not
> > found.
> > Text is a bit-stream whose subsets correspond to our view of their
> > representation. The rules for the flow of that bit-stream are entirely
> > external
> > to Euphoria. I would not recommend that they be internalized...that is a
> > thing to be accomplished with the algorithmic beauty of Euphoria. All I am
> > asking for is a recognition within the language of the intrinsic nature of
> > strings, and some primitives that will speed it's handling. It is a clear
> > subset
> > of the sequence in it's most general terms, but like so many other things
> > that are subsets of the sequence, they are not defined nor definable
> > subsets without huge overhead. Typing will not help here, because of the
> > overhead, and because it cannot be invoked globally without impeding the
> > natural flow of the language. Always having to invoke the low-order eight
> > bits of a word when we are really not supposed to know that it IS a word
> > is a most frustrating contradiction.
Many of your comments are well taken.. This one, however, I cannot understand.
Can you give an example (code, plz) of where it is necessary to "invoke thte
low-order 8 bits", etc. when manipulating strings in any useful way?
Or which useful string functions from other languages are missing from Euphoria?
Regards,
Irv
