Just out of curiosity: why in Heavens Name are you keeping the seven-bit restriction in the program?

Already in 1999 when I hacked the EXE-file and made it Y2K compliant, I removed that restriction. Those of you who have used my patch from then no longer have a robot that changes hi-ASCII characters to a "?".

So why, why, why? Why does this "new" version have this restriction?

Check out the listing of 2:203/6 to see what little progress we have made for soon to be 15 years. I've had that listing, properly spelled, for many years now, and I'm still just Bj?rn in the nodelist there... :(

--- Mozilla/5.0 (Windows; U; Windows NT 5.1; sv-SE; rv:1.9.1.16) Gecko/20101125
* Origin: news://felten.yi.org (2:203/2)

Hello Björn,

On Monday February 03 2014 16:45, you wrote to All:

Just out of curiosity: why in Heavens Name are you keeping the
seven-bit restriction in the program?

It doesn't. For some time now - on my suggestion - it supports this option:

Allow8bit 1

Check out the listing of 2:203/6 to see what little progress we
have made for soon to be 15 years. I've had that listing, properly spelled, for many years now, and I'm still just Bj?rn in the nodelist there... :(

There is just one little problem. Just as the ASCII only restriction is not fair to the world outside the part where the native US English speakers live, it would not be fair to restrict it to those that can spell their name or city properly in CP437, CP850 or LATIN-1.

I we allow more than ASCII, we should allow ALL Fidonet participants to have their name properly spelled. Unfortunately there is no single 8 bit character set that can fulfil this requirement, and having a different character encoding
scheme for each line is ehh.. unpractical.

The only thing that makes sense is to use UTF-8. That means the ö in you name will be encoded by two bytes in the nodelist.

When I see TWO questions marks in the nodelist for 2:203/6, we can discuss this
again.


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20110320
* Origin: http://www.vlist.eu (2:280/5555)

MvdV> When I see TWO questions marks in the nodelist for 2:203/6, we can
MvdV> discuss this again.

LOL! Yeah, I bet that takes at least ten years of twisting Ward's arm.

--- Mozilla/5.0 (Windows; U; Windows NT 5.1; sv-SE; rv:1.9.1.16) Gecko/20101125
* Origin: news://felten.yi.org (2:203/2)

My previous response aside.

MvdV> It doesn't. For some time now - on my suggestion - it supports this option:

MvdV> Allow8bit 1

I know. But I don't understand why we even need this. As I said, I completely disabled (made a conditional zero jump) that check in my 1999 patch.
Who would even want to *not* allow it? Get rid of all that code completely, for
fucks sake!

MvdV> There is just one little problem. Just as the ASCII only restriction is
MvdV> not fair to the world outside the part where the native US English
MvdV> speakers live, it would not be fair to restrict it to those that can
MvdV> spell their name or city properly in CP437, CP850 or LATIN-1.

Who cares? The nodelist will be totally backward compatible even if we put Chinese letters in it. Only stone-age editors, that tries to pick up recipients
names from the nodelist based on nodenumbers, will possibly break.

The nodelist is nothing but a database that is supposed to be read and interpreted by programs. Not to be read by low-ASCII brainwashed persons.

As nodelist reading programs evolves, so should what we allow into this common database of ours.

MvdV> I we allow more than ASCII, we should allow ALL Fidonet participants to
MvdV> have their name properly spelled. Unfortunately there is no single 8 bit
MvdV> character set that can fulfil this requirement, and having a different
MvdV> character encoding scheme for each line is ehh.. unpractical.

Nah, lets do it! Nobody think that the nodelist will grow bigger any more. So what if it suddenly becomes twice the size (as with every character 16-bit).
Not that it will ever happen, but still, so what?

--- Mozilla/5.0 (Windows; U; Windows NT 5.1; sv-SE; rv:1.9.1.16) Gecko/20101125
* Origin: news://felten.yi.org (2:203/2)

On Mon, 03 Feb 2014, Björn Felten wrote to All:

Just out of curiosity: why in Heavens Name are you keeping the
seven-bit restriction in the program?

did you forget to add the config option to allow 8bit characters? has everyone else also added that config option and set it to enabled? are you running the lastest version 3.4.1??

[quote="whatsnew.txt"]
98
99 ALLOW8bit:
100 ----------
101 ALLOW8bit is a new keyword that will cause MakeNL to allow char-
102 acters above 0x7F in nodelist entries. Normally, such characters
103 are replaced with question marks (?) when the file is processed.
104 This keyword is disabled by default. To enable it, add ALLOW8bit 1
105 to your MAKENL.CTL file.
106
[/quote]

)\/(ark

One of the great tragedies of life is the murder of a beautiful theory by a gang of brutal facts. --Benjamin Franklin

--- FMail/Win32 1.60
* Origin: (1:3634/12.71)

Hello Bj÷rn!

03 Feb 14 16:45, you wrote to All:

Just out of curiosity: why in Heavens Name are you keeping the seven-bit restriction in the program?

Already in 1999 when I hacked the EXE-file and made it Y2K compliant, I removed that restriction. Those of you who have used my patch from then no longer have a robot that changes hi-ASCII characters to a "?".

Just out of curiosity, what is high-ASCII ?

So why, why, why? Why does this "new" version have this
restriction?

Because tere is no consensus on what is to be represented by these so called
high-ASCII.

Check out the listing of 2:203/6 to see what little progress we
have
made for soon to be 15 years. I've had that listing, properly spelled, for many years now, and I'm still just Bj?rn in the nodelist there... :(

It is not the magic on how it is to be done.

The option for the control file to enable 8 bit characters is Allow8Bit and
should then be set to "1".

It was added as an option in v3.4.0 dated October 6, 2013 on popular request
by Michiel van der Vlist and possibly others.

As you will have seen, in many answers that have responded to messages that
you have written there is no single way to write Bj?rn.

Apparently many use what they call a codepage, that defines what their system
presents as high-ASCII.

In messages however, many claim, that there is a CHRS kludge in the message,
that tells you about the used codepage.

That is however of little use if you are generating a messages index and
for speed only read the binary header.

With nodelist there is no CHRS klude, so what is the default.

Now if you could arrange, that the sysops of fidonet come to a consensus
to what high-ASCII is, I am certain that the ZC's will cooperate and
enable the use of 8bit characters at their level.

Apparently I overlooked a message in REGCON.EUR, it seems that Ward has
enabled the option already

But I am still curious, what definition of high-ASCII will be used?

Kees

--- FPD v2.9.040207 GoldED+/LNX 1.1.5
* Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)

Just out of curiosity, what is high-ASCII ?

Usually known as the ASCII characters higher than 0x7f.

Because there is no consensus on what is to be represented by these so called high-ASCII.

No need to. Every region (outside of Z1 and Z3) -- meaning parts of the world using completely different languages -- takes care of it's own. The rest can simply either ignore it or take the recent technical developments and handle it accordingly.

many years now, and I'm still just Bj?rn in the nodelist there... :(

It is not the magic on how it is to be done.

No magic needed. My 15yo Y2K patch can handle it.

The option for the control file to enable 8 bit characters is Allow8Bit

and

should then be set to "1".

Maybe you haven't have time to read my recent comments yet? What I'm asking is, why this option is even needed. And if it is, why is the default to have it
disabled and not vice versa?

you have written there is no single way to write Bj?rn.

But there is! *I* write it the way I want it. In my Region20 file it's written the way I want it. That's simply the way *I* want it. Freedom of expression, anyone...?

But I am still curious, what definition of high-ASCII will be used?

None at all. If we regard the nodelist as the simple, binary file that it was intended to be from the very beginning, we can put whatever characters that
we want in there.

--- Mozilla/5.0 (Windows; U; Windows NT 5.1; sv-SE; rv:1.9.1.16) Gecko/20101125
* Origin: news://felten.yi.org (2:203/2)

Hello Björn,

On Monday February 03 2014 17:24, you wrote to me:

MvdV>> Allow8bit 1

I know. But I don't understand why we even need this.

Because the Fidonet community as a whole isn't ready for it yet. Fortunately those working on the MakeNlNg project were. So be glad that at least it has been made possible.

Some call this process diplomacy. Go in small steps. Don't try to break down the wall in one day. That won't work. First make a door but do not open it yet.
Then make a peep hole to raise the other side's curiosity. Then give them a key
so they can open the door. Small steps...

As I said, I completely disabled (made a conditional zero jump) that
check in my 1999 patch.

Eh.. if you completely disabled it, would that not be an UNconritional jump?

Who would even want to *not* allow it?

Is that a rethorical question?

MvdV>> There is just one little problem. Just as the ASCII only
MvdV>> restriction is not fair to the world outside the part where the
MvdV>> native US English speakers live, it would not be fair to
MvdV>> restrict it to those that can spell their name or city properly
MvdV>> in CP437, CP850 or LATIN-1.

Who cares?

I for one.

The nodelist is nothing but a database that is supposed to be read
and interpreted by programs. Not to be read by low-ASCII brainwashed persons.

So why don't we completely do away with the fields for the system name, the sysop name and the location. They serve no purpose whatsoever in regarding mailers making a connection.

MvdV>> I we allow more than ASCII, we should allow ALL Fidonet
MvdV>> participants to have their name properly spelled. Unfortunately
MvdV>> there is no single 8 bit character set that can fulfil this
MvdV>> requirement, and having a different character encoding scheme
MvdV>> for each line is ehh.. unpractical.

Nah, lets do it!

Let us do what?

Nobody think that the nodelist will grow bigger any more. So what if
it suddenly becomes twice the size (as with every character 16-bit).
Not that it will ever happen, but still, so what?

Indeed so what?

So I suggest you make the next step. Instead of using an 8 bit character set to
encode the ö for your name in the nodelist, encode it in UTF-8. This will show as TWO question marks in the Z2 nodelist.

Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20110320
* Origin: http://www.vlist.eu (2:280/5555)

On Mon, 03 Feb 2014, Björn Felten wrote to Michiel van der Vlist:

My previous response aside.

MvdV> It doesn't. For some time now - on my suggestion - it supports this
MvdV> option:

MvdV> Allow8bit 1

I know. But I don't understand why we even need this. As I said,
I completely disabled (made a conditional zero jump) that check in
my 1999 patch. Who would even want to *not* allow it? Get rid of
all that code completely, for fucks sake!

makenl_ng is used in other FTN networks... some of them do not allow the same things that may be allowed in fidonet or other FTNs... retaining this feature is good for them and their practises... making it an option is the best thing for all...

)\/(ark

One of the great tragedies of life is the murder of a beautiful theory by a gang of brutal facts. --Benjamin Franklin

--- FMail/Win32 1.60
* Origin: (1:3634/12.71)

MvdV> Some call this process diplomacy. Go in small steps. Don't try to break
MvdV> down the wall in one day. That won't work. First make a door but do not
MvdV> open it yet. Then make a peep hole to raise the other side's curiosity.
MvdV> Then give them a key so they can open the door. Small steps...

I know, Michiel. In 1999 I opened the wall. Many sysops adopted my Y2K patch. I told Ward about it over and over again, but he never remembered the previous time.

Now, I opened the wall 15 years ago. I then spent a year or so, some five years ago with this new project. This was some five years ago.

Yes indeed, I know all about the small steps and so on, but I am getting slightly frustrated with the constant brick wall I've been facing for all those
year.

MvdV> Eh.. if you completely disabled it, would that not be an UNconritional
MvdV> jump?

Dunno what that means. I changed JNZ xx to JNZ 00. In my book it's a conditional jump changed from somewhere further down the code into a jump to the next instruction.

Who cares?

MvdV> I for one.

Why?

MvdV> So why don't we completely do away with the fields for the system name,
MvdV> the sysop name and the location. They serve no purpose whatsoever in
MvdV> regarding mailers making a connection.

Depends. Most systems depend on them. But I agree that we could easily let go of the old, outdated FTS specifications, as long as we maintain the comma limited structure.

MvdV>>> I we allow more than ASCII, we should allow ALL Fidonet
MvdV>>> participants to have their name properly spelled.

Nah, lets do it!

MvdV> Let us do what?

What you suggest above.

MvdV> So I suggest you make the next step. Instead of using an 8 bit character
MvdV> set to encode the ö for your name in the nodelist, encode it in UTF-8.
MvdV> This will show as TWO question marks in the Z2 nodelist.

Good suggestion. But I think I'll stay with your one-step suggestion for the
moment. In five or so years, when we have a ZC that can handle the nodelist properly, I'll consider expanding my umlaut project in the nodelist. 8-)

--- Mozilla/5.0 (Windows; U; Windows NT 5.1; sv-SE; rv:1.9.1.16) Gecko/20101125
* Origin: news://felten.yi.org (2:203/2)

Hello Bj÷rn!

03 Feb 14 18:59, you wrote to me:

Usually known as the ASCII characters higher than 0x7f.

There are no ASCII characters higher as 0x7f. The 8th bit depended on the
parity sceme used.

Because there is no consensus on what is to be represented by these
so called high-ASCII.

No need to. Every region (outside of Z1 and Z3) -- meaning parts of the world using completely different languages -- takes care of it's own. The rest can simply either ignore it or take the recent technical developments and handle it accordingly.

Apparently they do.

many years now, and I'm still just Bj?rn in the nodelist there... :(

It is not the magic on how it is to be done.

No magic needed. My 15yo Y2K patch can handle it.

Your patch can only be used on an obsolete OS.

Maybe you haven't have time to read my recent comments yet? What I'm asking is, why this option is even needed. And if it is, why is the default to have it disabled and not vice versa?

I have read a lot of comments and I have engaged in discussions on this
subject, maybe not in echo's that you read, but still.

you have written there is no single way to write Bj?rn.

But there is! *I* write it the way I want it. In my Region20 file it's written the way I want it. That's simply the way *I* want it. Freedom of expression, anyone...?

Well so what is your problem? You have it your way where you decide,
others have it their way where they decide. I can assure you that the
nodelist looks very different, where I decide.

Have you ever looked at the distribution list compiled by Z3, it has been
totally different for many years.

But I am still curious, what definition of high-ASCII will be used?

None at all. If we regard the nodelist as the simple, binary file that it was intended to be from the very beginning, we can put whatever characters that we want in there.

It has never been a binary file and during the time I have been in Fidonet,
it was never intended to be a binary file. It is a structured file and that
brings limitations. You must always consider that there is 15 old software
that has to parse the file. Using 2 byte characters could confuse the parser.

Kees

--- FPD v2.9.040207 GoldED+/LNX 1.1.5
* Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)

Hello Björn,

On Monday February 03 2014 18:59, you wrote to Kees van Eeten:

Just out of curiosity, what is high-ASCII ?

Usually known as the ASCII characters higher than 0x7f.

There is no such things as "ASCII characters higher than 0x7f". ASCII covers the range 0-0x7f. If it is >0x7f it is not ASCII.

Because there is no consensus on what is to be represented by
these so called high-ASCII.

No need to. Every region (outside of Z1 and Z3) -- meaning parts of
the world using completely different languages -- takes care of it's
own. The rest can simply either ignore it or take the recent technical developments and handle it accordingly.

Everyone uses his own encoding? That's crazy.

But there is! *I* write it the way I want it. In my Region20 file
it's written the way I want it. That's simply the way *I* want it.

So it will show on your screen as intended by you and as garbage on anyone else's. What good would THAT do? YOU know how to properly spell your name. So that it shows OK on YOUR screen has no added value for anyone. The added value of allowing other characters than ASCII would be to inform OTHERS how to properly spell your name. And vice versa. That doesn't work if everyone uses his own encoding.

Freedom of expression, anyone...?

What you propose is not freedom of expression but the freedom to utter gibberish. I suppose you have that freedom, but it does not make sense when the
goal is to communicate. For that a common ground is needed.


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20110320
* Origin: http://www.vlist.eu (2:280/5555)