
Saturday 06-Feb-93 07:00:39

--650--
This is with CNet v2.60/R1 running here at FW:

FW BBS> Main> L
FW BBS>
FW BBS> Timeout error loading external program.
FW BBS> Sorry, program is not 3.0 format.
FW BBS>
FW BBS> Base, Feedback, Gfiles, Off, Pfiles, STatus, Time, Uploads, ?=Menu
FW BBS>
FW BBS> Main>

--651--
These are facts, not opinions:
1)  There's not a single Trapdoor purchase-sight anywhere in this country.
2)  You should allow upto *8* weeks for deliverly.
(Both of the above, are per TD's author, as of v1.83, the latest release.)

These are opinions, not facts:
Both statements are terrible ways to run a railroad.
I'd rather not deal with international currency exchange rates.
I'd rather not be at the mercy of international mail.
(Lost checks, lost software, and delays.)
If you purchased CNet TODAY, and started running it full-time immediately,
who wants to wait 4-8 more weeks to run Fidonet?

You can argue or disagree with my opinions.
But you can't change the above facts.

LH> Finally some reasons for why you want Ken to program a front end so
You haven't seen ANY of my other many reasons over the last 6 months?
Or other folks' either?

LH> Finally some reasons for why you want Ken to program a front end so
LH> badly... This could be solved by arranging to have FW as another North
LH> American registration site, something that has already been suggested.
I don't feel that having Ken help TD's author get rich by selling
TD software for him, is a 'solution' to CNet not having its own
FD.

LH> BA> BLH> BA> Stuck using Paragon-BBS JUST for its Fidonet FrontDoor.
LH> I do wish you would lose this tagline, I think you have outlived it's
LH> usefullness.
I feel it is still helping cut down on A LOT of the email I used to receive
regarding my very unfortunate "choice" of BBS software.  Feel free to
have any tagline you choose.

Len, thanks for a more reasonable, fair, calm, intelligent pro/con
discussion than others have given this topic.
It's the only way to go it.

--652--
P> And as Don M mentioned, this has been gone over at least 25 or 30 times.
P> about every 2-3 weeks someone asks about this warn code. I suggest
P> instead of calling up and bitching, that you start paying attention
This definitely needs to be in an online bulletin, or something.
Address it once, instead of 30x.

(And anytime a BBS tells callers to specifically notify the SysOp about
something, the SysOp should be informed about what action to take, or at
least what it means.  Not have to reply to users with, "I don't know
why you got that msg.  I don't know what it means.")

Here's a quote from Ken:

#2/2 from: Big Brother  (Ken Pletzer)
     orgz: Perspective Software
     on  : Mon 25-Jan-1993  8:41a

> Anybody get this before?  what's it mean??  (warn code 11 ; please notify
> sysop)

KP> First of all, these messages were NOT supposed to occur very frequently ...
KP> more of a way for me to tell when something REALLY BAD is happening.
KP>
KP> I have NEVER seen it happen in real life here on Future World (a micro-scale
KP> BBS, admittedly, by many measures).
KP>
KP> There are several variables in the PortData and MainPort structures called
KP> warn1, warn2, warn3, etc.  In a "warn code" the 1st digit is the structure
KP> (1==MainPort, 2==PortData), the 2nd is the location (the "warn" variable).
KP>
KP> These warn variables are normally 0.  If SOMETHING overwrites them, they
KP> become NON 0.  That something could be a bbstext line WAY too long, and
KP> overwriting one during a sprintf() or strcpy() command; could be a pfile not
KP> knowing WHAT it is doing and just randomly blowing away innocent memory; etc.
KP>
KP> In short (int, char, float, too), they were installed to notify the sysop
KP> (originally just me during "alpha" testing) that he did something very wrong,
KP> and the data NOW in memory may be inaccurate (such as the current user's
KP> CHARGES, etc).
KP>
KP> Ken
KP>

--653--
P> Maybe its me.. But some descriptions of how to call these functions
P> would be extremely helpful. Has anyone used all of these functions?
P> Could you possibly share this info??
Full docs are badly needed.

But, if Ken hasn't seen a need/time/desire after several years, I don't
think he'll write any over the next few years either.

I sure hope I'm dead wrong about this.

--654--
CNet> **** YANK REPORT
CNet>  Yank already running--you will be notified when complete.
CNet>  Press RETURN to continue...

Is there a way to have CNet generate the above text immediate after you
answer YES to "Mark messages using a background process [No]?"?

Or is there a reason to have a user type in all his KEYWORD-TEXT and
FILTER-TEXT and answer all the prompts and THEN tell him afterwards
that the YANK process can't even be started anyway?  (And he has to do
it all again in a few mins.)

I always feel like I'm missing something regarding things like this.

--655--
It would be a pain in the face doing all that 'per-sub' 'user-flag'
'carbon-copy' setting.  But using JOIN/DROP's "ALL" or "RANGING"
powers, would make it a breeze.

I'd like to be able to mark all my areas via DROP and JOIN.
Specify this as my 'new-scan' area list.

Mark all my areas via DROP and JOIN.
Specify this as my 'carbon-copy' area list.

You could even go 2 levels farther...

Mark all my areas via DROP and JOIN.
Specify this as my 'global-search' area list.

Mark all my areas via DROP and JOIN.
Specify this as my 'bundle and D/L' area list.

CNet would have 4 copies of its _buser file.
(You wouldn't really have to do all that 'DROP/JOIN' marking, you
could just specify "copy or swap" _buser files around.)

CNet would need 1 cmd that would produce a menu like:
CNET> Have your JOIN/DROP cmds effect:
CNET> 1) all lists
CNET> 2) 'new-scan' area list
CNET> 3) 'carbon-copy' area list
CNET> 4) 'global-search' area list
CNET> 5) 'bundle and D/L' area list
CNET> 6) Copy 1 list to another

By default (or users that didn't want to mess with any of this)
you'd be in the #1 mode.  (Currently CNet's only mode.)

Other callers would set their mode to 2, JOIN/DROP as needed.
Set their mode to 3, JOIN/DROP.
Set their mode to 4, JOIN/DROP.
Set their mode to 5, JOIN/DROP.

(See my suggestion #570 in KEN.LHA for more info on this.)

--656--
AK> Well, I on the other hand LOVE this new feature. I guess thats why each
I love this feature, too.
But not when grossly abused with tons of pointless graphic chars.

AK, what do your DOS filenotes look like?
Filename.LHA            23491 ----rwed 09-Sep-92 15:38:32
///----\\\\--- This File ///----\\\\--- Uploaded by ///----\\\\---

Is this useful?

--657--
I think Ken should NEVER mention a release date.
But ALWAYS mention a RELEASE RANGE.
Like stating "first half of Feb" "Last half of March", NEVER 01Feb93.
And to not mention ANYTHING, doesn't avoid the problem.

I think Ken should place his ReadMe under his gfile menu.
Or better yet, try out the new v2.60 "auto-grab type item".
Programmers could plan out their (current and future) ideas, keeping
the ReadMe text in mind.

--658--
DM> Bill and all.. Please remember that some of the Config stuff that you
DM> change around will not take effect untill  you reboot  your system.   Some
DM> of the stuff like the paths is not changeable untill a reboot and the info
I change a PATH.
I exit the PATH screen of the CONFIG.
I re-enter the PATH screen of the CONFIG.
My path #4 has now switched with path #5.

This is a 'must reboot 1st' problem?
I didn't even get time to exit CONFIG.  I didn't even get  a chance to
save anything.   When should I 'must reboot 1st'?

This isn't a bug???
I've not seen anything mentioned by Ken, about this being fixed, or
even being acknowledged as a bug.

--659--
Now that ^Q and ^Y call-up MCI input, the visual-editor needs to have
its HELP file updated.

It's help-file is showing:
CNet> Q   Page Up     Move up one page-full of text
CNet> Y   Bottom      Move to the bottom of text

Both are wrong.

Also, visual-editor's help-screen...
I don't know how many users (that REALLY need help) are going to be
smart enough to find it hidden under "CTRL eXit/file".  (In fact, the
1st time I used the editor, I was surprised to see so many additional
cmds hidden under "CTRL eXit/file".  Maybe "CTRL eXit+more" would be
better.)  And maybe a separate ^H to get help.

--660--
I like the briefer muffle/hide markers under WHO.  But is everyone
going to know what they mean?  Maybe at the end of WHO's output it
should display a ledend? (Or maybe if users type "WHO ?").
WHO> M=Muffled, H=hiding, etc.
(Who has a list of all of them?  Please post.)

Would anyone have use for a new feature?
A "?" next to a user's name in WHO, would mark him as one willing and
available to help out with questions.  When I'm calling local and have
time, (both rarely) I'd set my EP-help flag to ON.  Otherwise I'd set
it OFF.  This would set my WHO "?" marker accordingly.

Then WHO could tell everyone which users:
1) Specifically don't want (and can't) to be bothered/chatted. (M H)
2) Specifically welcome questions/chats. (?)
3) Feel neutral to #1 or 2. (no flags)

--661--
Ken, please consider passing "S" through the same 'range support' that
is allowed elsewhere on CNet.

>Mail Read
>Quit, Scan, New, Old, [All]  s41-45

Scanning from #1 (or the current #) instead of the one you want, can
be pretty annoying when you have 50+ msgs.

Also,
If a user is getting nasty email that they would prefer not to read,
is there a way to just scan-it, see in the scan-list that it's from
'HER' again, and delete it w/o reading it?

It would be a "[K]ill #-#" cmd, but from the scan prompt:
>Quit, Scan, New, Old, Kill, [All] >

--662--
Regarding 2400bps VE usage...

I tried to avoid having the visual-editor (VE) redraw the full 24
lines of my text each time, by setting my screen height to 10 lines.
Figuring I could still full-screen edit pretty well in a 'small
window' of 10, instead of 24.

No luck.

--663--

Regarding Bill's auto-re-dialer TERM arexx script...

Ken> May I add this command right from the term pull down?

Bill> A way to run an arexx script.  (Maybe a new CNet pull-down menu?)
Bill> while in the TERM-MODE.

Sure.
Maybe even do something like:
Pulldown Menu: RUN AREXX...
   SubMenu 1  (runs CNet:Rexx/1.rexx)
   SubMenu 2  (runs CNet:Rexx/2.rexx)
   SubMenu 3  (runs CNet:Rexx/3.rexx)

Pulldown Menu: RUN DOS...
   SubMenu 1  (runs CNet:c/1.exe)
   SubMenu 2  (runs CNet:c/2.exe)
   SubMenu 3  (runs CNet:c/3.exe)

Pulldown Menu: RUN PFILE...
   SubMenu 1  (runs Pfiles:1)
   SubMenu 2  (runs Pfiles:2)
   SubMenu 3  (runs Pfiles:3)
(Or whatever paths/filenames you think they should be called.)
(Or place both the paths/filenames and the descriptions that will
appear in the pulldown menus, into BBSTEXT.)

Any sysops could run any of the 9 at any time, with a pulldown menu.
(It would be handy from the term as well as the BBS itself.)

Give some power for local log-ins.
I'm sitting right here at the machine.

--664--
Has anyone heard of a new XPR protocol known as "HYDRA"?
Any info on what it has to offer?

--665--
CNet> Aborted or possible errors.
CNet>
CNet> Browse, Scan, Read, Download, Upload, Quit, ?=Menu
CNet>
CNet> (1) Suggestions for the future> WHY
CNet>
CNet> No error message has been set.

Shouldn't WHY give me "Aborted or possible errors."?

--666--

PT> Yeah... I PREFER the SINGLE MCI on/off key.... KEN!?!
Me too.
And I will always think of ^Q as xon/xoff usage.

And ^Y as:
CNet> Y   Bottom      Move to the bottom of text
as CNet v2.60/R1's help text is STILL showing.

JS>   There will be some resistance to the new MCI system, but I think once
JS> you start looking at how VERY much more powerful it is, you might feel
JS> better about the change!!
*VERY* much more powerful.

Are 2 MCI's startup keys needed?  (^Y ^Q)
I had originally thought they could just be just interchangeably.

But apparently:
TW> Well, I figured it out... Control-Y is for for color MCI and Control-Q is
TW> for running files...
we need to remember which is is which for different types of MCI codes?

(Also, there aren't really that many 'extra' unused ^keys left, to
'spend' 2, unless this was absolutely necessary for some reason.)

--667--
All my future posts now pertain to v2.60/R2....

I notice the "last-DL-date" is set to the "post-date" even if there's
been 0 DLs.  I think a file with 0 DLs should show a last-DL-date of
"00-Jan-00 00:00", like CNet does everywhere else.

--668--
Msg subs...

MY "DATA/_TEXT" FILES ARE ALSO CONTAINING THE PASSWORD OF THE POSTER!
DANGEROUS!
(It appears to be just the 1st 1 or 2 posts in each new sub.)
I didn't see this deliberately being a part of the _Text structs.
(Nor a need for it to be added.)
Maybe just junk-memory-data from a new, non-cleared _Text?

Also, can I safely delete all my old "post" and "response" dirs?
UP250 didn't remove them.
(Or maybe I answered the prompts wrong when I ran it?)

--669--
New MCI...

Who can give me a short class in Reverse Polish Notation (RPN) for the
new MCI?  How would I do something like (a*a) + (b*b)?

(It is very unfortunate that {M} couldn't have just used straight,
normal, readable, math.
{M60 (24/2)+45}
I don't know how often RPN is used since the invention of calculators
and parenthesis.  (Although HP's calculators have suck with it.)
So apparently someone is using it far more than I have since the 1970s.

I also would have switched around the "#" literal specifier.  "#45"
would mean "GetUser 45".  So users would know there is something
'significant' about it.  It's not the literal number 45.  But clearly
noted with the "#".

"10" would mean "10", as it does in real life.

1) {M60 24 #3 / +}            (current RPN)
2) {M60 #24 3 / +}
I'd just find #2 clearer than #1.
and
3) {M60 (24  / #3) +  60}     (non-RPN)
4) {M60 (#24 /  3) + #60}
#4 clearer than #3, still.

The new MCI is VERY powerful.

--670--
1) c:Version shows the BBS version # as "CNet 2.60/R2".
2) At log-in it says I'm running        "CNet 2.60/R1".

Please consider an MCI code that would allow #2 to display this
correctly 100% of the time. (We could display the correct CNet version
# during the next 100 releases of CNet.)

If this is impossible to add, please consider having CNet display the
exact "$VER: string" for #2, that #1 uses.  (Not have 2 different copies
that can easily be forgotten during a last minute update/release.  If that's
what, in fact, has happened here.)

--671--
Any plans on adding a LEARN key to CNet?
1) ^L turns it on.
2) CNet remembers your next 10-40 keystrokes.  (Text and ctrl chars.)
3) ^L turns it off.
4) ^M repeats those saved keystrokes.

Instant macros at any prompt for repeating cmds.
Powerful, full word-processing stuff in both editors.
Instant macros for words/phrases in chat/OLMs/group-confs/mail/etc.

--672--
DM> Live wire at the RESPOND or pass prompt, try entering S and see what is
DM> shown..Think it will give you what you want there..
Won't this just show the responses to the current single item?

He's trying to SCAN item headers, NOT response headers.

There would be only 1 way to make EVERYONE happy with this.
And that would be to do it the way giant, multi-line UNIX BBSes do it.
>>>The SysOp can define his own header-display with MCI codes.

Some systems even allow EACH USER to define their own header-displays.

Well worth the 5-10 new MCI codes needed.
(I don't think Ken has any intention of supporting this excellent,
powerful, but simple-to-do feature.)

JN> BA> Well worth the 5-10 new MCI codes needed.
JN> BA> (I don't think Ken has any intention of supporting this excellent,
JN> BA> powerful, but simple-to-do feature.)

JN>         Yes, I think this would be a GREAT enhancement to C-Net power!
JN> I'd like to see SysOps be able to do this not only in the SCAN headers,
JN> but headers in every other place!  For instance message headers, the
JN> userlist etc.

Well, if they were MCI codes (NOT c-formatting %#.#d codes), I
can't imagine them NOT being available at several different places.
CNet would need:
1)  A few new MCI codes.
2)  Lines in BBSTEXT in stick them in.

#1 is already available via MCI XT##### codes.
#2 is already available nearly everwhere w/CNet.
But they don't work together.

--673--
I just EDited the text in response #170.
After adding a few lines and resaving, I entered "170", which
displayed the txt from response #171, instead.  "169" displayed #170
txt.  "L" displayed msg #171, but with the text from msg #170.
Afterwards things straightened themselves out and reading 169-171 gave
the correct msgs.  (I hope I got that bug reported properly. I don't
remember if things were +1 or -1, but they definitely weren't correct.)

--674--
Ken, a short comment-line next to some of the structure definitions
in cnet.h would give as an incredible amount of info and help in
tracking these things down.

> struct MyStruct {  /* Mail:users/#/_my_struct used for <THING> storage */

--675--
On Fri 12-Feb-1993  9:50a, Goodman wrote:
G> I need help.  Whenever I use the {v7} (time remaining) command, I always
G> get an extra 0 at the end.  This makes things look a lot out of
G> proportion.  Anybody know how to just show the minutes left?  It used to
G> be 999.0 now it's 9990.  Please help!

I wouldn't call it "out of proportion", I'd call it "10x wrong".
(The docs state V7 is mins, not 1/10ths.  Nothing I could find in
any docs (or even the very latest ReadMe) that say it was changed.)

To divide it by 10, and show time to the nearest min...
{l60 #0}           (Clear variable)
{m60 7 #10 / +}    (/ by 10)
{v60}              (Show it)

I'd prefer {v7} instead of all that.

Or if you want 1/10 accuracy (999.0)...
{l60 #0}
{m60 7 #10 / +}

{l61 #0}
{m61 7 #10 % +}
{v60}.{v61}

--676--
Has anyone gotten the {%0date} MCI code to work?
{%01-1-90}
{%01-1-95}
Neither drops carrier or reduces the user's time.

(Reply:  A space is needed between %0 and 1-1-90)
(It's not mentioned in the new MCI docs.  In fact, the entire {%} {&} {+}
and other cmds, aren't mentioned AT ALL.)

Also, entering things like:
-----
This line disappears too.
Testing: {L7 #0}
Testing: {V7}
This line disappears too.
-----
into the editor, makes .L .M (or ^P in Vis-Ed)
show my msg is NULL.  Not even the text before or after the MCI codes is
ever shown.

Does .M .L ^P cease to work if a user's time is 0?

But he stays online anyway (because he's in the editor).
This *IS* a desirable feature, so he can finish entering his msg.
But if he can't see it?

--677--
Ken, currently if I want to do some math:
> {l60 #0}                   (Clear var)
> {m60 7 #10 / +}            (Divide by 10 and add it to var, store in var)
> {v60}                      (Display it)

Would there be a way to cut out 2 steps:
1) NOT have to clear the var 1st.
2) NOT have to 'add to it'?
   (Just default to "place it in the var", if no "add-to" is specified.)

> {m60 7 #10 /}             (Divide by 10 and load result into var)
> {v60}                     (Display it)

Just to stream-line things.

Or am I missing something?  I'm having an oddly difficult time
understanding the simplest of RPN concepts.
Anyone have any RPN tips?

--678--
On Sat 13-Feb-1993  4:05a, The Brazilian wrote:
TB> > Why couldn't we just have "(search for the \ and replace with the
TB> Ctrl-Q)"?

TB> Because the new MCI commands require closing (trailing) braces; you'd have
TB> to do more than simply search for a \ and change it; it would have to
TB> become
TB> smart to parse out MCI, find the ends, and complete it.
The new MCI docs show at ALL MCI codes need the "}".
But in actual practice, some do, some don't.

The original real question here had nothing whatsoever to do with "}".

But, here's the complete picture.
1)  Search for leading \ and replace with ^Y?
2)  Search for leading \ and replace with ^Q?
3)  Search for the trailing \ (if doesn't exist, add ^Y)?
4)  Search for the trailing \ (if doesn't exist, add ^Q)?
5)  Search for the trailing \ (if doesn't exist, don't add anything)?
6)  Search for the trailing \ (if exists, change to ^Y)?
7)  Search for the trailing \ (if exists, change to ^Q)?
8)  Search for the trailing \ (if exists, change to "}")?
9)  Search for the trailing \ (if doesn't exist, add "}")?
Which is REQUIRED?  How can anyone even attemp to write a converter program?

It is a shame these 9 questions even have to be asked.  The whole
change should have been just been a simple, single:
1) Replace all "\" with "^whatever".  Period.

Ken/Jim please end all this 'guessing' on our parts.  I'm hearing more
and more rumors about this.  (Please, no more "I think it works this
way", by all of us.  Including me.)  What's the 100% full scoop
between ^Y and ^Q usage?  Why 2?  Used where?  Answer the above 9
questions, please.  Which MCI codes MUST have the trailing coding?
Which don't?  Which seem to work either way?

I'm 100% favor of the new MCI changes and the power it brings to CNet.
But the changes need to be fully explained.

--679--
GD> I type 'EL' and the drive scans for a second and then says "ERROR READING
GD> VDE FILE"

Ken, please consider changing this to read:
> ERROR READING VDE FILE: Cnet:SysText/VDE/Subboard (or whatever).
You'd NEVER have to address questions like this again the rest of your life.

Do I think it is *THAT* important?  Worth the effort?

YES!  Seeing as, a user's VDE will not work AT ALL, if that 1 file is missing,
yes, it is worth the small amount of added code, added work, added time.

--680--
TR> It would have been nice to have gotten just an .LHA archive with all the
TR> files that have changed since the last major release (v2.42e) here on this
TR> sub-board, but unfortunately we have to trudge through a whole DMSed disk
TR> to see what has changed and make our own modifications from there.

This was proposed, and 1 CNet release was done like this.
It was great.
1)  No digging out floppies to unDMS onto.
2)  Much smaller archives to DL.
3)  Saves time, LD phone costs, HD space.
4)  You know everything in the LHA is only stuff that has changed
    since the last release.
5)  Ken could prepair the whole thing by executing 1 DOS script.
6)  No boot-block DMS "worries".
7)  You could backup your old data, and UN-LHA the new update, automatically
    placing everything right into their correct directories with a single cmd.
8)  No creating and formatting exact floppy-sized ram disks.
9)  View inside the archive online and offline.

After that 1 CNet LHA release, it hasn't been done since.
Can't imagine why.
Callers could DL the LHA 'update' or the DMS 'full release'.
Whichever they needed.  But I bet 95% only needed the 'update'.
I'm not sure why those 5% that needed both, full DMS disks,
prevent the rest of us from having all these 1-4 advantages.

Ken, please consider LHA again.

Could LHA be considered just for 1,4,6 purposes on minor updates?
And 2,3,6 on major updates.

The phone company is rich enough.

I'm not asking for an LHA release to INCREASE time/work.

I'll agree, for major updates like v2.60, #2-#3 will be minimal improvements
(but still slightly better.)

#1,#4-#9 will ALWAYS be better with LHA instead of DMS.

Could you list 9 (or more) reasons for DMS over LHA?

--681--
Shawn!!!!  Where are you???!!!??
Please.  Another 'getuser' list.

(Are you generating that 'by-hand'?  An arexx script?  If so, are you
going to release it?  Then we could make are own 'getuser' list.  Or
Ken could even execute that arexx script right before he DMSes the new
master disks, and include the 'getuser' list right in the archive.)

"Instant new getuser list" with every release!!!

It could even go 1 step farther and automatically generate a
"translation/conversion list", too.

>  v2.60    Item            v2.42e
> ------------------------------------
> 1234567   User's name     1229876
> 1255555   User's address  1212333
> 1244444   User's city     1238331

Even someone without any "arexx getuser" knowledge could edit scripts,
replacing old values with the new ones.

--683--
Ken, please consider adding an MCI code that would call a VDE data file
(CNet's or our own).  Mainly for BBSMENU use, but it could be used
elsewhere also.

> MyVDE | {[SysText:VDE/MyVDEdata}

I'd gladly do and release a few VDE's of my own.
Ones that can act as:
1) "Replacements" to CNet's VDEs.  (Simpler or more powerful.)
2) "Optional alternatives"
    (Keeping both CNet's and the new ones online.  Take your pick.
    Or different ones for different access levels.)
3) "New".  (Edit other CNet structures, currently not supported.)
4) "Custom". (Edit 3rd party CNet pfile data, etc.  It's a shame to
    make each author design his own data/cfg-editor when VDE is
    does it all.)

Currently, without this new MCI code (or another method), I think we
can only do #1.

Unleash The Power.

--684--
SM> BB> Instead, he is seeing about 5-10 users that HATE the idea of a CNet
SM> BB> and are saying so LOUDLY.
SM> Bill you have many great suggestions, NONE of which would be implemented
SM> until 1995 if Ken started work on a mailer today.
This is NOT the case, according to Jim S.  The closest person to CNet,
other than Ken himself.  And Ken's not talking.

Why would *ALL* new features have to stop, if Ken chose to work
the same x hours per day on CNet.  And 1 additional hr on the FD.
This would force *ALL* new stuff to stop dead?  Most?  Some?  None?
The correct answer (if Ken did it this 1 way, of many possible ways)
is *NO* CNet development would have to stop, at all.  0%

Or he could devote 10% of his time.  10% of the new features would stop.

Or 50%.  Or devote 100% of his time on a FD.

So, it would stop new features somewhere between 0% and 100%.
Quite a difference from your assumed 100% stoppage.  Ken could choose.

Other BBS authors (no where near the talent of Ken) have added a simple,
FrontDoor and Fidonet itself in 6 months.  I don't know why you assume it
would take Ken 5 times longer than them.

SM> You do understand that with Ken writing a mailer, he MUST write
SM> something that will completely replace TrapDoor for all of us.
Why?  You run TD, the other 90% of CNet's will run Ken's FD.
To each his own.  I honestly think it WOULD be a 10% VS 90% ratio.
Just my opinion.

SM> Im not going to "pay" for Ken to develop a lame mailer that
SM> I wont use. ANd with Ken writing a mailer, a nodelist compiler
SM> is also required, there's another few months development time.
SM> When CNet internal fido support, (netmail, etc is complete), when cnet
SM> has a tosser, and when CNet hsa internal UUCP support, that
SM> would be time to start a mailer.
It's a Catch 22 situation.  I honested don't think Ken is EVER
going to support/add/rewrite/modify things so that OTHER FD's
will work better.  He just doesn't sound that interested in Fidonet at all.
I don't think he'll even EVER consider RUNNING a Fidonet BBS.
How much of the MASSIVE # of v2.60/R2 improvements were geared towards
better Fidonet support?  After months of hard work.

Now, if he had his OWN FD, I think that would be the incentive he needs.

Maybe, just maybe, we are all going to have to go through the back-door
on this one..  Ken writes a FD, THEN supports/adds all the features that
the BBS needs, because of it.  I know you (like a VERY small % of others)
feel it should be the other way around.  Maybe it should.  Maybe not.

Or at least it's worth asking yourself:
If you were a BBS author, would you more likely support someone else's
FD or your own?

--685--
The new variables are greatly appreciated!

2.60> 179.  Several new item variables are available from the item AT
2.60>       command in a subboard:
2.60>  Notify uploader at DL:  The uploader will receive a "blank" message
2.60>       from the downloader, with the simple title of
2.60>       DOWNLOADED: <filename>

Shouldn't the ULer himself (not the SysOp) have the power to set this
one?  To avoid making all your callers mail the SysOp, just to have
him (the SysOp) MANUALLY start or stop this file's 'auto-mail' process.

I'd love to be able to config which variables are "SysOp settable" VS
"SysOp or maintenance" settable.  With 50 of them available, we'd never
all agree on which should be which.

Make "ED#" allow #1 cmds, if the user is UPLOADER.
Make "ED#" allow #2 cmds, if the user has REMOTE MAINTENANCE POWERS.
Make "ED#" allow #3 cmds, if the user is SYSOP.

Could this be done similar to the way CNet v2.60 now allows MCI groups
1/2/3 to be SysOp defined?

Also, I'd really like 'maintenance powers' to have the 'grouping'
method also.  Just too much power to give 'all or nothing' when you
just want to give someone "TEst" only powers.

The SysOp would define which maintenance powers fall into groups 1/2/
3.  Then each user would get assigned a group 0-3, instead of a
'all or nothing' YES/NO maintenance-flag.

--686--
Be thankful.  Here's how Starnet SysOps change their screen colors:
(Direct quote from Starnet's author.)
>There was questions about changing the colors of the screen or text ect...
>Well those of you into modifying the program with newzap will like this...
>Do a text 'Search for "Spock"' after the 'K' there is a 00 byte this ends
>the string Spock, the following 24 bytes contain the colors in 8 sets of 3
>as what would be in a setrgb4(); function call. The first set and the last
>are the ones that affect black and white.

Handy, huh?  Binary edit the giant executables, searching for Star-
Trek character names.  How professional, huh?

And I'm *STILL* running it.  I need Fidonet.  And Paragon/Starnet comes
with a built-in, free, no 6 week wait, instant, no international-mail,
simple, microscopic memory usage, easy-to-use, easy-to-setup, no
reload 100 times per day, no exit 100 times per day, not 3rd party
shareware, frontdoor.  (And yes, it's even bug-filled.)  But it's
still better than CNet's no-front-door-at-all setup.

*THAT'S* how badly I need Fidonet as an out-of-the-box part
of a BBS.

--687--

According to a 1991 published estimate there were over:
50,000 Usenet systems,
With a total of 30,000,000 readers.

We are talking GIANT network.  I'd love to see CNet support
it directly and fully some day.

--688--
Nice feature, but I don't think we'll see it in too wide-spread usage
until your suggestion is implimented.  Just too hard to set this flag
file-by-file after each UL.  But a flag on per-sub basis would do them
all.

>Notify uploader at DL: The uploader will receive a "blank" message
>           from the downloader, with the simple title of
>           DOWNLOADED: <filename>

Maybe also some mention that this mail was 'automatically' sent by the
system, so you don't think the DLer wrote you a blank msg.  (You know
callers will start replying to these.  "What did you want?"  "Why did
you mail?"  "I didn't mail you."  "No, you mailed me 1st.")

I also think something like the following would be better than a
"blank message" anyway:
CNet> Your upload (FILENAME.LHA) was DLed 15-Feb-93 03:12:32.
CNet> You were awarded 3 file credits and 30,922 file bytes due to this.
CNet> Bringing your credits upto:  17 files @ 341,039 bytes.
CNet> ---This msg auto-generated by CNet-BBS.

(Or better yet, make it configuratable via BBSTEXT.  Or is it already?
Which line #(s)?)

ULer's getting a handful of these msgs, and watching their credits
grow, would be greatly encouraged to UL more and more, better and
better quality titles.  (I'm looking forward to getting some of these
auto-credits myself at systems I call.  But we'll never know exactly
what files/bytes bonuses we got.  Maybe big.  Maybe small.)

--689--
Config idea....
CONFIG> Mail maintenance warnings to user ID#: ______

You could send them to user #1 or any user specified.
(ID #0 would turn them off.)

CNet mail you reports like:
> UDBASE6: is 98% full.
> Sub "Arexx Files" isn't allowing any more posts.
> 84 dropped carriers yesterday.
> 23 failed password attempts.
> 16 no-connects at log-in.
> Reported Warning #11, 6 times.
> UDBASE:MyDir/MySub is missing.

Why beg your users to tell you about this stuff?
Why wait until something fills?
Why have a problem for 3 months before you find out?

--690--
If I kill a file that's OFFLINE or NOT ONLINE, CNet still asks:
CNET> Delete file also    [Yes]?
this prompt most likely not even need be asked at all.

Also, for files that ARE ONLINE, it would be nice to know if CNet
successfully found and successfully deleted the file.
(Pfiles/gfiles/news files etc, also.)
CNet> Delete file also    [Yes]? YES
CNet> Deleted UDBASE3:MyDir/FILENAME.LHA successfully.

(I do a lot of file juggling.  There's no telling if the system
can find things when I get through with it.)

--691--
Right before a DL, CNet says:
CNet> Bytes: 21K  Time: 0:23  Credits: 13 files, 54K  Charge: $0.32

I think CNet should be smarter in its wording.

If the user is GAINING something:  Always use words like 'credits',
'bonuses', 'rewards'.

If the user is LOSING something:  Always use words like 'costs',
'debts', 'charges', 'dues', 'prices', 'expenses', 'mark-up'.

So users don't think, "Hey look, I'm getting credits each time I DL.".

> Costing: 6 files @ 64392 bytes.
> Actual:  2 files @ 12227 bytes.     Time: 1:83mins     Charges: $0.33

Or you could do 'The Works' and have SS and DS show these costs on a
per-file basis.  You might have 15 files selected.  Many are free.
Many costing you 1 DL.  Except #4, which is costing you *23* DLs and
big bytes.  "Yikes.  I really don't need that file that bad."
Currently, you'll never know which one it is.

>#   Title          Actual      Costs      Description
>-----------------------------------------------------------------
>1. TEST_A.LHA     59K  1:23    1@ 59K   Filenote goes here...
>2. TEST_B.LHA     13K  0:53    0@  0K   Filenote goes here...
>3. TEST_C.LHA    117K 10:53    0@  0K   Filenote goes here...
>4. TEST_D.LHA     73K  0:45   23@832K   Filenote goes here...
>-----------------------------------------------------------------
>5  Total         192K 13:34   24@891K   480% file; 464% byte mark-up

Ken, was there a reason why SelectType must be just a small subset of
ItemType?  Couldn't SelectType be the full ItemType structure + the
file's full path.  I don't think marked-files are too large a disk-
space waster.  AmigaDOS sees a 1,024 byte files as the same disk-usage
as a 100 byte.  I'd love to see 3rd party SS replacements if we
had all needed info in the SelectType structs.

--692--
BB> It will only print 80 characters ... that's what's stored on disk.  To
BB> display the rest would require more time to search index files, etc!
Ken, since FIND only searches for filenames, wouldn't the index
really just need FILENAME and a method to point to the full file info.
No filenote (80 or full), size, or anything else in the pointer index.

But is WOULD allow full filenote displays, true UL dates, etc.
And might be even faster, to boot.

BB> AK> 2) It also does NOT display the BIN that the file is located it.
BB> No?  I don't know why that would be.
I didn't see it at first.  The sub name is right over the filenote.
(Not the best place to see it.  And not really too helpful when
trying to find exactly were is sub "Graphics".)

--693--
Is there an MCI code (absolute or XT#####) for displaying the
total calls to the system?  (And I mean grand-total since day #1.)

The MainPort has variables like Nums[5];      /*  # of accounts, etc */

What are these 4 other "etc" variables?
Ken, is this documented anywhere at all on planet Earth?

--694--
Ken, would it be possible to have SAM also display the "# of calls"
per last/period/setup/total?  If we see any of the *70* other
variables as being unusually high or unusually low, we could also see
the # of calls that might have caused this.

As well as just being 4 additional, useful variables that AM could
display.

(I stared at the screen long and hard.  I can't believe "# of calls
per last/period/setup/total" was left out entirely, but tons of other
info was included.)

--695--
And we can't really allow callers to use ANY math cmds themselves because
it also means they'll be able to change their own ratios/time/access
limits too.  So I guess it NO math at all.  (Therefore no accurate
display of these variables either.)

Ken, would there be a way to allow callers access to MCI math only
with vars 60-70 and 70-74?
So they can do "limited", "safe" math.

Likewise, would there be a way to allow callers to access {V} for 0-9999
but NOT the XT##### codes?
So they can display various things about THEIR account, but NOT
see names, ratios, addresses, phone #s and even passwords of other users.
(A XT##### warning should be in the manual about this.)

Allow {J}umping forward, but no backwards, to avoid endless loops.

Maybe distinguish between UPPERCASE {V} {M} {J} (full usage)
and
lowercase {v} {m} {j} (safe/limited usage).
(Eventually you're going to run out of A-Z MCI characters and going to
have to start using a-z, anyway.)

And I think I'm the only person in the world that knows we can now
specify what MCI codes pertain to MCI levels 0/1/2.
(This VERY POWERFUL, VERY NEW feature is not in the ReadMe anywhere.)
Ken, which lines of BBSTEXT pertain to this?

(Answered: lines 4-5.)

--696--
Has anyone tested backwards MCI jumps?
The following code doesn't work if the user picks 1 and then 1 again.

{9}                  (Label #9 marks top)

Choice: {G0}         (Get choice)

{T70 #1} {JE 1}      (Test if '1', jump forwad to label #1)
{T70 #2} {JE 2}      (Test if '2', jump forwad to label #2)
{T70 #3} {JE 3}      (Test if '3', jump forwad to label #3)
{JA 7}               (All others,  jump forwad to label #7 for exit)

{1} Picked 1 {JA 9}  (Label #1, and jump back to top)
{2} Picked 2 {JA 9}  (Label #2, and jump back to top)
{3} Picked 3 {JA 9}  (Label #3, and jump back to top)

{7}                  (Label 7 marks exit point)

Does the above all look correct?
(I've got far too many MCI codes and various computer languages going around
in my head at the same time.)

For clarification, definitions:
1) "Read from disk":  Saved as a text file, run as a text pfile from
                      the pfile menu.
2) "Read from mem":   The .M or ^P cmds in the online editor.
Do I have the above assumptions correct?
(Or can you give some better/other examples?)

I did ALL my "back-loop testing" with #1, not #2.
Still didn't work.  I'll keep trying.

You see what I'm trying to do?
Give a user a menu list, and have him pick, until he chooses to exit.

Or maybe you can think of a different way to do this, in a MCI
text file, but without the need to back-loop at all.

--697--
Has anyone been able to open any kind of window on top of CNet's
"public" screen, ever?

CLI> NEWSHELL CON:0/0/640/40/WindowName/CLOSE/SCREENCNETSCREEN0

I've been trying through MANY releases, including v2.60/R2.

--698--
J> BB> So, what you are looking for, is a CLI command that just start with
J> BB> "login:" and goes from there?
J> Yes Yes, it is just like this :-)
Hey, maybe even non-Internet sysops could find this handy.
Don't even open a CNet screen or window at all.
Just log-in and log-out right in your current CLI.

Ken, could us non-internet users pass this CLI cmd some parameters
that would, in turn, be interpreted by CNet right from the log-in prompt?

CLI> BBSCLI "SysOp;MyPass;MR;1;Q;O!"

Jamie, would this interfer with what you are trying to do?

--699--

RE> messages. (especially in the latest versions of CNet).  First off, you
RE> would have to write a conversion system that would import the uucp
RE> messages into a CNet style layout.  Most usenet groups receive hundreds of
RE> messages daily.  You would need LOTS of harddrive space to keep both
RE> copies on the system at once before you delete the UUCP messages.
Do UUCP msgs come in 1 by 1 or in bundles?
If bundles, wouldn't it be possible to go directly from the bundles into CNet
format WITHOUT the 2-step process?

RE> They come in packets.  In the packets, you have messages for any
RE> newsforum, or the packet may include internet mail.   I suppose IT IS
RE> possible to go directly from the packet to a CNet format. This would mean
RE> re-writing UUXQT.
But if 10,000 msgs need to create 10,000 files, just to have them deleted
after converting them to CNet format, it might be worth going directly
from bundles into Cnet format, instead.

We are talking *MAJOR* time/space/CPU-load/HD-fragmentation improvements,
here.

Likewise for UUCP msgs and Fidonet msgs.

--EOF-- Tuesday 16-Feb-93 17:21:05

-Bill "Mr. BBS" Beogelein, 313-473-2020, 2-line HST 14.4k USR DS, 1:2410/207
