Friday 23-Oct-92 02:14:25

--351--
Does anyone really need to give someone 47 3/10s of a min accuracy?
Wouldn't 47 or 48 mins be fine?

Users would gain or lose at most 30 secs if it was rounded up/down
to the nearest min.  And users and sysops alike wouldn't have to
remember which vars are in mins and which are in 1/10 min.

--352--
> If you enjoy CNet's SCAN-FOR-NEW at logon, you must now enter the
> command "NS" into your logon macro.  You do this my (SIC) using the
> EP command.

How do I get log-in macros like "WHO;NS;U;1;1;etc..." to work?

"NS" must come near the start because it can only be executed from
the main menu (ugh).

But "NS" seems to prevent anything from appearing afterwards in
the macro.  (It'll just be ignored.)

--353--
I often send the same "congratulation" email when a user is 1st promoted
to level n.  Or the same "thank-you" text when a user donates $$$ and
is raised to level n.  Or the same "sorry" info-text when the accounting
system demotes a user.  (Or I manually demote someone.)
Sure would be handy if:
> SysText:NewUserMail      (Sent to a user when he 1st registers)
> SysText:NewUserMail.n    (Sent to a user when he 1st becomes level n)

--354--
BB>  I am working now to re-do the subboard system once again.
Ken, would it be feasible it get suggestions from *100* users *BEFORE*
you have to go through this code-re-writing again and again?

You have an incredible brain-trust of information on-hand for the
asking.  Several 100 years worth of combined efforts in coding,
suggestions, algorithms at your finger tips.

It's called "Future World's userbase".  Don't be afraid to use it.

--355--
If Ken goes with a word-based getUser() call, I recommend he just
take all the element-names from the structs and use those.
The cnet.h file itself would then act as the docs for the word-based
getuser variables.

--356--
Instead of just allowing us to specify which types of arcs get
transformed or don't, maybe...

As always, the #s refer to their countparts in BBSarc:
>BBSARC transformable: 1-3,7,9 (Only convert these types.)
>Preferred ARC type  : LHA     (Change them to this type.)

If sub #12 is Amiga, you might want TO convert .ZIPs to .LHAs.

If sub #47 is IBM you might want to specifically NOT convert .ZIPs,
but convert .LHAs TO ZIPs.  Let's face it, different machines will
always have a different preferred packer.  And I might 'prefer' a
certain arc-type that isn't even supported at all by certion machines.

This represents, what I feel to be, a much more "real-world usage" on
per-base arrangement.

I know you can have a different arc-script in EVERY area, but that
might be 100s.  And I know we can specify our 'perferred ARC' method
in BBSTEXT (I can't imagine why this is in BBSTEXT and not BBSCONFIG),
but 1 'perferred ARC' isn't the same as 'perferred ARCs for each base'.

--357--
LH> Bill, if you change your BBSmenu file where it says things like:
LH>  YANK NEW PREVIEW to just YANK NEW then you won't have to read the
LH> previous response everytime. If the users want to do it that way, just
LH> alias P to PREVIEW and that should allow them to do things like YNP and
It was 1 of the 1st things I changed.  Works fine for my system, but
not the 100 other Cnets that I call that have PREVIEW as the default.

And I really hated to remove my PREVIEW because it does serve a very
useful purpose.  As a 'lead-in' to the current discussion.  Unfortunately,
PREVIEW isn't 'smart'.  It can't take into account things like:
1)  I've already read msg "n-1" 3 times.
2)  I *wrote* msg "n-1", I know what I said.
3)  Msg "n-1" is 100-200 lines long.  (ugh)

Actually, my 'read-each-msg-twice/thrice' was only part PREVIEW
trouble.  And part me doing RA, getting interrupted, and having to
do another RA.  (Maybe 2-3 RA's per call)

Remember, a BBS with *1000* Usenet subboards is still considered
a 'partial Usenet feed'.

--358--
Arexx New User Applications could do things like...

Check if the user entered a validate areacode.
Check if the user's state, matches his zipcode.
If a user can't enter his own name in 10 tries, drop carrier.
Does the user address have at least 2 words in it (# and street name).
Is his password too simple to guess because it was just found in
the online dictionary.
Does his age show he is <3 or >150 yrs old.
Check if his areacode is correct for his specific state.
Compare phone # with those just received via Caller ID.

As well as 100 other things that sysops now all do 'by hand' when
they are faced with the question, "Should I raise/lower or kill
this account".

And probably the most important thing to Ken...
Never again have to try and please 50 different sysop's methods.
Add/remove/change Cnet's code over and over again.
Ken could just say, hey, it's an arexx script, do it any way
you want.

--359--
Does anyone get annoyed by folks quoting 40 lines in replies, only to
post a 1-2 line response?

In local bases, wasting valuable disk space.

And in Fidonet/Usenet bases, when some of these mail packets get very
large, very quickly, even w/o large amounts of quoting.  (As well as
all the $$$ that sysops pay in phone charges.)

Would anyone like to see EL's quote flag be made tri-state:
 > Allow quoting: YES/NO
changed to:
 > Allow quoting: YES/NO/CONFIG

"NO" would allow NO quoting.
"YES" would allow unlimited length quoting.
"CONFIG" would use a global value specified in BBSCONFIG.
35% <- Percent of quote-text VS user-text allowed in replies.

If a user tried to save such a msg in that base, he might just see:
> Can't save msg, it contains 92% quoted-text.
> Limit for this base is 35%.
> Please trim the quote before attempting to re-save.

When you multiply 50 lines, in 50 replies, in 50 echos, fed to 200
systems, the *25,000,000* lines of (previously-posted!) quotes, start
to add-up.  (And I don't think my 50*50*50*200 is over-exaggerating
things.)

(A user's signature-lines would NOT be counted in the % totals.)

--360--
1> Kill all messages [No]? No
2> Kill [A]ll msgs or [O]ld previously read msgs?
I think #2 (dropping #1) would help encourage callers to clear out
their mail boxes.  Fast, easy, 1 key-stroke clear-out of all old/read
mail.  Instead of K4 [RETURN], K9 [RETURN], K12 [RETURN], K7 [RETURN],
Kn [RETURN], opps, I deleted the wrong one.

And, no, Cnet doesn't allow #-#,#-# ranging here either. :-(

--361--
The full power/speed/disk-space of every Cnet msg base and file base
is not something that should be left as "2nd best, next to Xenolink"
or ANY Amiga BBS, for that matter.  (Nor should Ken have to write it x
times over and over again.)  Cnet already has the POWER.  But is not
*THE* fastest, but is close, but is WAY behind on disk space usage.

Log into any Xenolink system.  Type: "D MyFile.LHA" and watch
how fast 100 subs and 10,000 files are searched.  (Even if the files
themselves are on the slowest CD-ROM made.)

Cnet doesn't even allow global file marking or D/Ling at all.
But, how long would "SG 'MyFile.LHA'" take under the same
circumstances?  (And then mark or DL 'by-hand' after matches are
found.)

Ken sounds like he is very interested in revamping things.
It's not going to be easy to code.  But once in place, it should serve
Cnet's needs for many yrs to come.

--362--
H> The thing is there is a flag to turn off imediate testing in a particular
H> sub, but I can't find anything to turn off testing in a particular sub
H> when maint runs. I have a Sub Beemer Uploads I only want the testing to
I see.  (Finally)
1)  Turn testing on/off after upload.
2)  Turn testing on/off for maint runs.
Should these both be controlled by the same flag?
Are separate (new) flags needed for each?

--363--
DM> I use the file size limit for a users mail box to encourage them to delete
DM> their old mail.. Set it so they can only keep four or five normal pieces
DM> of mail and they will soon get in the habit of clearing out their old
DM> stuff..

I always like to try and make things as easy as pie, 1st.  And see
if that encourages it.  (It usually does.)  Then and only then, start
limiting things, lecturing my users, making threats, dropping access,
killing accounts, etc.

I'd be glad to allow a very large # of pieces of email.  (User's that
don't call but every 2 weeks might need such.)  As long as they
have the attitude, "It's so easy to clear out my old, already read mail,
I might as well."

--364--
UN> BA> >base #40, 17 posts, 110 replies,  421 reads
UN> BA> >base #41,  7 posts,  12 replies, 9032 reads
UN> BA> >base #42, 33 posts, 274 replies,  109 reads
UN> BA> (I'm sure someone would write a pfile for this.)

UN> This is not a bad idea, as long as it would not take each sub any extra
UN> time to tabulate the # of reads. It already takes long enough when
UN> entering a sub.
I was thinking more like a separate pfile-cmd that a user or sysop
could run.  NOT something done everytime, every caller enters
every sub.

--365--
UN>  2. A command to vote step-by-step on all topics not yet voted on.
I'd even settle on x-x,x-x ranging to pick which topics I wish to
vote on.  (Not supported yet.)

UN>  3. Ability to "hide" the results of a topic until a certain date or
UN>     percentage of users have voted on it. Then as an option, automaint
UN>     could post the results as a news item.
I hadn't thought of that, but I guess sometimes seeing the results
of others CAN cause a band-wagon effect and make undecided voters
sway, resulting in inaccurate results.  (Imagine entering a voting booth
and seeing an up-to-the-min result tally for president BEFORE you
vote.)

If "until a certain date or percentage + amaint" is too tricky
to add, I'd settle for "only see results AFTER you've voted" for
right now.

--366--
UN> How about an extra command at the
 UN> : Abort, Quit here, Post, Drop, etc...
 UN> prompt to CANCEL THE NEW DATE for that particular sub? So if you do a
 UN> RA and finish reading a sub, you can have it's NEW DATE canceled.
 UN> If you quit and later do another RA it will only apply to the subs that
 UN> you haven't already read and canceled.
 I could go for that.  (Although my 1st choice would still be a user
 acct flag to make 'auto-reset date on completed RA subs'.
 Set-it once as desired and forget it.)

 Your [C]ancel idea would also be consistent with NSAL's [C]ancel.

--367--
I'd like to see Cnet's editor be able to receive chars with 0/0
line/char delays set on my end.  (For ASCII U/L or ASCII send modes.)

The manual states that sending chars too fast "my cause extra
LINE NOISE to occur".  (I don't know if this is actually CAUSING line noise,
or the 'editor just can't keep up'.)

--368--
J> Is CNet's custom screen a WB 2.x public screen? If not, please make
J> it so as I am very happy to hear that CNet requires KS 2.04 or up.
I haven't been able to get any of my 'public screen lister' prgs
to find the Cnet screen, so I guess it isn't public.  And nothing
mentioned upto and including the v2.40 readme about public screen
support.  But I too hope it is coming soon.

I haven't worked much with public screens, but I was under the
impression that it was as simple as adding 1 word to Cnet source
code screen definitions.  "PUBLICSCREEN".

But if it was that simple I'm sure Ken would have already added it now
that Cnet is WB2.0 only.

--369--
Would anyone (other than me) like to have a value in each BBSPORT
config that would set the priority of all background tasks spawned by
Cnet to x instead of 0 for that port?

You could set them all to -1.  Or maybe have port 0 (or whichever
one(s) you use for local log-ins) set to +1.  Or maybe have
paying/non-paying only nodes set higher/lower than the others.

(Faster background tasks would be yet another incentive for users
to gain higher access levels.)

--370--
Is TurboText available for D/L here?  Please U/L.

I'd love to see someone bundle 10 proven-Cnet compatible alternate
editors into 1 pkg.  (Nearly every Cnet I call allows 0 alternate
editors.)  I've got nothing against the 1s built-into Cnet, but
always think choices are best.

--371--
Ken's algorithms request...
Can we each tackle just the items we know about?
(And leave it to you to tie them all together.)

> #1 New stuff in any/all areas.
When a user posts or uploads the date/time is also stored elsewhere.
Next poster/uploader overwrites it with his date.
If a base has 10,000+ posts, only this separate date-element need
be checked to tell if the entire base has new posts/files.

A 2nd global-last-post-date (done similarly to the above) would
instantly tell if there's any new posts/files anywhere online.

2 disk reads instead of 10,000 tell if there are/aren't
new files/posts anywhere online.

--372--
> #2 Search for SPECIFIC messages to/from any user.
Most BBSes just search each post until match is found.  Slow.
A faster (but more complex) method would be to keep 2 separate,
numerically sorted, files containing the following structures for each
msg:

struct ToData {          /* BaseX:_ToData */
   long ToIDnum;
   long MsgIDNum;
};

Let's say a sub has 10,000 posts (just to show this method's speed).
A user wants to find all posts TO a certain user.  He enters a user
ID #, or handle or real-name (Methods already a part of Cnet convert the
latter 2 into a user ID#.)

A quick seek is made right to element #5000 in the _ToData file.
Its ToIDnum it is shown to be greater-than the 1 the user is searching
for.  So element #2500 is checked, ToIDnum is shown to be less-than
the 1 the user is searching for.  So #3750 is checked, then #4375.
Comparsion are made, and distances halfed, repeatedly until a match is
found.

You can find an element in a list of 10,000 with a MAXIMUM of 12 disk
reads.  Instead of 10000 reads, like linear searches need.
12 VS 10000 not too bad an improvement.

100,000 needs only 16 reads.
1,000,000 needs only 19 reads.
10,000,000 needs only 22 reads.
You get the picture.
(The MAX # of reads needed will always be the Nth root of the # of
msgs until results <2.  Or however you math guys say it.)

Of course a separate file would be needed for:
struct FromData {          /* BaseX:_FromData */
   long FromIDnum;
   long MsgIDNum;
};

10,000 of these 8 byte elements would be an 80K file.
10,000 elements of the 400 byte full-msg-headers (if this index/
pointer method is NOT used) would be a 4,000,000 byte file.

12 reads through an 80K file OR
10,000 reads through in a 4,000,000 byte file.
Again, not too bad an improvement in speed.

--373--
Ray Akey (Heavy Metal), it's been 4 wks, any more info about
programmer's docs?

--374--
Maybe some day Ken will have to break Amiga-Cnet up, and assign 1-2
people each to cover different parts of things.  I hope he does it
BEFORE he becomes so bogged-down with work/school/coding/wife/kids
that he just throws his hands up in disgust and walks away from Cnet.

 1 VOTE source
 2 BBSLIST source
 3 User docs
 4 Programmer docs
 5 Updated manual pages
 6 Each update's conversion executables (if any)
 7 Install-scripts for updates
 8 Answering user's msg posts
 9 Amaint source
10 Fidonet
11 Usenet
12 Filling mail orders
13 Filling mail update requests

A team of 5-10 would leave Ken much more time to do Cnet itself.

I hate to see Ken waste his time and talents doing things like writing
#6 when it could be 'farmed-out' to someone that knew even basic C
coding.  OR #7, simple DOS script writing 101.

(Of course, everyone writing source code would have to submit it to
Ken for checking before he (Ken) compiles it.)

--375--
Does the accounting system's monthly-charge only get assessed for
REMOTE calls?  Or LOCAL too?  (I think it should be both.)

I've tried both while logging on/off, while changing the date 01Oct,
01Nov, 01Dec.  Even when I get the "Monthly charge has been assessed."
string, the system still reports the same original $10 balance each time:

> Item                  Charge    Session
> -------------------- ------- ----------
> Monthly charge       $ .8300 $   0.0000
> -------------------- ------- ----------
> Connection time charges      $   0.0000
> Other activity charges       $   0.0000
> Total cost of this call      $   0.0000
> Remaining acct balance       $  10.0000
> Minimum balance allowed      $   0.0000

(I haven't checked daily or weekly assessments.)

--376--
1> From Cnet:bbsCharges...
1> 83    <-- MONTHLY Assessed 1st call in any month   (cents)

2> Online...
2> Item                  Charge    Session
2> -------------------- ------- ----------
2> Monthly charge       $ .0083 $   0.0000

The former shows amounts in cents.   (83 of them)
Online it becomes 0.0083 dollars (which is .83 of 1 cent)

--377--
Cnet- 1)  Very powerful accounting system.
      2)  Very powerful access-user-group template system. (Privs struct)
#2 doesn't allow you to set a default starting-$value for #1.
So you have to EA 100s or 1000s of accts 1-by-1, by hand, setting each
user's default starting-$ value.

I thought the whole idea of the account system was to start-out with
$x and slowly spend until 0 is reached.  (Although charges can be
assessed up or down.)

Cnet- 1) "Very powerful accounting system" becomes total useless
         and unusable *TO ME*.

There are MANY parts of Cnet that are like this.  1 little forgotten
thing negates an entire, super-powerful feature.  Big :-(

I'm sure Ken will add it someday, but if I had run Cnet for the past
3 years, 1000s of hours of EA'ing wouldn't have been fun.

--378--
> I was creating new subboard in Upload via online and I accidently
> entered '23' in "Enter the partition range :" (using AL command)
> and it should 0, so I go ahead finish remaining fields and then
> BBS locked up when I am done. (whole system lock up).  Could I abort

Since countless places in Cnet use the 0-23 values, 1 single new
subroutine could check them all.

BYTE check(BYTE n)
{
   return((n<0 || n>23) ? 0 : n);
}

Well worth "all the extra code" to stop GURUs in 20+ different
places.

--379--
DS>    Now that 2.04 is only being supported wouldn't it be nice to hit a
DS> button
DS> and bring up a screen and check each option like you've described for
DS> AMaint:)
Maybe something like:
Amaint would get its options via the CLI like "-t -n -s 47 -h" or
via that same string stored in a file.

A point & click GUI could create/edit this options-file for you.
(I'll write that code also.  No charge.)

--380--
LH> Brings us back to wanting a pointer set to your last message in the last
 LH> base that you didn't finish. I maintain that the amount of pointer space
 LH> would not increase appreciably.
 I'm not sure how Cnet's msgs are (or will be) #'ed internally.
 But what would happen when msg #6 is deleted and the next new post
 is 'inserted' at the #6 place?  (When your last-msg-read in that base
 is 48.)

 LH> BA> Remember, a BBS with *1000* Usenet subboards is still considered
 LH> BA> a 'partial Usenet feed'.
 How'd you like to read all those twice/thrice on each call?

 I've been getting email on the subject from several people.
 (Please post public, I can't possibly reply to it all, and I also
 want everyone's comments to encourage PUBLIC discussion of these ideas.)

 Suggestions like, on a 2nd RA attempt, just type:
 RG since 09-09-92 12:23
 instead.
 (After I've check my exact date/time of log-in, remembered the SINCE
 keyword, remembered the (in my opinion) very odd Cnet date format,
 remembered if it's 1-12 or 0-23 clock, etc.)
 I can just imagine telling all my novice users to do all that.

 Others suggestion,
 noting all the subs, subdirsubs, subsubdirsubs, etc (maybe 100s)
 that have "*", and then list all those:
 RG 1-3,7,12,29-35...
 (I'm not sure on the format, nor if subsubdirsubs will be searched,
 nor if "*" correctly marks subsubdirsubs, etc)

 So anytime so many sysops are trying to come up with workarounds
 (and drastic ones at that) it should be a msg to Ken that this
 matter needs to be looked into and fixed.

 I appreciate all the above opinions that say "all is fine, there
 is absolutely NO problem with the multi-RA matter", but also let those
 that disagree get their msg to Ken.

--381--
LH> to 100% as possible. Version 2.30 was close but there were some little
 LH> bugs that were truly annoying in it that ever got fixed before you went
 LH> on, and I thought that they should have been fixed. Thanks for your time!

 I think the real issue is:
 1)  Updates for the purpose of bug-fixes
 2)  Updates for new features.
 If we wait 6 weeks for each of #2, so be it.  (But 6 wks can feel like
 forever.)

 But what is the 'release schedule' for #1?  Personally I think it should
 be 'as soon as bugs are found/fixed'.  (Of course, if #1 has a few #2
 great-new-features, I'm not going to complain.)

 Am I correct in assuming that none of the bugs from v2.30 12Sep92
 ever got fixed and released?  50 days with some systems taking 100
 calls per day, you can really 'feel' the bugs.

--382--
Jamie, I hate to get off-track here, but Ken sounds like he's
very interested in caller-id and voice support.

But I have a question about those features themselves.
Namely VOICE.  What are few of the things I can do with it?
Can the Supra answer the phone, detect a voice caller, and re-ring
the call through to a standard voice phone or answering mch?

I'd love to have all my BBS lines double as voice lines too.
(At 0 extra $$$s per mo.)

I've tried to do this with several of those fax/modem/voice
black-box auto-switchers, but they all had hoaky schemes that
required the voice user to already know to type 222, etc.

I need 100% transparent usage for all voice and all BBS callers.
They call with a modem, they get the BBS.
They call voice, they get me.

--383--
Does anyone know a way to make all quoted-text in msgs appear in a
different color or style (underlined, italics, bold, inverse) than
non-quoted text?  (If you have color/ansi turned ON, of course.)

I guess it would just need a 'start-quote' and 'end-quote' entry in
BBSTEXT.  (As always, forwards slashes are backslashes.)
> /R1       /; inverse ON at start of quote /;
> /R0       /; inverse OFF at end of quote  /;

(No, MCI comment-codes "/;.../;" aren't supported.  Just an example.)

--384--
UN> 1> Kill all messages [No]? No
UN> 2> Kill [A]ll msgs or [O]ld previously read msgs?

UN> Note that you will only get that prompt if you haven't killed any of the
UN> mail during the mail read, and it's easy enough to kill mail as you read
UN> it.
I was mainly interested in helping those users (I'm guilty myself) that
had a pile of old mail (the real disk waster) and didn't want to go
through reading every piece again for the express purpose of getting
rid of it.  (Would you ever take out the garbage if you had to examine
(re-examine actually) every single piece of it?)  But didn't want to delete
ALL their mail (even stuff they hadn't read at all.)

Am I correct in assuming it is currently:
1)  Delete all your mail (new, old, read, unread, wanted, unwanted).
2)  Re-read every piece of your (old, read, unwanted) mail before you
    can delete them 1-by-1.
Choose 1 or 2.

That's where "Kill [A]ll msgs or [O]ld previously read msgs?" comes
in.
It in no way prevents the current #1 or #2 choices either.  (Just
adds a very powerful, useful #3.)

--385--
 BA> Cnet doesn't even allow global file marking or D/Ling at all.
 BA> But, how long would "SG 'MyFile.LHA'" take under the same
 BA> circumstances?  (And then mark or DL 'by-hand' after matches are
 BA> found.)
 UN> BA> Cnet doesn't even allow global file marking or D/Ling at all.
 UN> You can do a "BG 'MyFile.LHA'". Then you can get a list of all the matches
 UN> complete with descriptions and decide if you want to mark them. Then just
 UN> do a "DS" afterwards. Perhaps CNet should have Download Global and SElect
 1) Browse or scan
 2) Mark them
 3) D/L those marked
 "Download Global" 1-step instead of 2-3.
 "DG" or "*G" would be for those of us that already know what we want.
 Especially on multi-disk D/Ls, so you don't miss any sides.
 "DG FontSet*.LHA"
 I often see the 5 disks scattered around instead of listed together
 during BA scans.  (I have my sort set to newest-1st so I don't know
 what causes that, either.)

--386--
Would anyone find this a handy way to help find a user's name, as
well as save a lot of typing?  At any prompt that expects a user's
name (ie MS), as you start typing, the BBS fills in the remaining
characters with the closest matching name possible.  The more of the
name you typed, the closer these guesses would get, maybe only the 1st
3-4 chars would be needed, the match would appear, and you hit return
to accept it.

Or would these as-you-type-guesses confuse/scare the user?  Or
would a new account user-flag be needed to make everyone happy?
> Do you want user input expansion ([Y]es)?

--387--
Yikes, I never would have been able to track down the source of
CPS slow-downs as 'amaint hung'.

> AMAINT would freeze right at this point!
N-O-T-H-I-N-G should make a bulletproof executable freeze.
(Especially the 1 that we run automatically, unattended as an event
more than any other executable.)  Ken?

--388--
If Ken has the disk space he might want to let 1 of his msg bases
get extremely large to check in real-time how fast/slow things
are running as new search/sort methods are tried.

--389--

Anytime a file exists in a user's dir named 'SkipPosts.txt',
containing a list of upto 10 handles, all posts by those users are
automatically skipped while you are RAing.

You can skip users that post too long posts (me).
Too many post (me again).
Or skip those users that you just don't care to read for whatever
reason.

(You could even skip your own posts, since you already know what they
say.)

Large UNIX BBSes call this feature a 'twit-filter'.  I haven't seen
it available on any of the 53 Amiga BBS packages yet.

T> cases. And as a result the user with that opt/flag enabled would really
T> miss out on some great discussions.
But it would be THAT user that specifically chose to 'miss out'.
And if he chose to, I can't see them being 'great discussions'.
(1 person's GREAT DISCUSSION is another person's I DON"T WANT TO
WASTE MY TIME/MONEY READING THIS)

What about all the users that chose to drop an entire msg base?
Shouldn't they be forced to read all that stuff that doesn't
interest them because if they don't, 'they're missing out
on some great discussions'.

Let the user choose, and you can't go wrong.

--390--
>READ GLOBAL SINCE 14-10-92 UNTIL 16-10-92
Shouldn't this read everything posted between 14Oct92-16Oct92, inclusively?
It's reading replies posted as recently as 01Nov92, too.

--391--
DM> All this is a nice suggestions, but remember one thing.. Not everyone will
DM> be updating from the same previous versions numbers..
These are the users that will get to D/L the full 1,200,000 at 2400bps
DMS versions.  Everyone else gets the much smaller 'LHA' release.
Ken can create all DMS and LHA updates by executing the same, single
DOS script.

DM> Ken changes so many things so often, I am sure that it is difficult for
But they are released only every 6+ wks 'often'.  So I don't see this
as a problem.

Ken runs a script.  (I'll even write it.)  It does 4 things:
1)  Reads a 'last time packed' date.
2)  Packs all files that have been changed since that date.
3)  Resaves current date for next usage in #1 above.
4)  Automatically creates 'Update.txt' file, short/long descriptions,
    telling users 'this date' and 'last date' for this update.

Does several things we can't do with DMS:
1)  View inside the archive online.
2)  Save 300 D/Lers about 600,000 bytes each.  (Why tie up the system
    D/Ling 18,000,000 bytes THAT YOU ALREADY HAVE SITTING ON YOUR HD?)
    (Why pay the phone company if you are calling here from BFE?)
3)  Want to know exactly which files have/haven't been updated in
    this release?  Easy.  Every file in the LHA.
4)  Unpack directly to any sized partition on your HD.
5)  Selectively unpack only any 1 (or more) files that you need.
(See my AutoPack.rexx file.)

JS>   Still, the POINT is well taken.  All we have to do is establish a proper
JS> METHOD for insuring that all files are of the correct version.
All of Cnet (and all pfiles) should be doing this already (but don't)
regardless of DMS or LHA packing methods.  The end user should NEVER
have to worry about which of the 5 executables go with which of the 20
updates (100 prgs total).  (Why wait until Cnet is 25 executables &
libraries and there have been 300 updates.  7500 possibilities for
wrong combos.)  Let the software keep track of things like that.
Report/abort if it can't find a version of a prg that it needs.  Don't
just GURU for unknown reasons.  Here's a excerpt from a BBS I was
going to write in 1987 and abandoned:

struct MasterVersions {
SHORT HouseKeep_ver; /* Version of Master House Keeping Task (MHKT) in use */
SHORT HouseKeep_rer; /* Version of Master House Keeping Task (MHKT) in use */
SHORT Master_ver;    /* Version of MasterBBS in use */
SHORT Master_rev;    /* Revision of MasterBBS in use */
SHORT Arexx_ver;     /* Version of Arexx needed (min) */
SHORT Arexx_rev;     /* Revision of Arexx needed (min) */
SHORT library;       /* Version of libs:Master.library in use */
SHORT userdata;      /* Version of user-data structures in use */
SHORT msgdata;       /* Version of msg-data structures in use */
SHORT filedata;      /* Version of file-data structures in use */
SHORT menudata;      /* Version of menu-data structures in use */
SHORT padding1;      /* Reserved for future expansion */
SHORT padding2;      /* Reserved for future expansion */
SHORT padding3;      /* Reserved for future expansion */
SHORT padding4;      /* Reserved for future expansion */
SHORT padding5;      /* Reserved for future expansion */
};

L> P.S. You can make a copy of your RAD: device in your mountlist,
L> duplicating it and just call it RAD1: or something. Then just mount it
L> like RAD:
Regardless of conflicting unit #s and LowCyl settings?
Regardless of wrong sized HighCyl settings?
These are the values I need.

--392--
> What file and what format is used for the charges-files?

Should be cnet:bbsCharges0.10 for example.
If you have a CHARGE cmd in your events, change the filename's
"0" to match it.

The 10 affects only level 10 callers.

Be sure and grab a BBSCharges file from your latest master disks.
Much has changed there.

There were still many things that it can't do.
1)  I couldn't get $-per-month to work at all.
    (I hope no one was basing their last 20,000 calls on this amount.
    Or you just lost big money.)

2)  I couldn't get it to 'do-nothing' when a caller's $$$ runs out.
    (You can raise/lower his access or cut him off immediately and
    permanently lock him out.
    Yikes, I just wanted to send him some mail about it.)
Maybe Ken will support:
-2 <-- NEWGROUP   New AG if balance prevents logon (-2 == do nothing)
(But I'll be able to tack an MCI arexx script into the BBSTEXT
line that gets displayed.)

3)  Some $$$ claiming to be cents are really 1/100th of 1 cent.
    (Quiet a difference.)  Or was it the other way 'round?
    (Again, you just lost/overcharged your last 20,000 callers 100x
     more than you meant to.)

The part of your BBS that brings in the $$$ is no place for missing
features, typos, or bugs.  Does anyone know approximately when the
accounting system was 1st installed in Cnet?   It should be rock-solid
by now.

--393--
Suggestion #393...

New value added to Cnet:BBSport#...
60 <-- Calls <x secs don't count in anyone's calls-per-day limit.

Callers remaining online <x secs would benfit by:
1)  Allowing users to quickly check mail/posts/new-files and log-off
    freeing up the system for other callers.
2)  Help callers that are accidently cut-off for whatever reason.
3)  Auto-callback validation.  (Currently uses up 2 of your calls/day.)
4)  Setting value to 0 remains 100% compatible with current Cnet,
    "count-all-calls towards user's limit".
5)  Doesn't count calls where "I'm getting line-noise, I need to
    log-off and call back."
6)  "I wanted the 14400 line, but got 2400, I have to call back."
    (SysOp's:  PLEASE mention your phone#s and modem-speeds BEFORE
    I have to log-in.  (And search all over for this info.))
7)  Upcoming event kicks me offline early.

All of the above currently counts against you in your calls-per-day
limit.  (Even if your call was 1 sec long.)  This makes me set all my
callers call-per-day limit artificially high (ie 10-20/day) to
compensate.  This new proposal would allow truer settings.

"3" and "60" would allow 3 calls per day as long as they were
more than 60secs each.  (This is per port.)

--394--
I take it I can only [W]rite a short description later, not a long one.
(Other than replying.)

Suggestions #394...
W45
>This item already has a short description!
>This item already has a long  description!
OR
>Write short description (Yes):
>Write long  description (Yes):
(Which ever one is missing.)

We'll always be certain that the original U/Ler's long-note is at #1.

--395--
D>    Also, The "bbs.idata" protection file has your serial number(325)
D> stamped within it's framework as well.
This whole discussion would come to an end if Ken/Jim would come
out and officially state, "your ser # need not be guarded with your
life" or "never tell your ser # to anying living soul", as the
case may be.

In cnet's quote-buffer, I see:
D> ~[0m~[10C~[33m----------------~[7C~[31m---------~[7C~[33m----------------
D> ~[9C~[42m  ~[34m(716)/823-9892  ~[35;40m <~[36m*~[35m> ~[44m  ~[37mDotoran
D>  ~[35;40m <~[36m*~[35m> ~[42m  ~[34mFRONTIERS BBS!  ~[0m
D> ~[10C~[33m----------------~[7C~[31m---------~[7C~[33m----------------~[0m
But in the original post, as well as in my quote-add, all is fine.

--396--
JS>  There are a lot of pitfalls to letting the idea go FURTHER, but it is
JS> intriguing!  Can you imagine how interesting it could be if a binary file
JS> could be 'attached' to ANY text message created with the text editor?
JS>
JS>  I think that idea is a little TOO powerful, and I am not asking for it,
JS> YET!
JS> But I WOULD like the expanded file mail ASAP.

Concerns like: 100s of private files changing hands just for the
benefits of the 1 person (in eachcase) that it's addressed to.  Hard
drive space.  In many cases I'd say, 'if you are going to UL something
why not let all of us DL it?'.  (Except in very private/personal
situations.)

I was thinking more along the lines of 'attach files to
pre-existing public posts'.  (Heck, even if it can't be made so
OTHERS can attach, at least let the original msg poster add/remove/
update a file attachment.)

--397--

All future posts here by me are regarding v2.40e (or better)...

I see that a NULL CMD is no longer supported in the CONFIG's event handler.
I assume it was used pre-v2.40e to do 'clear-line', 'don't answer',
etc without doing anything else.

--398--

Does anyone know why:
"NEWSHELL CON:0/0/640/40/WindowName/CLOSE/SCREENCNETSCREEN0"
opens on WB instead?

ListPS 1.01 (27 Jul 1992) by Tomas Hellberg lists all screens:
> Workbench (Workbench Screen)
> CNETSCREEN0 (PRIVATE)

Do I have to do anything to set Cnet up as PUBLIC?

--399--

During "AC DETAILED" you can't answer NO to "More[Y,n,c]".

--400--

BBSTEXT:
> Daily charge has been assessed.   Hit any key \g1
> Weekly charge has been assessed.  Hit any key \g1
> Monthly charge has been assessed. Hit any key \g1
Why doesn't this hold for a key?

--401--

If I have AG set as the default idle-display for a node.  When the node
is opened, the top 1-2 lines of the graph are hidden under the stats window.
Closing + reopening = all is fine.

--402--

No need to display:
> Item                   Charge   Session
> -------------------- -------- ---------
> -------------------- -------- ---------
if there aren't any charges anyway.

--403--

I'm assuming the "------------------------" in CONFIG's archiver
setup is showing the archiver #.  It's displaying 2 as
--------------------2--
instead of
--2--------------------
(In fact, can't it display just "2" in English?)

--404--

How do I set different colors for each node under 2.40e?
I always ran all 2-color screens.
Node0: Black on white (local)
Node1: Red on black   (remote)
Node2: Green on black (remote)
Node3: Blue on black  (remote)
Node4: White on black (local)
Color-coded so I would know right where I was at all times.

--405--

In BBSCONFIG3 .LZH is missing its trailing "double-quote" for TESTING.

--406--

CONFIG: When setting up 2 or 4 color screens the gadget correctly
displays COLOR 0/1 or 0/1/2/3.  8-16 color screens attempt to go
by color name (which has no meaning since we all use different ones.)
If this was done on purpose, maybe mark them "BLACK 0" "RED 1" "GREEN 2"
instead.

--407--

I'd love to see a color sample while I run the color sliders in CONFIG.
(But I guess it would mean opening a 16 color SCREEN, so I don't know
if it's worth it.)  But without it, who knows what text-color 5/1/14
looks like on a 8/12/4 color background?  Maybe the color-sample screen
could show the word "CNET" colors^2 times, showing all colors of text
on all colors of backgrounds combos.  No point in showing callers text/
background combos that may come out invisible or near-invisible when
online.)

--408--

It's common for a screen full of string gadgets (like CONFIG's) to
move-on to the next gadget when RETURN is hit.  And move back/up
1 gadget if SHIFT-RETURN is hit.  Really fly through set-up.

--409--

CONFIG:
Please consider having SAVE, just save.  (Like it says.)
And a close button to exit.  (Like intuition says.)
When editing 100s of values it's good to SAVE work-in-progress often.

Currently it's:
SAVE (which also exits) then re-run CONFIG.
SAVE (which also exits) then re-run CONFIG.
SAVE (which also exits) then re-run CONFIG....

--410--

When toggling a screen between clock/AM/AG they all nicely scroll up
from the bottom.  Except when going *TO* CLOCK.  It starts to scroll and
then plants it over the existing display.

--411--

> Item                   Charge   Session
> -------------------- ------- ----------   <misaligned in BBStext
> -------------------- -------- ---------

--412--
So this LHA *REPLACES* the full disk #2?
Or is it just offered in additional, as 'only these things changed
in v2.40e'?

Personally, I think both are needed:
1)  2 full disks DMS of everything.
2)  2 LHAs of just updated files.

User's previously running v2.30 would just need #2.
Those running earlier releases would need #1.

--413--

>This base is for suggestions, not questions/comments/etc...

1) Bug reports
2) Suggestions
3) Questions
4) Comments
Sometimes ideas start out as one of the four and end up being
several of the others, also.  I'd really rather just keep all my
lengthy babbling in 1 item and everyone can quickly find/read or
skip/PASS it.

NS> I'm on my high horse about things being in their proper subs today, so
NS> excuse me, but this belongs in the "Suggestions" subboard.
NS> On a hypocritical note, I agree with Jim ... I've wanted "File Attaching"

I know all too well the importance of keeping items on topic and keeping
topics in their correct subs but I've always felt the #1 priority
was:
1)  The open, free, welcome communication of ideas by all.
   (I see so many sysops begging (sometimes try to force) more
    msg base use and less file base use.)

With secondary concerns about:
2)  That paragraph is 65% suggestion and only 35% question, therefore
should be posted elsewhere.

--414--
CONFIG IS SLICK!

Ken, please consider adding another BOOL-checked-gadget to CONFIG's
event scheduler.

It would be a simple ON/OFF switch for each event.
Allowing us to keep a massive # of events "available" and decide
which to turn on/off at the flip of a switching.
(Instead of NEW/ADD/REMOVE'ing them each time.)

(ACCOUNTING might need it too.  That's a lot of info to NEW/ADD/REMOVE.)
Or maybe effect ACCOUNTING differently:

A new EVENT CMD parameter for CHARGE: "-1" which allows a "free-time"
(weekends or maybe just major holidays) when all charges are waived.

Even if I use only 1/2 of the available charges, and 1/2 the available
access groups, and only 2 schedules, that's
(70/2)*(24/2)*2
1680 values I need to type in.
Pretty hard to clear/re-enter too often.

Looks like the accounting system can quickly be turned on/off just
by having a schedule with all values set to 0, and switching to it.
Not so easy for EVENTS (without clearing that event).

--415--
A> Thanks (still looking for those 2.04 chips)

Are these hard to find?  Another WB2.0 only BBS (no names please) is
offering the chips for sale also.  Ken/Jim, are you interested in selling
2.04 ROMs?  (I guess profits would be minimal, but users could order
a CNET+ROM package or either item individually.)
(Either way, users could at least get their hands on the 'now
necessary' chips.)

--416--
LH> Perhaps if it could be a compromise of these... could we not just have it
LH> in full disk .lha format, if the only files that changed doesn't work out.
LH> This should still be smaller than the DMS archive and would be easier to

So you mean offer:
1)  2 full DMS disks
2)  2 'just new files' LHA disks
3)  2 full LHA disks
Users would D/L any 2 of the 6 files.  As long as everything is clearly
marked.  But I think things would start getting more confusing instead of
clearer/easier (which was the original idea).

--417--
What would be causing my log-off to look like:
> Other activity charges       $   0.0200
> Total cost of this call      $   0.0205
> Remaining acct balance       $  13.8795
> Minimum balance aQcr3,I,Z  ! __   1w~
> NO CARRIER

Looks like it's dropping-carrier before everything is displayed.
> "The remaining line(s) of accounting info."
> "@Connection closed"
> "@Log-off complete"
(and whatever else is 'normal log-off'.)

--418--

Does Cnet have log-support for the accounting system?
>28-Oct 18:01  Port-0  Sched-1  MONTHLY .50  John Doe (134)
>28-Oct 18:01  Port-1  Sched-2  DAILY   .10  Jane Doe (291)
>28-Oct 18:01  Port-3  Sched-0  DLOAD   .04  Pat Robertson (69)
>28-Oct 18:01  Port-4  Sched-1  DLOAD   .04  Jim Baker (344)
>28-Oct 18:01  Port-1  Sched-2  UDBASE  .02  John Q. Pubic (893)
>28-Oct 18:01  Port-0  Sched-3  BASE    .06  Henrietta Doe (532)

I'm sure someone would write a pfile that would parse this log and
present the user with a detailed, itemized bill at the end of every
month.

(As well as tally-up the total # earn per item, per user, per day/wk/
month/etc for the SysOp to keep an eye on.)

--419--

Would anyone like to see an option in the accounting system that would
present the user with text when he attempts to execute a cmd that is
going to ring-up charges?
> This command costs x cents to use.  Continue (Yes)?
> This command costs x cents per min to use.  Continue (Yes)?
> This command credits x cents to your acct.
> This command credits x cents per min to your acct.

Perhaps a new BOOL-checked-gadget in CONFIG marked:
> "INFORM USER FIRST [ ]".

--420--
> Mark this item as file-mail  [No]? Yes
> Enter a subject for your message.
> : The file you needed.
>
> Mark this item as file-mail  [No]? Yes
> Enter a subject for your message.
> : The file you needed.
>
> Mark this item as file-mail  [No]? Yes
> Enter a subject for your message.
> : The file you needed.

If the user responds YES to file-mail, could Cnet display a different
line of text from BBSTEXT.  Like:
> Enter the name of the file to attach:

Or better yet, allow independent "Subject"s and "FileName"s.
(It's not always desirable to have them be exactly the same.)

--421--
Jim, I just noticed I can do private file-mail.  But it looks like only
from LOCAL.  Is this right?  It's already partially installed in Cnet?

--423--
MC> your sub-board.  YOU DONT.  you get charger wether you are a sub-op or
MC> not.....
1) Sub-Op
2) Unlimited D/Ls
I'd always like to control these independently.
(And since there's alreay separate flags for each, it's ideal to keep
them that way.)

1) Sub-Op with Unlimited D/Ls
2) Sub-Op with limited D/Ls
3) Non-Sub-Op with Unlimited D/Ls
4) Non-Sub-Op with limited D/Ls
Set things e-x-a-c-t-l-y like you want.  Just set both flags if you
want #1.  (But the other 3 are still available if the need arises.)

--424--
BB> User Input Expansion (UIE)...
NS> I agree. Visualizing how it would work in my mind .... I think it would be
NS> a major CPU hog. Picture CNet trying to do this:

Per ken:
> 118.  CNet now maintains a new format for the old "bbs.alpha" (alpha
>    ordering of users by handle).  Now, this primary key information
>    is actually sotred in MEMORY (as it should be), and handle searches
>    are now BLITZ fast.  No more waiting several seconds on large systems
>    just to have the system tell you an account you are mailing is not
>    there.

I don't think UIE would be a CPU hog or slow.
Let's put this 'BLITZ fast' action to the test.

What are your ideas on my main, original concern about this?
The user types "The Ha" when trying to send email to "The Hacker" and
"The Hard Salami"'s name pops up out of no where.  Confusion or handiness?
(Of course the user would have had to physically turn on his acct's
UIE flag, so he'd be getting what he asked for.)

--425--
 AK> Ken, this is a KICK ASS update. I LOVE the new feature where I can MODIFY
 AK> the configuration file and then NOT have to re-boot the BBS to have it
 AK> take effect. That feature ALONE is WELL worth the download. (Now just ad
 AK> that to BBStext and BBSMENU and it will be PERFECT).

 CONFIG'izing BBStext and BBSMENU looks like it's going to be harder to
 do.  I'd be happy with:
 1)  Use an editor to make the changes (the online editor or my favorite
     word processor offline.)
 2)  Run a DOS pfile (online from remote or offline from the CLI) and
     it reads BBStext and/or BBSMENU and activates the changes.

 I'm always torn between:
 1)  Slick intuition point & click prgs.
 2)  Being able to make changes to the system from remote.
 Anyone have any ideas on letting Cnet do both?
 (I thought that was the whole idea behind VDE.  And thought Ken would
 have done a VDE CONFIG instead.)

--426--
Does CONFIG need its separate CREDITS and COSTS for Netmail?
I thought we just had to specify + or - charges.
(These are most probably for another purpose.  Any docs for CONFIG,
planned or present?)

--427--
Ken, I love the new, expanded arexx cmds.  Please consider 3 more that
would greatly encourage arexx authors to allow BBSTEXT, as well as,
their own BBSTEXT-style external string-files.  (User-customize and/or
foreign language support.)

BBSTEXT {n}
  RESULT will contain line #n of BBSTEXT file.

TEXTLOAD {drive:path/filename}
  Loads this text file (for later TEXTLINE calls).

TEXTLINE {n}
  RESULT will contain line #n of current text file,
  (previously loaded with TEXTLOAD call).

--428--
Does anyone think Cnet should allow the marking of files that
aren't currently online?  (Currently it doesn't.)
1)  Mark them anyway.
2)  Request the sysop to put them back online.
3)  They're already marked & ready for D/L on your next call.

--429--
#128.> When using the DS command, it is possible to remove RANGES of
#128.> files.  Also, files removed here will "re-appear" after the
#128.> download is complete.  This allows downloading PORTIONS of the
#128.> select buffer at a time!

This can be very handy.  Might want to change the line in BBSTEXT
to reflect the VERY different way SS and DS now work.
The former, PERMANANTLY removing marked entries.
The latter, only TEMPORARILY.

(Allow 2 different lines in BBStext, that is.)

--430--
David Prothero writes...
>I really had to jump through some hoops to accomplish this (at least being

How many people think "reading files within archives" is an
important enough, and commonly needed/used feature, to include it
as a standard part of cnet?

In addition to VIEW & TEST, CONFIG/ARCHIVERS would also have a READ
string.

I'd love it.
SysOps not needing it would just leave those lines blank.

--431--

>My BBS can no longer find my logs.

Read #134 in Ken's ReadMe.
> SYSDATA: must contain one new subdirectory entitled LOG.
He doesn't say anything more about it, but guess what is now
kept there?

--432--
D> "calls" log. What would happen is the activity of the logged on user and
D> YOUR activity would be combined and intertwined with each other. The way
D> around it is to create the "port" file. When a user logs off, the entry
D> for THAT port is assembled in the "port" file, then the "port" file is
D> appended to the "calls" log, and lastly, the "port" file is deleted from
D> the system.
This is indeed how this works.
But what would happen if this tmp port# log wasn't deleted AFTER every
call, but BEFORE the next call?  You could see what the last caller
on each port did rather than trying to read the last x lines of a giant
log (that had info on 8 ports).

If I want to save this current portX log by adding a DOS COPY cmd to
a string in BBSTEXT, which line do I edit?  (That is, what line of BBSTEXT
is displayed immediately before PortX-log gets deleted?  And it would have
to take into acct, O, O!, dropped-carriers, exits via pfiles, exits after
D/Ls, events-cut-offs, sysops cut-offs, system GURUs, system lockups, etc.)

--433--
>It needs to be moved to SYSDATA: and renamed to...
Ken, until a bug-fix release is issued, could a few of these things
be mentioned in NEWS AT LOG-IN to those with registered-sysop access
here.

If it's too time consuming, maybe you could appoint 2-3 trusted
callers to the task, and give them ADD-NEWS access.

--434--
JS>   This would be a good thing to address.  I recall in the deep dim past of
JS> the C64 version I had a couple of sysops who had hard drive crashes and

If Ken is going to add the ability to edit a few system vars, I think
it would be better to pick the 20 most likely ones and add this support
'all at once' instead of modifing the code 20 different times.

There's room to squeeze 1 more "Sys Vars..." gadget into CONFIG's main
screen.  What would you most likely need to edit?

# of calls
Feedback
Mail sent
Mail to #1
Posts
Responses
Gfiles read
Pfiles launched
NewUsers+Guests
Upload files
Upload Kbytes
Download files
Download Kbytes
Minutes use
Minutes idle
Charges

--435--
DS> I think his complaint is that you can no longer set your macro up to do
DS> multible tasks to each base.  I miss it to, but such is life.
Maybe Ken could allow MCI or ANSI codes to be easily inserted into
online-macros and func-keys so they could still jump the cursor
up/down while in a VDE screen.

LH> What I would prefer is to be allowed a choice between point and click or
LH> the older interface. This way those that like it can have it thier own way
LH> without forcing it on the rest of us... <whichever way they like it>.
Maybe if each called an external prg.
Those wanting VDE would only need have that executable.
Those wanting the old way would only need have that executable.
Those wanting both (me) could add both.
(Ken, do you still have the old-way source?)

--436--
MC> thanks ill give that a shot....  I think I was using modem 3  did the
MC> numbers change from the old spiral bound manual?
No charge.

I can't even find it mentioned in the spiral manual.
I've taken it and penciled in everything in every read-me, every new
feature, and every post here I can find.  I should have in my hands the
most complete Cnet manual on the planet Earth.  And there's still a lot
of stuff missing.  Each passing week Ken is just making it harder and
harder for him to ever issue a 100% updated, 100% complete manual.
New features just don't get fully explained (and some not mentioned
at all) in the Readme's.

--437--
Is the following "AM" line in BBSTEXT someplace?
>SETUP  11-Nov-1992  6:27p PERIOD  7-Nov-1992  8:47a TOTAL  14-Apr-1992  4:19p

I wanted to change its spacing into something I find more readable like:
>Setup: 11Nov92 6:27p         Period: 07Nov92 8:47a       Total: 14Apr92 4:19p

Which could you read easier at a quick glance?

--438--
"List last 10 callers" has to be 1 of the most used, most popular
log-in displays.  Exactly what am I supposed to do with all this info?

Does it really matter that John Doe called, 5 callers ago?  But Ron
Williams called 7 calls ago?  And what/why should I need to know, 10
first/last names, 10 handles, 10 cities, 10 states, 10 log-in dates,
10 times, 10 call lengths, and all the other info that many times is
display?  And it really gets good by my call to my 56th BBS today.
(And I've had to sit through it on all of my last 97 calls to it.)

I feel like I'm missing something here.
And this is "The Most Popular Log-In Program Ever Written"?
I don't get it.

--439--
Am I correct in assuming that anyone with the "edit-own flag" set can
edit their previously posted responses at any time?

Anyway to make "ED" smarter with any 1 of more of the following?

Users can only edit their own responses:
1) if no one else has yet responded after them.
2) if less than x hours/days have passed since they posted.
3) only during this current call.  (or maybe as late as 'next call'.)
4) the header will be marked "(EDITED)".
5) if no one else has read it yet.

Anything that will stop callers from posting 1 thing and then changing
it and claiming that they never said it.

I am NOT trying to eliminate the very handy feature of 'correcting text'.
Just putting some SysOp control over it.

--440--
Everything in SysText:* root dir, help/*, menu/*, and new/* subdirs
are all text files.

But all SysText:VDE/* files are data, not text.

SysData:VDE/* would have been a better place for them.

--441--
I'd like to see Cnet get horribly, horribly specific about what/where
a user is at any given time.
(Especially as systems grow larger and larger.)

Either keep the existing DOING-string as-is and add a more specific
FULL_DOING-string.  (Even if Cnet itself doesn't do anything with it,
but external WHO cmds, arexx, pfiles, etc could read/write to it.)

Or break DOING up into WHAT-strings and WHERE-strings (and maybe
even SINCE-strings).

WHAT           WHERE                              SINCE
----           -----                              -----
D/Ling        FileName.lha in udbase #23          12:23pm
VOTING        On topic #3 "The Election"          12:34pm
U/Ling        GreatGame.zoo in base #12           11:09pm
BBSLISTing    Adding 313-473-2020                 10:12pm
CHATing       With SysOp (1)                      08:12pm
OLMing        Sending to Bill Beogelein (123)     12:31pm

For 8-10 (and growing) line systems "HE'S 1 OF 10 USERS THAT ARE
CURRENTLY UPLOADING SOMETHING, SOMEWHERE TO SOMEONE, OPEN AND GO
SEARCH ALL YOUR NODE WINDOWS TO FIND OUT WHO/WHAT/WHERE/WHY YOURSELF"
just won't be enough some day.

--442--
Would anyone find it useful if a SHIFT-CLOSE-GADGET click on the control
panel would instantly shut everything down?  No verify requester, no
icon-mode.  Just instant shut-down.

--443--
Has anyone seen an Amiga executable that will display a crude (but
usable) ansi display when given an IFF or GIF picture file?

I thought I saw this as part of an IBM BBS long ago.  It allowed users
to get a rough idea of what a picture-file looked like, right online,
before D/Ling it.

--444--
Ken, any plans on allowing WB2.0 style '1st-char' keyboard equivalents
to CONFIG?

You could re-word a few gadgets where necessary...
 [A]ccounting->[C]harges
 [E]vents->[S]cheduler

--445--
Ken, any plans on adding a new "/Tn{v}" test var to MCI?
It would return 0 or 1 based on if current caller was LOCAL or REMOTE.
We could skip/access from REMOTE certain strings/prompts/executables
as need be.

--446--
Could CONFIG be given a "smart-save"?

If you try and exit-without-save (CANCEL) after changes have been
made it'll inform you of such.

If you try and SAVE without any changes having been made (thinking
you have), it'll also inform you.

--447--
Does anyone think U/Lers should only get byte-credits based the "after
transform" size of their U/Ls, not the "pre-transform" size, like
is does now?

ie.  U/L a 200k .ARC, that is auto-transform into a 150k .LHA, the
U/Ler just gets the 150k credit (or % thereof).

(Currently, users actually get rewarded MORE byte-credits if they
use .ARC instead of .LHA due to its (arc's) lower compression.)

It wouldn't have to be done per-user, or even per-sub basis, just 1
single, global-variable for a system-wide effect.

> CONFIG
> Re-compression adjustment: 25%

U/Lers would get x% of the difference (old_file_size - new_file_size)
subtracted from their byte-credits.

SysOps not wanting this feature would set it to 0%.

SysOps wanting to subtract the full amount of the difference: 100%.

SysOps wanting to penalize those that U/Led old/out-dated/low
compression-types could set it to 150% (or more).

--448--
>What does 'subboard use rate do?"
>Does anyone know?

I'll take a guess.
You use EL to set it to 1,2 or 3 ( or 0 for none).

You use CONFIG/ACCOUNTING to set:
Per minute using subboards (rate 1)
Per minute using subboards (rate 2)
Per minute using subboards (rate 3)

If you EL-set your area #47 to rate #2, then users are charged that rate
as listed in CONFIG/ACCOUNTING.

Personally, I think the same wording should always be used when
things like this need to be cross-referenced.

If Ken has EL call it "subboard use rate", then CONFIG should call it
subboard use rate 1
subboard use rate 2
subboard use rate 3

Or if he wants CONFIG to use "Per minute using subboards", then EL
should call it "Per minute of use rate", or something like that.

Cnet is extensive, the last thing we need is to remember *2*
different terms for each feature.

--449--
> 'BBSidentify' BBS
> 'transmit' RESULT

Gives "CNet 2.36" when running v2.40e.

Ken, would it be possible to have just 1 single master-version string
within Cnet?  It would be used in every window/screen titleBar, MCI, Arexx,
shown-at-login, everywhere.

Just update that 1 string for each realease and you'll be certain that
it's updated everywhere.

                                      -

> 'BBSidentify' SYSOP
> 'transmit' RESULT
Isn't working.

                                      -

Also, doc correction:
> NAME		: Your BBS's name   (bbstext line 9)
> SYSOP		: Sysop's name	    (bbstext line 10)
These are no longer lines 9-10.

--450--

NS> Maybe SHIFT-CLOSE-GADGET would be good. I personally like the requestor
NS> because I'm always moving so many windows around at the same time that I
NS> might accidentally ---log everyone off, no matter what they are doing
NS> (uploading, downloading, posting, sending mail) and kill CNet---
Would this 'accident' be likely with SHIFT-CLOSE-GADGET?
(Or did you mean something else?)
Holding down SHIFT with your left hand while simultaneously hitting
that tiny close-gadget with your right hand clicking the left mouse
button.

NS> Additionally, I'd also like to see CNet get a HOT-KEY and the ability to
NS> assign HOT-KEYs to any of the ports would be nice as well.
You mean from the control panel?
Like hitting 0-9 would pull up those nodes.
SHIFT 0-9 would get 10-19.

NS> The EDit ability is Tri-State (as it should be) ... Only give users the
NS> ability to edit posts/responses in area >where you don't mind them being
NS> edited.< Obviously, a political debate forum would not be a place to allow
NS> this, but a poetry forum!? Or a Sysop->User information forum!? We've
NS> already got plenty of SysOp control over the edit function (by user and by
NS> subboard) ... I don't think any of your 5 ideas will make EDit any better.
YES/NO or SUB's YES/NO really isn't too flexible when it's going to
allow the 'rewriting of history'.

>edited.< Obviously, a political debate forum would not be a place to allow
>this, but a poetry forum!? Or a Sysop->User information forum!? We've
That's just it.  Any 1 (or all 5) *WOULD* allow it to be used in more
places.
I'd love to know what % of everyone's callers have this set to
NO (or SUB/NO) just because it can't be controlled better.

NS> BA> Users can only edit their own responses:
NS> BA> 1) if no one else has yet responded after them.
NS> BA> 2) if less than x hours/days have passed since they posted.
NS> BA> 3) only during this current call.  (or maybe as late as 'next call'.)
NS> BA> 4) the header will be marked "(EDITED)".
NS> BA> 5) if no one else has read it yet.
Heck, I'd give 100% of my callers full editing ability, if Cnet
did #5.  (Who cares if some one changes responses if no one on the Earth
has yet read the old response?)

Or at least (very least) allow #4.

NS> I LIKE this idea, Bill! (Btw, would you consider starting to make new
NS> threads instead of adding to the, now monster, Random Q&A thread? It takes
I guess I should, but last time I checked, this sub had already hit the
max items.  (And I guess this 1 is next in line to get removed.  :-(

Maybe I'll start posting 3-5 things per response.

-Saturday 14-Nov-92 20:42:57
-Bill "Mr. BBS" Beogelein, 313-473-2020, 2-line HST 14.4k USR DS, 1:2410/207





