home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Share Gallery 1
/
share_gal_1.zip
/
share_gal_1
/
HR
/
HR007.ZIP
/
RAMBLING
< prev
next >
Wrap
Text File
|
1986-01-12
|
5KB
|
102 lines
WA7MBL BSS - POSSIBLE ENHANCEMENTS & MODIFICATIONS - 11 Jan 85
Distribution: WA7MBL BBS Beta Testers, Users
Just about everyone!
First, a few assorted rambling comments......
There exist what I consider to be major problems in the current BBS structure.
The majority of these are not due to poor design by W0RLI (in fact, I think
Hank has done an excellent job of finding ways around as many of the potential
problems as he has!)
The main problem is that the TNCs are NOT designed for use with host computers.
(WA8DED has made a major contribution in this area, BUT his software runs only
on the TNC-1 and lookalikes.) The biggest problem is that there is no easy
way to tell if all of the packets sent to the TNC have been received and
acknowledged by the user. There is also no convenient way to tell what mode
the TNC is currently in.
Currently when a message is selected to be read by the addressee, it is marked
as having been read at the time of selection. Ideally, it would not be flaged
as read until the last line had been sent and acknowledged. A possible (but
very kludgy) method would be to not mark it until the next command had been
received from the user (even this wont work if the user decides to save time
and send the next command early!) A sysop wants to be able to clean messages
that have been read off the board, but there is presently no guarantee that
it was fully received. Everyone should at least be aware that even if the
flag is set as read, it aint necessarily so!
Also, there is a problem if the user is not using a C/R as the 'SENDPAC'
character. If there is a partial line of text, then a *** DISCONNECT appears
without a leading C/R, it doesnt get detected by this software. Parsing for
a *** DISCONNECTED anywhere would mean that you cant upload this message to
a BBS. While it would be possible to use pin 8 of the TNC for connect status,
this brings up other problems. The real solution again is a better setup on
the TNC side. Maybe when the proposed PC plug-in TNC board appears, we can
solve these plus other problems. In the mean time, I thought it wouldnt hurt
to make everyone aware of some of the existing problems.
I also have concerns with the method of breaking out of the Gateway. If a
person connects to me on 2m, uses my 20m Gateway to another BBS then uses
that Gateway to connect to someone else but makes a mistake he has to either
retry out or send a cont-W which takes down the entire chain. If he had to
wait because of a busy station, it can be a frustrating experience. Dunno
if its anything to worry about, but we should consider it. If binary file
transfer is ever going to occur across a Gateway, there will be another
potential problem of how to pass everything transparent, yet be able to
control the Gateway (kinda hard to send a Break via Packet to get back to
control mode!) A binary protocol which converts control characters (like
KERMIT in 7 bit mode) could be used, but seems like a waste of extra chars
and acks. Just something to think about.
..............
The following modifications and enhancements have been proposed by various
parties, also I've thrown in a few of my own ideas here and there. I'd like
to hear comments from anyone on these proposals (as well as any additional
ideas). Your comments could also include a priority list of what you see as
the most important to get done first.
1 - Be able to abort during a Forward.
2 - Reject all mail not addressed to a local user, a user we forward to in
FWD.TNC, or an @BBS in FWD.TNC ??
3 - Reject all mail not addressed to the owner. [In both cases, a reject
would have to be done by disconnecting, after sending an appropriate
message, otherwise a station doing auto-forwarding may think the message
was received ok.]
4 - Output a character to the parallel port both as a command and from the
FWD.TNC file. [I'm working on this one!]
5 - When sending the same message to a number of people, allow a special
command in the message text to indicate that text is to be included
from a file.
6 - Be able to request a file during forwarding.
7 - Add a KM command, Kill my messages that have been read. [The problem
with this is that messages are marked as 'read' as soon as someone has
selected it to be read, as per earlier rambling]
I've could make a bigger list, but this will get things started. Multi-User,
binary transfer, DOS commands, lotsa possibilities. I'll read thru my notes,
letters, electronic mail, etc., and try to put together a master list of
everything that has been suggested. I know we will never all agree, but at
least we can discuss it all.
I'm still digesting all the comments about security (user level) and file area
structure and commands for best use on a hard disk.
Vote for (or against) your favorites.
73 - Jeff