The Wave V1.xx

   Page Updated 25/07/2001    

The Wave is now at V1.0 and available from:

  [Maurice_Randalls_Site]

And is now also available in German from

  [Werner_Weicht's_Site]

Below is some interesting information of the various tweaks and new options available disscused in various Emails between myself an Maurice.


News and Tweaks for the WaveV1.0

Uploading through telnet.
I managed to figure out a way to improve on this again. While using my laptop with a null-modem cable, I figured out that Linux would lose an occasional incoming packet and then the file transfer would get out of sync since there was now an outstanding unacknowledged TCP packet and also a missing YModem packet.
Sometimes it would fix itself and sometimes it wouldn't. Half the time, the transfer would fail. This almost always occured on an even 32K point in the transfer. I could see that Linux was writing the incoming data to the floppy at that point and would miss an incoming packet. (who says these PC's can write to the floppy at the same time they are receiving data on the serial port! Baloney... it's not perfect)

I spent some time figuring out a way to detect this and to correct it and I think I've got it fixed now. I haven't had a failure on my end yet. I'm sure this will also help when packets are lost (for whatever reason) on the internet. We don't seem to have any problem with downloads because The Wave doesn't miss incoming packets like Linux does. :)

This also includes a bunch of little tweaks to other parts of the software, you probably won't notice much of the improvements. However, the TCP software is a little faster now. It's able to detect when data is waiting to send and won't acknowledge the current incoming packet. It will wait until the next packet carrying data goes out and combine the acknowledgment with that packet. This speeds up data transfer considerably since it eliminates 40 outgoing bytes for every incoming packet. You'll notice an improvement when typing online such as when using Pine on a telnet connection.

Some NEW choices in the various menus

EC in the Telnet options...
EC stands for "erase character". It's the telnet sequence of bytes that tells the remote host to delete the character to the left of the cursor. A proper telnet site uses this instead of the common decimal 8 like most BBS's would use. Some telnet sites are not done correctly, so for those, you can now set an 8 or a 127 depending on the site.


Auto box...
When The Wave starts up, it first tries to find the SL/T232. If this is set for AUTO, it will look for it. If AUTO is not selected, then it will use whatever address the user chooses. This allows the user to select something like $d700 in case two SwiftLinks are connected. If AUTO is selected, The Wave will use the first one it finds. Otherwise, the user can force it to use a particular one by turning AUTO off.

Now, for the BPS speed... The Wave tries to determine the fastest speed your modem can handle between the SL/T232 and the modem, as long as AUTO is selected. If AUTO is turned off, then it will use whatever speed you have chosen. In most cases, both AUTO settings will work correctly. In a rare case where they might not, then the user can pick the settings that work.

A null-modem user must turn AUTO off since there's no way to determine how fast the connected PC can go using the methods I'm using. In that case, the user picks the speed that the PC is capable of.

Noisey Phone lines..
I've thought of a way to alter my SL/T232 driver to help counteract a noisy phone line.
The only thing I've been able to determine is that noise on the phone line occurs and for maybe a small fraction of a second, the carrier goes away. This results in a change of carrier detect at the SL/T232 which causes an interrupt to occur. My driver sees this immediately and signals to the rest of the software that we no longer have a connection. So, the next time you click on a link the software thinks it has to redial out to the ISP and it does so.

I'm going to change it so that the change of carrier is no longer dealt with in this way, but is still checked when an actual byte comes in. Chances are, the carrier will be present when the byte comes in and the connection will remain up and running. That momentary loss of carrier will not cause a problem in between bytes.

There are still other parts of the software that check to see if a carrier is still present so we will still have some usefulness there from the signal without losing any functionality.

[Return_to_the_Wave_Index_Page]