Re: ADTPro beta - protocol change

[Tristan Mumford <xtristan.xmumford@xgmail.xcom>]

schmidtd wrote:

> While doing destructive testing on ADTPro 0.1.5, I noticed that
> dropping an acknowledgment packet (entirely possible via UDP or Audio
> transports) can result in corrupt image transfers.  The problem is
> that the host, sensing a timeout, resends the packet.  If the Apple
> had already acknowledged it but that ack got lost, the resent packet
> was written as the next one in line.  Definitely bad.  I added block
> numbers to the protocol so now we know _which_ blocks are getting
> acknowledged or sent.

I welcome the change with open arms.
I love the ADTs because they are one of the most awesome things ever, but I
have found that data integrity has been an issue.
Recently I transferred maybe 50 doublesided 5.25" floppies to my PC. Every
single one was unusable but I was given no warning of this.

Can I make another suggestion?
The client and host end exchanging the number of blocks before transfer.
I've dumped an image to a IIgs RAMdisk a few times, and have to manually
kill the client end of the transfer once it's done. Not amazingly
inconvenient but it's a possible extra feature. It would also provide
another basic form of integrity check.
Ie if expected number of blocks != total number received, something is
wrong.

Tristan.
-- 
-----> http://members.dodo.com.au/~izabellion1/tristan/index.html <-----
=====       It's not pretty, it's not great, but it is mine.       =====