In a message to Anthony about ram disk, Rod said:

AM> Do DV apps usually become _very_ sluggish when run under DV?

RS> No, never ever. Which suggests that Winfred isnt at fault since it
RS> works fine with real DV. You cant do it wrong in the app anyway,
RS> you just say 'ok DV, I've finished with the rest of this
RS> timeslice, you can give the rest away to another task, I have
RS> nothing useful to do'

AM> Surely he'd know at what places in the program it's logical to
AM> make calls like that?

RS> And he must be doing it right since it works fine with real DV.

Well, you know that, and I know that, but I had to ask.  Stranger
things have happened.

AM> It doesn't interfere with downloads,

RS> Thats essentially because they are interrupt driven.
[...]

Yes, that's true.  In fact, I have telemate set up to communicate in
fossil mode, using SIO's VX00.sys direct interface to SIO.SYS which
would make this effect even more pronounced, as telemate would only be
executing to work on large blocks of data, rather than on an interrupt
every character or every 8 characters basis in the case of 16550...

AM> but mainly while dialling, and the interpretation of result codes
AM> like BUSY or connects.  Often the modem will detect busy, hang up,
AM> send the code, and Telemate will still just sit there until it
AM> finally notices the busy characters (often ten to thirty seconds!)
AM> or worse, until it times out after the 60 seconds I set, not
AM> seeing them at all...

RS> They are all just extra manifestations of how fundamentally rather
RS> pathetic DOS high speed comms is under OS/2. Some manifestations
RS> are more grotty than others. In some circumstances you cant even
RS> get the modem to display its stored profiles properly. Pathetic.

The more I look at it, the sluggishness is only evident when telemate
has to exchange control info with the modem, like during dialling or
waiting for result codes...  I'll do some more checking to narrow the
problem down at bit, like going back to native comms, fiddling with
OS2SPEED...

'tnt, Anthony.

 * KWQ/2 1.2 *