[hylafax-devel] Re: Class2/2.0 vs Class1

[Dr. Harald Pollack <Harald.Pollack@Datanews.at>]

On 20 May 2000 10:48:57 +0400, Dmitry Bely wrote:

Hello Dmitry!

>> Timing consideration are of less importance, because the are in the magnitude of several 
>> seconds (in contrary to class 2/2.0, where often some milliseconds decide for death or 
>> alive :-)
>
>Very interesting. Can you point to a part of the protocol where Class2/2.0
>behaves worse than Class1?

It is NOT the behaviour of the 'class protocol', it is the behaviour of the (internal) firmware 
in faxmodem and the assumptions, that the developer of firmware has already taken and which are 
therefore constant and unchangeable!

It is a very hard job, together with the real firmware developer, make class 2/2.0 protocol as 
transparent as possible (from the point of class 2/2.0 fax software). I know exactly the 
problems, because I have worked (and still do) together with developers of firmware. Normally, 
they do their job, test functionality with one or two Win?? fax software and if these passes, 
they assume, their job had finished (which may be correct for some Win?? faxprogs :-)

In real worlds faxmodems, you DO NOT have real class 2/2.0 protocol! All you have, is a kind of 
protocol, which works (often fine) together with Win?? faxsoftware protocols!!!

But in more detail:

In class 2/2.0 protocol, there is one point of interest where T.30 protocol could (almost) never 
be translated to class 2/2.0. It are the two commands AT+FDT (sending) and AT+FDR (receiving). 
These are a kind of 'multipurpose' commands, and nobody knows very well, which purpose the follow 
in a certain situation!

I will describe this on the example of AT+FDT:

"The FDT command requests the DCE to transmit a Phase C page. It is issued at the beginning of 
each page, either in Phase B or in Phase C."

This (from class 2.0 standard) is very unclear, because nobody (except the developer of firmware) 
real knows in normal class 2/2.0 faxprog, which state (phase B or C) is active at a certain 
moment.

In fact, the AT+FDT command does a kind of 'frame release' (TSI+DCS or MPS,EOM,EOP) in 
consequence of T.30 but WHICH ONE is slightly unknown. And, additional, faxmodem does the 
repetition of the commands, and you DO NOT know, how often it has already repeated! Only in case 
of phase B after phase A (or after RTP or EOM) you can see DIS+CSI from called station as 
response to repetitions of TSI+DCS. In case of phase B after phase C, you never see the MPS-MCF 
sequence and how often it was repeated or not! All you can do is, to believe (or not) the +FHS 
response after success or failure :-)

Normal implementations of class 2/2.0 do all the neccessary repetitions automatically if no +FDT 
is supplied in time. But A LOT OF faxmodems behave mysterious, depending AT WHICH MOMENT fax 
program send AT+FDT to faxmodem!

But I am not able, to describe all the possibilities here, becausemy english is too less. All I 
can do is, to report my experiences. And those are: If treating class 2/2.0 exactly like class 1 
(watch carefully the state of T.30 protocol) and send commands only in those moments where they 
should be in class 1, than class 2/2.0 works perfect. Maybe that's why I had implemented class 1 
first and than class 2.0 and long after this I have done some work on class 2.


Mit herzlichen Gruessen / Yours sincerely

Dr. Harald Pollack

Harald.Pollack@DATAnews.at



____________________ HylaFAX(tm) Developers Mailing List ____________________
 To unsub: mail -s unsubscribe hylafax-devel-request@hylafax.org < /dev/null