Mail To: Michael Opel <i381@stio1.fh-wuerzburg.de>
Subject: Re: Stik/Sting
----------------------------------------------------------------------------

On Mon, 25 May 1998, Michael Opel wrote:
>Hi,
>
>my name is Michael Opel from the Running Design Team, the
>guys who developed the game Running for the Atari Falcon.

Hi Michael,


>I got in contact with you via Anthony Green, the australian 
>distributor of Running.
>
>Well, I think you know that already. Also that we plan to
>use the Stik API.

Yes, he mentioned your discussions of this in the STiK beta
tester's mailing list, asking for our response to this.


>But first of all I need some general informations about that
>API. You mentioned the STING.HYP - where can I get it ?

STING.HYP is the ST-Guide formatted hypertext document for the
STinG TCP/IP stack.  It can be downloaded from the STinG home
page "http://www.stud.uni-hannover.de/~perot/", separately or
as part of the main STinG release package.

On those pages you will also find many other useful tools,
and links to some other STiK/STinG-related sites.

STinG was developed by another member of the STiK beta-testing
crew, Peter Rottengatter, as a logical extension of STiK into
a fullfledged networking system.

Where STiK is simply a one-port TCP/IP stack, allowing dialup
connection to use client programs through ISP servers, STinG
is a real networking stack allowing both local and remote use
of both clients and servers.  And yet, from the point of view
of the client programs there is little difference, because the
STiK API is used as the basis for that of STinG, and most of
the 'normal' clients do not need to use extended functions.
So these clients may be used with either STiK or STinG active.

Servers, on the other hand, often do need more functionality
than the simple STiK interface allows, so some of them need
to be  used with STinG.  In many cases their use with STiK
would be meaningless, since STiK only allows a single port
for ISP connection, but no real local networking.


>After that I hope being able to assess the usage of that API. 

I hope so, but feel free to ask any questions needed to settle
any doubts that may come up.


>You said MIDI is a subnet - this would offer the possibility
>to use the STIK API for all communciations only, I agree.

Yes, Midi is a subnet of STinG, since STinG can use all the
serial ports supported by HSMODEM, as well as any additional
ports for which someone has written a driver module.  As yet
there are not many such, but I have myself written one such
module that uses the centronics port.  This allows fairly
fast transfers even on an ST/STE.

The point is that if you use the API for all access, then you do
not even need to know what port is used.  The user can configure
that with his normal STinG tools, without your drivers in the
game program even knowing that it happens.  So it does not matter
to your program whether the connection is a local one, using Midi
or PLIP, or even a transcontinental Internet connection via ISP
modem dialup.  Your API access needs for all cases are identical.


>This is my personal opinion. I brought up the idea of STIK support 
>in our team recently - hmmm - to make a decision we need a little 
>bit more information ...

Yes of course.  Much of what you need to know will be in the HYP
file mentioned above, and even more will become clear if you install
and use the STinG package for a while.  Many of the client authors,
as well as the authors of STiK/STinG and their modules, are also
available in the mailing list if you need to ask how something of
particular interest was done.


>After getting started with the whole story, I will register myself
>in the mailing list ...

Ok, I look forward to seeing you there eventually.
