SF for Windows development (fwd)

Eugen Leitl eugen@leitl.org
Sun, 19 Jan 2003 00:03:01 +0100 (CET)


---------- Forwarded message ----------
Date: Sat, 18 Jan 2003 23:45:16 +0100
From: John Walker <kelvin@fourmilab.ch>
To: M.W. <debug@dk2net.dk>, speak-freely@fourmilab.ch
Subject: Re: SF for Windows development

M.W. wrote:

> After being dormant for quite some time this SF list again is active - 
> for Unix/Linux etc.
> 
> However, there have been many plans for a radical renewal of the Windows 
> program interface and documentation. These plans may be hibernating or 
> they may have been discontinued. My question : Is anybody continuing the 
> development of the Windows program interface and documentation?


At this time there is nobody working on any renewal, radical or
otherwise, of Speak Freely for Windows.  When I find the time,
which may not be soon, I plan to build a Speak Freely for Windows
which includes the CELP CODEC introduced in Speak Freely 7.5
for Unix (4800 baud data rate with GSM-like audio fidelity) and
the extended robust/redundant transmission options included in
that release.

Basically, I consider Speak Freely done, which doesn't mean
finished.  It was intended as an OpenSource (although that word
hadn't been coined when it was born) network telephone incorporating
the best performing audio compression and encryption algorithms,
explicitly developed to serve as a "code mine" for folks interested
in deploying them in other applications.

Since about 1998, I consider this goal achieved.  Largely due to the
improvement of Internet bandwidth due to "dark fibre", most of the
problems people encountered in IP telephony in the early days are
now historical footnotes.  One of the reasons I haven't made CELP in
the Windows version a priority is "who cares": GSM compression is
more than adequate and works over most Internet connections about as
well as a typical mobile phone.

Many things could be done in opening Speak Freely as an Internet
telephony server, but the means of accomplishing that seem to change
more rapidly than one can write code, and it's far from clear that
there's a large number (or, in fact, any) people waiting to exploit
such an interface, whether deployed via OLE, COM, .NET, or alien
implants.  In fact, the existing Speak Freely source code, ugly as it
is, has been used as the basis for a number of projects, most of
which wouldn't have benefited from a rigid API interface.  I do not
argue in favour of ugly code; in fact, some of my more recent scribblings
have been cited as examples of literate programming:
     http://www.literateprogramming.com/fexamples.html
but one must balance effort against potential benefit, and in a mature program
such as Speak Freely, the burden of proof rests upon demonstrable benefit.

Should somebody go and develop an API interface for Speak Freely for Windows
(or a shared library for Unix), I'll be glad to integrate
it into the distribution and make it available.  But history shows that
such projects work best when coupled with an effort which *uses* the
API being developed; absent that, it's almost impossible to identify
the lacunŠ which are obvious in retrospect.


                      * * *

To unsubscribe from this mailing list, send E-mail containing
the word "unsubscribe" in the message body (*not* as the
Subject) to speak-freely-request@fourmilab.ch