[BRLTTY] Re: Information about status cells through BrlAPI?

[Milan Zamazal]

>>>>> "SH" == Sébastien Hinderer <Sebastien.Hinderer at ens-lyon.org> writes:

    >> It is not.  (Only XEmacs can call binary code from Emacs Lisp
    >> some way.)

    SH> Hmm, too bad.

Not really.  While using foreign interfaces may save some work
initially, it's better to use native implementations in the long term.
Using foreign interfaces makes a risk of crashing the application,
problems with debugging, necessity to care about garbage collection and
potential binary incompatibilities (and in case it's necessary to write
C wrappers instead of direct calls to a shared library, e.g. in Python,
additionally the need to make software distribution architecture
dependent).  This is why I like BrlTTY's socket API.

    SH> As Samuel and I already explained, we don't think converting the
    SH> protocol to text is a good idea.

If all the problem is the raw mode (IIRC), then it's not a problem at
all, it's easy to arrange binary data transmission within a text
protocol (see HTTP for an example).

    SH> However, on intermediate solution could be to write a little
    SH> gateway which would take text as its entry and communicate with
    SH> BrlAPI.

The question is whether it's worth the effort (now), I've no idea how
many clients other than speechd-el access BrlTTY directly through the
socket API.  While I'd welcome if I could communicate with BrlTTY using
a text protocol because it saves implementation and debugging work on
the client side, it's not as important to me as some new features (such
as access to status cells).

Regards,

Milan Zamazal