Robert Lipe wrote:
> On Tue, Mar 31, 2009 at 1:41 PM, Paul Cornett
> wrote:
>
> > Hi Robert,
> > I saw a message from you on the DeLorme PN-40 forum, and wanted to
> > let you know I have been working on support for the DeLBin USB
> > format. On
I've opened an SF patch for this, and uploaded what I've currently got
for you to look at.
I've found device transfers to be highly unreliable. Reading a track
with hundreds of points will frequently fail, due to what seems to be a
lost packet. I've read that people have had trouble with the cable,
maybe thats the problem. Anyway, the code is alot more complex than I
would like, due to allowing for retrying almost everything.
I'm confused about the interpretation of the waypoint shortname,
description, and notes. It appears that some modules (like garmin)
ignore shortname when writing, and create a name on the fly from
description. What is "best practice" for dealing with names,
descriptions, notes, comments, geocache information, etc?
> If we produce a driverless kext to keep the OS/X HID driver out of
> this device, we can use libusb, just like you did on Linux.
>
> Probably we'll want to use the "real" Mac interface at some point,
> but that has a cost. The per OS-piece isn't tricky as long as we
> build in a good dividing line between protocol bits and hardware/OS
> bits.
Right now I've just assumed that a simple init/deinit and packet
read/write interface will be enough.
> > I haven't looked at what you checked in on the delbin-branch, but I
> > noticed some C++ files. Is it acceptable to use C++ for a GPSBabel
> > module? I would certainly prefer that.
> >
>
> This is one of two outstanding projects I've goofed on in C++. The
> other *requires* C++.
>
> When I started GPSBabel in '01, the state of C++ was too sad to be
> usable. While I see no payoff in replacing existing GPSBabel modules
> with C++ versions, I'm warming up to the idea of using C++ in
> GPSBabel.
Well, that was not exactly a "yes", so I've stuck with C.
|