[Lcdproc] VLSYS L.I.S 2005 Driver written

[Daryl F]

Forgot to CC the list. Here's my reply to Peter.
-Daryl

-------- Original Message --------
Subject: Re: [Lcdproc] VLSYS L.I.S 2005 Driver written
Date: Sun, 01 Apr 2007 10:51:43 -0500
From: Daryl F <daryl@rainyday.ca>
To: Peter Marschall <peter@adpm.de>
References: <46057666.1040406@rainyday.ca>
<200704011358.42525.peter@adpm.de>

Peter Marschall wrote:
> Hi Daryl,
> 
> On Saturday, 24. March 2007 20:05, Daryl F wrote:
>>   I wrote a driver for the L.I.S 2005 (not the L.I.S 2) by tracing the
>> USB traffic on a WinXP system with VLSystem's (vlsys.co.kr) driver and
>> application running. I've inquired with them if I can release the driver
>> to open source but have received no response.
>>   Can I release this driver for lcdproc without their permission? Does
>> it break copyright law to trace a datastream and write a driver without
>> decompiling or disassembling their programs?
> 
> First of all: I am not a lawyer, so this is no legal advice.
> 
> If you only looked at the data stream transferred over USB,
> found what it does and then wrote an LCDproc driver without
> looking (disassembling, ....) at the original drivers,
> then it looks pretty much like what the Samba team is doing
> with the SMB protocol.
> 
> Maybe they can help you with your question.
> 
> In any case it does not matter whether you release this driver
> "for lcdproc" or "for" any other purposes: you are the one
> who may get under fire first should they consider the driver
> illegally created.
> When the driver gets included in LCdproc and they take legal action,
> LCDproc (and the ones who added the driver to LCDproc - me) may
> also be affected.
> 
> So, I'd rather stay on safe ground and will currently not include 
> the driver.
> 
> Could you ask VLSystem a 2nd time?
> Maybe they react this time.
> 
> For curiosity's sake:
> How did you snoop the data off the USB wire?
> 
> Regards
> Peter
> 
> 

I certainly can't afford any legal defense so I'll hold out for
permission before I release any code.

I'll ask VLSystem again. They haven't said "no" yet so maybe I'll ask
someone else there. Hopefully other people will start asking too.

The snoop I used isn't exactly off the wire (ie. hardware). There is a
software tool at http://benoit.papillault.free.fr/usbsnoop/ I used on a
WinXP laptop to trace the protocol. Installed the software from
VLSystem, turned on the sniffer and plugged the display into a USB port.

I then used usbmon in the Linux kernel to trace what was going on when I
tried the lis2 connectiontype for the HD44780 driver. It was easy to see
that the lis2 uses different baudrates and terminal attributes.

I ended up using libftdi ala the ULA200 driver to have better control
over the device. The usbserial stuff in the Linux kernel would require
changes to support an ioctl() for the brightness. That seemed really
tacky, going way outside the 'serial port' pardigm.

With usbmon tracing on linux and the copy of the log from WinXP it was
easy to feed the device what it needed.

Thanks to you and all who've contributed to lcdproc.

-Daryl