Date : Thu, 18 Nov 2010 15:15:44 +0100
From : rick@... (Rick Murray)
Subject: [OT] Who'da thought...
On 18/11/2010 14:48, Jules Richardson wrote:
> to give up just for instant Asians and naked kittens, or however you put it.
<giggle>
> I really, really hate over-complicated sh*t that doesn't give me any real
> benefit. I'm not a luddite because I'm all for technology - but technology
> that's simple, well-documented, and solving well-defined real-world problems
In today's IP-obsessed world, the level of documentation is going down
at an alarming rate.
Example. My Neuros OSD runs on a TMS320DM320. It is a sort of pre-OMAP.
ARM core, DSP, bunch-of-stuff. Now, the "OSD" in OSD actually stands for
"Open Source Design". The aim of the OSD is to be completely open, and
on Neuros' part, it was. I have the complete source tarball plus schematics.
It's everybody ELSE that let the project down. I can understand Indigent
(sp?) not being open as DSP codecs are their income, we get the
binaries. Not perfect, but understandable.
On the other hand. Texas Instruments. Release the tech docs under NDA.
For what purpose, exactly? It's a pretty standard datasheet saying what
the chip does, registers, timing, blah de blah. So I contact TI given
this part is "so old" it is no longer on their website. I get a reply
basically saying "I'm neither a corporate nor educational user, get
lost" (paraphrased!), written by what must have been a non-native
English speaker as the message said the person had pleasure in declining
my request. Nooooo, you don't make people happy with faux pas like that.
So a bit of epic abuse of Google Translate plus a lot of head-scratching
to figure out slang written in Chinese (!), I finally found the three
parts of the DM320 datasheet. Which says all over it to be released
under NDA. I signed NOTHING with TI (I would have, once) so instead I
put it up on my site in case there is anybody else interested in
developing for the OSD. I count three (G, pffft, and myself). It is a
quite capable PVR. My old (crappier but much faster) PVR runs the same
chip so it shows what sort of response times ought to be possible. But...
The most Neuros could ever hope for is people screwing around with
trivial bug-fixes in the firmware (a hacked around ARM Debian), for the
DSP is completely undocumented (so no scope for additional codec
support) and until recently :-) so was the chip. No hacker with an ounce
of pride is going to stand for "documentation? read the source!".
It was (still is, if you ask me) a great little concept. It is,
primarily, a PVR. But with SD/CF/USB and a 10/100 network socket, it
could easily be repurposed into a lightweight web server. It isn't,
really, so different to a Beagleboard, except for a slower CPU and a lot
of video-related stuff, plus PAL/NTSC output (not (S)VGA). But in terms
of toy-value, I am seriously interested in the idea of attempting a
(basic) RISC OS port.
But, then... until I dug in up in the country of the red flag, WTF was
the documentation? NDA? Why? Don't they *want* developers? Or are they
scared of something deeper?
Sorry, it is a long rant - but your "well-documented" kinda hit home.
> without a billion bells and whistles to get in the way.
I believe this is the modus operandi of Apple Corp.
Best wishes,
Rick.
--
Rick Murray, eeePC901 & ADSL WiFI'd into it, all ETLAs!
BBC B: DNFS, 2 x 5.25" floppies, EPROM prog, Acorn TTX
E01S FileStore, A3000/A5000/RiscPC/various PCs/blahblah...