Letters to the editor

Letter to the Editor

To the Editor,
 
I have started a grass-roots campaign aimed at convincing IBM to open up and
free the programming documentation for its STB (Set Top Box) series chips
http://www-306.ibm.com/chips/products/digitalvideo/products/settopbox.html .
You can find the campaign home page at
www.users.bigpond.net.au/mysite/freestb.htm
 
My personal outlook is that the documentation and drivers need to be freely
downloadable to encourage people to develop. Developers have to be pretty
motivated to embark on a quest to engage with an IBM reseller to sign an NDA
and eventually get access to the documentation. IBM, Microsoft, Sun and all
other software companies are very open with their software documentation and
programming documentation for software API's, but why not hardware?
 
I'm not sure what the origin is of the practice of hiding chip documentation
behind an NDA. Surely companies like IBM don't think that an NDA would in
some way keep them one step ahead of the other chip manufacturers?
 
I think such information hiding is a relic of the days of proprietary
computing and chips haven;t felt the wind of open source, so to speak.
 
This is a sub $100 TV connected Linux machine which uses an STBx25xx chip
http://www.hauppauge.com/html/mediamvp_datasheet.htm
Here are the hardware specs of the Mediamvp machine.
http://www.shspvr.com/forum/viewtopic.php?p=19411
 
Also http://www.netgem.com/ has an STB chip.
 
This forum shows people who are trying to build their own version of linux
to run on the MediaMVP http://www.shspvr.com/forum/viewforum.php?f=38 I
think life would be much easier for these people with full documentation and
drivers available.
 
Projects like this would benefit from public and free access to the
documentation and drivers.
http://sourceforge.net/projects/mvpmc/
 
I'm hoping to spread the word and build public pressure on IBM to address
the issue. Any help that you might be able to give would be appreciated.
 
Regards
 
Andrew Stuart
astuart@bigpond.net.au

Comments (2 posted)

Release Early, Release Often

Dear editors,
 
The one thing most software developers forget is that
if the project is dead to the world, it will often
become simply dead.
 
At the moment, I'm looking at software routers - an
area notorious for slow to non-existant releases.
Zebra, for example, has a commercial offshoot, and the
Open Source version has since come to a halt. Who,
though, is going to buy the commercial product, if
they perceive the project as dead?
 
Click, a router from MIT, is better in that they have
just made a release. A very large, bulky, and no doubt
bug-ridden release. That's the reason for releasing
often - bugs breed in the dark, and die off in the
light.
 
Mind you, if you are after a software router, Click is
the only one out there with any decent releases at
all. It's also very fast and does support a lot more
than the others ever did. All the others are sleeping,
comatose or dead.
 
If you want to encourage Open Source - and I think we
all do - then release early and release often. LWN
does, with its headline news items, though you could
argue it's not really software. The point is, though,
the model works and the alternative doesn't.
 
Jonathan Day

Comments (4 posted)

XNotesplus - calendar support (and comments on Grumpy Editor column)

 
Anyway, there were some comments made about XNotesPlus that I need to
add some clarification to.
 
1. XNotesPlus is shareware. There have been 1000's of downloads of it.
To date, I think I've had 5 people pay for it. Anyone complaining about
it not being free ought to see it from my end. It'd be nice if people
paid for it, but I'm not tracking anyone down who doesn't. I just ask
that they do. I just didn't want it being added to GNOME or any
distribution without my getting something for it. I designed it
specifically for end users, not for distributions. It doesn't even
install in system directories, it installs in the users HOME directory
(something I may change in the future, however, so it can be used on
university systems).
 
2. The post from "utidjian" was just plain silly. XNotesPlus builds on
just about any Unix platform if you have the proper prerequisites (and
what desktop application doesn't have prerequisites?). The "pspell"
issue is configurable in the config.h (it doesn't use autoconf cuz I
just haven't gotten around to learning that yet - I still use imake) and
is not required but is turned on by default. The RH7.3 binaries are
available for purchase on CD but then so is the source and so are - now
- the RH 9 binaries. I use RH 9 at home and have built XNotesPlus on
probably 20 different platforms. And the RH9 binary version is even on
the web site, available for download *WITHOUT PAYMENT*. So this guy is
just whining without really doing his homework (and I certainly have not
received any email from him asking about these issues).
 
The calendar feature of XNotesPlus was just added to the 3.6 release and
may be buggy. I found a crash today that I need to fix. But I use the
calendar at work and at home and, from the KISS perspective, it works
fine. Its not full featured - it doesn't do group scheduling or the
like. But its not meant to. Its meant to be the home users calendar.
Eventually I may add groupwise support (its on the grand plan) but I'm
not anywhere near adding it yet. It also doesn't allow breaking an
appointment (ie the "exceptions" that the PalmOS allows) yet. I'm
considering how to add that feature. The hooks are there, but the UI is
not implemented.
 
Anyway, I wish people would spend a little time sending *me* the
feedback on problems with XNotesPlus instead of whining in public. I
have answered *every* email with request for help in getting it running
and, to my knowledge, have only failed to get it working once (with
someone who was new to programmning). Just cuz its shareware doesn't
mean the author isn't willing to help. I'm just asking you to offer an
honest dollar for an honest days work.
 
Feel free to post this as a reply if you want, or anywhere else. Like I
said, I tried to reply but couldn't for some reason.
 
--
Michael J. Hammel The Graphics Muse
mjhammel@graphics-muse.org http://www.graphics-muse.com
------------------------------------------------------------------------------
Bumper Sticker: Some people just don't know how to drive... I call these people
"Everybody But Me."

Comments (3 posted)

It's a step in the right direction, but only a step

 
    http://computerworld.co.nz/news.nsf/NL/6671974C513F31E8CC256E5B00723C21
 
> And in what might be a first for a senior Microsoft executive, [Peter]
> acknowledged that Linux is not going to be a passing fad.
 
> "Linux is going to be part of the future. It's going to be like Unix was."
 
While I appreciate the message in there that Linux is going to rule the server
landscape, and am frankly flabbergasted that a Microsoft exec would openly
confess as much (bonus points for so doing, Peter), I don't appreciate the
innuendo that Unix is in some way a "has been" or that Linux is going to join
it in has-been land.
 
I can plug a Linux CD into one machine, and a minute later have a fully
functional Linux workstation and server going there, with extensive office,
networking and diagnostic capabilities - all without disturbing what's on the
hard disk. I do so regularly while repairing virus-savaged MS-Windows LANs.
The staff can be editing up documents and getting on with their lives while
I'm still repairing their system.
 
I can issue one command and reboot the rest of the office into the same
software within a very few minutes, without any extra CDs (hurrah for PXE and
caching). I can batch-process information supercomputer-style on this
network. I can permanently install the software onto the machines' hard disks
while they're running and being used for day-to-day work. This is not the
substance of a has-been, and I CAN'T DO ANY OF IT without a great deal of
effort in MS Windows, and a great deal of licence-counting.
 
> For each of Red Hat, Mandrake and Debian, their websites reported more
> than double the number of security advisories of Windows 2000 and XP,
> Moore said, and while the Linux security advisory rate was rising, that for
> Windows was falling.
  
I can speak to this with authority on Mandrake. First, account for the
*nature* of the patches. Very few of them are for show-stopper issues. Think
CodeRed. If what Peter infers from this were true, there should be twice as
many attacks through Apache as through MS IIS, but day after day my Apache
web logs show stuff like this MS IIS probe and no Apache probes:
 
    GET /scripts/..%c1%9c../winnt/system32/cmd.exe?/c+dir
 
Next, account for what's being patched. Mandrake 10.0 ships with over 1800
packages including three different equivalents to MS Office, three different
equivalents to MS Exchange, two different equivalents to MS SQL Server, three
different equivalents to MS Outlook, three different equivalents to MS
Internet Explorer and so on ad infinitum. One would expect to see roughly
three times as many updates based on this factor (more choice) alone.
 
Microsoft supports an organisation trading as "Software Choice". I hope you're
not going to turn around and claim that more choice is now a Bad Thing. The
workstation I'm typing on has 1458 packages installed; some for me, some for
my wife, and some for my children.
 
Even allowing for the observation that those packages are generally more
granular (call it the equivalent of roughly 500 MS Windows software
packages), just getting all of that software installed together on MS Windows
at one time without having it "tread on each other's toes" would be a minor
miracle. When even such basic issues haven't been completely solved, security
must by definition take a back seat to not rocking the boat.
 
>"Security is an industry issue," Moore said, "and we're getting better."
  
Security is a multifaceted thing, and blaming it principally on "the industry"
denies that you're (Microsoft) putting sufficient weight on more important
issues such as basic software architecture.
 
Fixing security aspects such as this would require Microsoft to bite the
bullet and make statements along the lines of "OK, so the MIME handling in
Windows is broken, and that Outlook application is a house of cards from keel
to crowsnest. We're going to re-engineer those, *without* building in more
DRM hooks and other junk and lock-ins designed to help us and our market
image at the expense of customer utility."
 
The people best placed to help you face that are your MVPs, who are as close
to a genuine Open Source community as Microsoft (so far) gets.
 
If Microsoft doesn't do something radical along those lines, and very soon,
Linus Torvalds' flippant quip, "Really, I'm not out to destroy Microsoft.
That will just be a completely unintentional side effect." will come to pass.
Really. And then what of the customers stranded by Microsoft lock-ins, but
without any source of security updates?
 
Meanwhile, there is no such single point of failure in the Open Source world.
 
Cheers; Leon

Comments (none posted)