[linux-dvb] [PATCH] Userspace tuner

Tobias Stoeber tobi at to-st.de
Sat Aug 18 01:37:02 CEST 2007


Hi,

Dâniel Fraga schrieb:
> 	Well... if I was a developer, I would knew why I'm doing this.
> And I would accept the challenge ("Man who plays golf in rain get wet
> balls"). So, developing drivers isn't just write code, but to interact
> with users and see what's their problems. We, the users, depend on you,
> the developers.
> 
> 	And Linux just make sense if it support the hardware the users
> have. Otherwise, people are unable to use Linux completely.

Well, sure it would be nice, if linux would support almost any device, 
which is out there. I am using linux for over 10 years now, and I am 
used to check in advance which hardware is supported.

Unfortunately there is so many hardware out there, that's - well - just 
crap. And even in the case it is not, you wonder why you get a fully 
packed CD or DVD to install windows drivers + additional sofware of 100+ 
MB to get it working....

For example, I have a small webcam, which I can use with v4l framework 
and a driver of 22 kB. Using Logitech'ss driver package (a 30 MB 
download) it uses around 60 MB and won't fully work, if I remove certain 
"unnecessary" parts.

I also encountered a few times the problem, that a specific hardware had 
already been replaced after 2 or 3 months by a new one ... same package 
design, same look ... but completely reworked, completely new drivers. 
Just because many manufacurers "just" by their hardware from someone 
other ... recognize it's not what they hoped / paid for ... get another 
manufacturer ... and keep the valuable marketing efforts by using the 
same name and design.

So, yes, it is desirable to support as much hardware as one can, but it 
sometimes is just impossible or wasted time, because devices may 
disappear as fast as they went to market.

>>Why would a device be obsolete?  Why would no one still be using it?
> 
> 
> 	For example, in version 2.6.17, I was amazed by the fact that
> there's support for an old and completely obsolete adlib FM card:
> 
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=cf40a310a7aaf1944eea3e01e9c120b31850c3b6
> 
> 	And I ask: who in this world will be using such an old card?
> It's worth the effort?

So you think, there may not be users out there with such adlib cards? 
Because the ISA bus is obsolete, not found anymore on modern PCs?

Well, there are so much kinds of hardware, that are no longer to be 
found, RS232 serial ports, parallel printer ports, even agp ports etc.

But believe me, it's sometimes a real pain to use usb adapters, where 
you have used a simple RS232 in the past. Or there have been and are 
great ISA measuring boards (also with regard to programming), where you 
today really waste a lot of time with problems related to pci / usb, 
need for a GUI etc.

Another reason is "hardware re-use". Just recently I built a small 
system for my home office from old PC/104 hardware (which uses an ISA 
bus btw). Just a cpu board, a network card, a 2-slot pcmcia card + modem 
/ isdn card and a notebook harddisk. The main reason for this were just: 
(a) hardware already available / no additional costs, (b) using standard 
x86 debian and (c) production system consumes ~7 - 10 W power (depending 
on hd and modem activity).

Sure, you can't what tv or play movies on such a system ;-)

> 	Now, we have the cx88 chips, successor of the bt87x chips...
> and it isn't supported completely yet. When it's finally supported,
> maybe it will be obsolete as an adlib fm card is today. So there's a
> time limit to develop a driver. If you take too long, it's better to
> dedicate on another kind of hardware...

Well, please bear in mind that linux always had and has a tradition to 
be used by people who don't belong to the group of the so-called 	"early 
adopter's" ... for various reasons (re-use of "old" hardware, lack of 
money as a student etc.). So maybe not all work is wasted.

>>Why? What's the rush?  Is there a race on? 
> 
> 	I think there's a race between Linux and Windows. If you don't
> support the hardware people are using, they will go back to Windows...
> or give up on Linux (not me of course, I use Linux for about 10 years
> and I can live with the lack of support on some hardware, but most of
> users don't). If it isn't a problem, ok, but I think that Linux will
> lose many users on the multimedia front. 

On the server part I think, Linux has found it's way and is recognized 
as an OS commercially.

Sure, on the desktop part there may be a kind of "race". Most problems 
resulted in the past from installers, that demanded to much "input" from 
the people (note: Windows comes pre-installed ;) That's been solved I 
think and I know quite a lot of (now) linux users who apreciate this 
(and who surrendered due to install / configuration problems in the past).

Windows to some extend must be loved by many hardware manufacturers. 
Take Vista for example ... almost anyone I know who has to use it for 
some reasons, had to upgrade some hardware. But for what (real) 
improvement? Take way the new fancy look and all those silly things, 
there's not much to justify the investment - just my personal opinion.

Did you ever have been so "kind" to help your relatives, friends, 
neighbours, ... to re-install windows. What a "pleasure"!). So if 
someone would build and sell pre-installed desktop linux systems on a 
large scale, they would use hardware that is supported or convince the 
manufacturer to release specs/docs or develop a driver. This already 
happened for the server use of linux (e.g. raid controllers).

People ... well "users" are some crazy folks. The "mainstream" user 
really is receptive to marketing. Users don't (always) prefer a good 
(the best) solution. So the world used VHS for home recording a very 
long time (and still, actually to only be beaten by whole new digital 
standards). In the end it was marketing and money that made the race for 
VHS then.

> 	Well, anyway I don't want to start a flamewar. I just want to
> express my opinion. I respect your opinion, but I really insist on the
> point that there's some parts of Linux where the development is too
> slow. I know you won't agree with me, but it's ok.

As a user, well, I can really agree to some point with you ;)

But comparing linux drivers (often re-engineered without docs from the 
manufacturers) to many windows drivers (which should not have much 
problems with singning NDAs etc. - and you only built devices upon 
hw/ic's you have specs for!) the quality of linux is amazing!

Cheers, Tobias



More information about the linux-dvb mailing list