Not logged in
Log in now
Create an account
Subscribe to LWN
LWN.net Weekly Edition for May 23, 2013
An "enum" for Python 3
An unexpected perf feature
LWN.net Weekly Edition for May 16, 2013
A look at the PyPy 2.0 release
Posted Jan 24, 2013 1:40 UTC (Thu) by tialaramex (subscriber, #21167)
But in the other groups I'm left with a lot of questions - why for example network chipsets but not storage controllers? Why webcams (hardly the most trivial devices) ended up UVC compliant almost down to the very cheapest products on the market while DVB remains an endless series of incompatible chipsets doing much the same things but all in very slightly incompatible ways?
And why the timing? Was AHCI somehow perfectly timed for an opportunity to standardise storage controllers? Is there some clear reason such a standard couldn't have emerged a year earlier? A decade earlier?
USB mass storage shows that you can crank out a standards document even if the contributors can't agree on anything of consequence. Six different ways to do either of two different things, but only six, and not sixty or six thousand. And it also shows that having done that the market explodes.
Posted Jan 24, 2013 9:43 UTC (Thu) by robert_s (subscriber, #42402)
This may be a rhetorical question, but I'm answering it anyway. Microsoft decreed that webcams had to implement UVC to get its' "designed for windows 7" badge.
Posted Jan 24, 2013 12:12 UTC (Thu) by nye (guest, #51576)
Does that mean the the only webcams you can confidently assume to be Linux compatible are those with 'Designed for Windows 7' written on them?
That amuses me.
Posted Jan 24, 2013 21:31 UTC (Thu) by rahulsundaram (subscriber, #21946)
Posted Jan 25, 2013 5:01 UTC (Fri) by rsidd (subscriber, #2582)
Posted Jan 28, 2013 12:25 UTC (Mon) by tialaramex (subscriber, #21167)
In the past I have given the same advice to people that rahulsundaram mentions below, I'm aware of how this came about and there are plenty of examples, Microsoft also effectively mandated AC97 and HDA, and I think they had some influence on the AHCI situation. But again, why these things and not everything else?
Did somebody at Microsoft consciously say "Oh, we need to standardise webcams, but DVB doesn't matter" ? I'm guessing no. So how did we get here, and what can we do to improve things?
Posted Jan 28, 2013 16:17 UTC (Mon) by mjg59 (subscriber, #23239)
Posted Jan 29, 2013 16:08 UTC (Tue) by tialaramex (subscriber, #21167)
The driver for something like a Realtek Gigabit Ethernet chip slapped onto a mid-range home PC is nothing very clever and gives no signs that the hardware is clever either (not to say I could build one, I suck at digital electronics and am a danger to myself and others with a soldering iron).
Having a class driver for NICs (even just for wired NICs) would make a lot of people's lives easier, probably even Microsoft's but it hasn't happened even as NICs went from an obscure add-on card you could buy from a specialist to a standard piece of equipment that draws attention for being absent from devices like the MacBook Air.
Posted Jan 29, 2013 16:14 UTC (Tue) by mjg59 (subscriber, #23239)
Posted Jan 30, 2013 13:57 UTC (Wed) by tialaramex (subscriber, #21167)
Most of the utterly terrible decisions seem to have already been tried (explicit polling even when idle, PIO data transfers, making the driver do lots of bit-shuffling in software, even physical hardware that claims to be 64-bit but disregards the top 32 bits silently has been tried already) so I'm not sure what we'd have to do to really screw things up, but I have confidence that somewhere at an IHV there is a programmer who is stupid enough to invent something so terrible it can bring about the bright new world that I hope for.
Posted Jan 30, 2013 14:36 UTC (Wed) by mpr22 (subscriber, #60784)
Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds