Skip to main content

Linux on a USB stick and desktop unreadiness

I got my Dell Latitude D600 to boot off a Sandisk 1GB CruzerMicro USB stick  running Knoppix 5.1. It booted much much faster than my normal Windows XP2 (which is a real pain). Booting off a USB was easier than I thought it would be. The screen passes my eye-candy requirements and the icing on the cake is a sophisticated female voice, arguably on the bridge of a USS Enterprise, saying "Initiating startup sequence".

 

If you try this, you may need to change the BIOS bootup sequence, which may confuse Windows when you get back to your normal, boring life.

 

I had a good mind to boot up from my cell phone by extracting the Knoppix image on the phone itself. However, with just 512 MB of expandable memory, it would not have made the cut. Interestingly, I have Motorola A1200 "Ming", which runs  a customised Linux distro from Motorola.

 

I could not get my WiFi to work with pump or Kwifi Manager. That was the sad part. I wish these things were easier. Is Linux ready for the desktop yet ?

 

My answer is a thundering no. Brief reasoning.

 

  • Display and Fonts. Windows font kerning rules anyday. You may attribute it to better driver compatibility owning to non-disclosure agreements that Microsoft signs with most major hardware vendors. But at the end of the day, viewing characters on Windows is far more pleasant than on Linux. One can argue subjectivity, one man's witch is another man's Moonielizzo.  However, I hold a fair amount of confidence that if you sample 300 unbiased people, the jury will rule in my favor.  you will have to arrange for the logistics though :-)
  •  Device Support. Support on Windows for common devices is way much more better than on Linux. Once again, the non-disclosure element come into play. Specific hardware device details are not available to the Linux developer and so the device drivers operate at a generic level unable to utilise vendor-specific optimizations of feature-improvements. The other problem is that for many vendors, Linux is not considered mainstream enough to warrant writing a device-driver. Take my M-Audio FastTrack USB guitar recorder for example. It comes with only a Windows and Mac driver :-(

 

The second one could change. I am not optimistic enough to say "soon", but it could. Earlier in January this year, Linux kernel maintainer/developer Greg Kroah-Hartman made an astonishing announcement. It was an open announcement addressed to the large number of companies that have devices and corresponding drivers on Windows/Mac/ Sun but not on Linux. His rather generous offer to these companies is that :

Linux Kernel developers will develop complete and working device drivers (free-of-cost) for their devices and the source will be distributed as a part of the Linux kernel. In return (or rather, to assist), the company would have to provide:

  1.  some kind of a spec of the device
  2. an engineer with who Linux kernel developers can talk to, send emails etc for clarifications and deeper understanding (XP: The Customer is always available)
  3. a sample device to test and debug ( methinks its so that the dev knows whether it is a stun-gun or a washing machine (s)he wrote the code for.

The good news is that apparently Greg and his team has figured out some ways around the non-disclosure agreements.

I would recommend you checkout Greg's Monkey Log for a very well articulated description.

Comments

Popular posts from this blog

Barcamps over the world: BCB3/Minnebar

I think I am kinda special. Not quite like Paris Hilton special, but getting a  chance to attend two Barcamps separated by 8000 miles in a span of 3 weeks has got to be some kind of special stuff. I think the big guy above is smiling at me. Invest in my equity. This is an article outlining some of the interesting differences I saw between the barcamps in Bangalore (BCB3, 31 March -April 1, 2007 ) and Minnesota(Minnebar 2007, 21 April). This is not an article intended to compare or pass a judgement. Just throwing up some observations, fwiw . I am not offering explanations, I am not a socio-anthropology by training. Some of these do not require a degree to arrive at the reason of causation, but I want to keep this blog close to what I saw, not what I think. At most, some "could-be"s. Both the barcamps have a local flavour and preservation of local flavour to me, is inherently good.  Consider food, for example.  A predominantly South Indian buffet spread for lunch at BCB3 a

Bambi 2.0

Bambi is a small coding-fest that we organise in our group at GE Healthcare. It was inspired by Yahoo Hackday after I heard about it at BarCamp Bangalore last year. I still remember, I came back all charged up after BarCamp and with some help from Arun B, we put together the first version of Bambi. Ours was a small team, roughly about 60 people, so spreading the news was not much of a problem. Getting people out of their workload was a bigger problem. The load is high and the work is, I guess, somewhat exciting ;-)  It is sometimes tough to lure people out of writing indexing algorithms for proprietary image databases or mitral-valve plane adjusters for segmentation of the human heart.     Today we had the demos for Bambi 2.0 The quality of demos were much improved and people came on the last day with some utterly cool demos. Unfortunately, I do not think I can write about them in detail owing to Intellectual Property issues but a mash of  Biometrics, Bluetooth, MRI scanners and

Talk at Barcamp Bangalore 3

Rakesh and I shared some thoughts on unstructured innovation such as unconferences and codejams in large, well-structured (and somewhat paradoxically) very innovative companies.       Technorati tags: barcampbangalore3 , innovation