If Linux Didn’t “Suck” It Wouldn’t Exist
by Wyatt WalterThere was a great presentation that made it to the top of digg in the last couple of days on how Linux “sucks” and how we can fix it. It was a very good down-to-earth overview of the problems with Linux, but the solutions that were suggested disturbed me a bit. Now, I love Linux. Just like the presenter, I’d like to get that out of the way. I use it every day on both the desktop plus manage several Linux servers at work. I also realize that there are many weaknesses with Linux. Wireless, audio, and video controls are sometimes sketchy at best. Examples would be that dual monitor setups are commonplace, yet sometimes Linux distros struggle with configuration of them. That’s a reality that I don’t deny.
I whole-heartedly agree with the presenter’s list of problems in Linux. They are 1) audio inconsistencies, 2) hardware driver issues, 3) display configuration issues, 4) a lack of commercial backing of software, open or otherwise, and, finally, 5) software packaging differences. Numbers two and three I whole-heartedly agree with. We still are fighting issues with hardware compatibility as well as display configuration is a trainwreck. At the risk of sounding like another fanboy, the latest version of Ubuntu has made some serious strides in the display configuration department, but there’s still room to improve.
Numbers one, four, and five I have a real problem with the solutions presented. Problems one and five take the freedom out of Linux that its creators and biggest advocates have come to love. It takes away a developer’s freedom to decide that the way something was done was inefficient or ineffective and do it a way that they prefer instead. If we are handed a package management software created by some arbitrary company and told not to ask questions, then what makes Linux different than, say, Windows? Linux offers that freedom which is what has made it into what it is today. Take that freedom away and most likely instead of having multiple distributions of Linux, but you’ll end up with different OS’s on top of the Linux kernels other than GNU or even new kernels. Then we won’t just be duplicating a little work, but all of it. Why does this happen? Because the way I do things is better than the way you do things. You might think that your way is better but it’s not. Trust me, it’s not.
Another problem with trying to bring all Linux distros under one package management standard is that it turns into a very slippery slope. If we bring all Linux packages into the .rpm format, then who pays Canonical to switch Ubuntu? In fact, why not just have Novell dump Yast and Debian drop Aptitude? Why stop there? Why not just move everything to Red Hat? That’ll make sure we don’t duplicate work. Okay, so I’m taking it a bit further than suggested, but it’s the freedom and competition that we like about Linux. Take that away and I’m not sure how many would stick around.
The fourth problem becomes a bit interesting. It’s a chicken before the egg problem, really. Companies who are going to have to have a market before they can afford to create a product for Linux and there has to be a product before they have a market. This problem could be solved with some good marketing and venture capital if companies are willing to take risks, but it may or may not happen. I would agree with the presentation that the open source ecosystem aren’t as sensitive to the fact that software costs money. Often a lot of people assume the wrong definition of free software and insist that software should never cost money. Having the ability to customize and redistribute software has been fundamental driving factor for Linux and I’d be afraid of the way it would change if a lot of the software wasn’t released under some sort of open license.
The presentation is a refreshing, pragmatic way of looking at the status of Linux as a whole. The hardware compatibility and display issues need some love. The other issues.. well the solutions presented seem like double-edged swords to me. The purpose behind Linux has been freedom and if it changes, we will see a transition away from Linux to a new OS which in reality would just be Linux by another name.
Tags: linux
Filed under Tech Trends :
Comments (1) :
May 1st, 2009
Follow whatan00b on Twitter
May 1st, 2009 at 12:16 pm
To the extent that he touch on a series of broad topics he did pretty well 45min. Two things though that he misses –
1) Canonical, Novell, Red Hat have carried the ball IN THEIR NICHE about as far as they can. Their niche of course is as OS providers NOT application layer purveyors. So though it is suggested the ball is in their court. The real matter is its not their court to be playing. The apps that are there in the distro exist out of their graciousness not their need.
2) Risk Mitigation. Remember it. Lets face it, companies are where the bucks are at. So if you want to be successful you need to be able to sell to them in a big way. It is key to Red Hat’s success. Many say Red Hat will not touch the desktop market. Au Contraire. They do sell a ‘Commerical’ desktop to corporate clients.
RH’s key to success has been Risk Mitigation for business. Any Fortune 10K buying a product quickly heads into the weeds of ‘the contract’. Price? Negotiated? Yes, but that is generally 10% of the effort, the other 90% revolves around who takes what risks and the defined support response taken on by the vendor. Bottom line — companies are risk averse and base their purchase decisions on how much of that risk they push off to the provider.
We all know that as an OS Windows is a less than stellar product but companies buy it in droves for the primary reason that by doing so they push much of their risks onto Microsoft via a support contract.
Solutions:
* The OS providers need to get together and develop a uniform risk mitigation program for Linux. HP did it 2 years ago for their HPUX –> Linux push.
* Push not the cart about Linux when it comes to peripherals. The horse is the problem — ie the hardware mfrs. The Linux Foundation should push for a development that implements a cross platform middle layer. Yes the driver layer may be closed source. But the Linux platform would have faster adoption by the vendors to provide those drivers.
* Application products. Quite honestly it would be in business’s interest to fund projects. An example would be the relationship Mozilla foundation has with Google. GIMP if it could find 10 Corps willing to provide $50k each would have the funds to push GIMP to Photoshop levels. The cost avoidance by those companies not having to pay for $150 per seat charge is the payback.
Mere observations.