The Linux Question… and Answer

Steve Sims thinks Linux can win. He sent me the following plan this morning, and it was too interesting to keep to myself…

I’ve used Ubuntu for maybe four months now. Before that I used Windows. I haven’t always been into computers; until a few years ago I still used Yahoo!’s IE shell for web browsing, and then I discovered this blue E that didn’t take forever to sign into. My father later introduced me Firefox but I rejected it at first since I couldn’t figure out how to see pictures (I think he disabled them for some reason). So, unlike old-school Linux users I know what it’s like to not be great with computers. From this experience, here’s what I think Linux needs to dominate PC operating systems:

  1. Have a package manager that can convert RPM’s and other commonly used packages into debs. Debian’s package manager is more reliable than RPM so it should be used. Since most packages are now released under both RPM and deb this is less of an issue than it used to be, as is the KDE/Gnome compatibility issue.
  2. Install both Gnome and KDE libraries by default, so that KDE programs can run on Gnome and vice-versa. The GUI can be built into the kernel like in NT if this becomes undoable, but it’ll probably work and the slower performance won’t turn too many people away.
  3. Either include proprietary codecs and pay licensing fees or try to defeat the DVD Forum / Fraunhoffer and Thompson Multimedia (mp3) in court. At least with DVD’s, removing the copy prevention circumvention technology from libdvdcss2 would probably make it legal. I think the first option is better since many patent holders (i.e., Thompson, Microsoft) make FUDish claims about open source products without ever taking their producers to court.
  4. Hide the command line, even at the expense of security. With my password, I should be able to modify root’s files with the GUI in my regular account without first going into the terminal and typing “sudo nautilus”.
  5. Install all programs into the same folder. It could even be called /Programs, like in PCBSD.
  6. Linux is not Windows. People won’t want a cheap Windows imitation. They might want something better and different. I can tell you from seeing so many college Mac users that consumers are willing to try different products and will remain with those products if they decide that they’re superior.
  7. Specialized applications. Gyms, lawfirms, etc. all rely on a few specialized Windows applications. These applications are often data management applications, which aren’t too complicated. It seems that “companies will adopt Linux when specialized applications for it exist, and specialized applications for it will exist when companies adopt it.” Regardless, once a Linux company has enough money from home sales it can begin to tailor custom applications to specific industries. This part will not happen overnight, so desktop Linux should first be targeted toward home users.
  8. An IDE that combines the GUI with the code. Microsoft Visual Studio is very easy to use and probably makes developing specialized applications much cheaper. A Linux Visual Studio, especially one that could output in Gnome and KDE, would make developing these applications much easier.
  9. And a few changes in strategy! Embrace, extend, and extinguish proprietary software. Too many open-source projects seek compatibility with proprietary stuff without making it much better. It might be less bloated, it might be a little more stable, but except for Firefox few go beyond what the proprietary programs do.
  10. Patentleft. Red Hat licenses their patents royalty-free as long as they’re used in free software. More patents licensed this way would prevent proprietary developers from implementing open source extensions.
  11. Make it work. At first, limiting hardware selections will be the only way to deploy Linux. People don’t want to tinker with it in order to get it to work.
  12. Ignore the nerds. They can play with Gentoo all they want. They aren’t experts on usability.

The only thing left is devices. The main devices for home users are cameras and printers. Most digital cameras use some type of flash card and newer computers often come with built-in flash card drives.

As for printing, many printers work for Linux while many don’t. This is the biggest technical roadblock to Linux adoption, at least for switching an existing Windows computer to Linux. It would be solved, however, if printer companies open sourced their drivers. This wouldn’t give their competition any extra edge and they could censor the parts that include patents. In fact, it would help the printer companies because open source driver developers would rewrite the printer companies’ drivers (off the parts of the drivers the companies are permitted to distribute) to not include royalty-incurring patents. Thus, the drivers won’t incur a single cent for patent licensing and it won’t cost the printer companies a cent to develop these drivers.

In short, Linux can’t do what Microsoft did because it lives in a different world. I read somewhere that at first Microsoft didn’t replace Apple/Amiga/etc. initially, but instead it put computers on desks that didn’t have them before. Now almost every desk has a computer. So, the biggest place for Linux to grow is in newer markets, but it can still do it on the desktop.

So, that’s Steve’s take on the situation – what’s yours?