M$ had a huge headstart on GNU/Linux locking in OEMs, retailers and ISVs for years. In 2008, Apple started its “app store”
tm and Android/Linux started it’s “market place” a few months later. Now, Apple has 2/3 of the app market and Android/Linux has about 1/3 with nearly one million mobile apps between them.
Will GNU/Linux ever catch up on the apps? Probably, after many more years. GNU/Linux is growing in usage and ISVs will soon see it in their best interests to serve the community. Besides the short lead-time of Apple over Android, that situation was totally different because Google and the OEMs really pushed the OS. We are not there yet with GNU/Linux. RedHat, Novell, Dell, IBM and Canonical are doing a lot but it’s nothing compared to M$’s partners. Still, M$ cannot kill Android/Linux nor GNU/Linux so they will keep growing share until there is some equity in the market place. Then ISVs will port to GNU/Linux if they want to keep share.
I expect other regions than USA will generate apps for GNU/Linux. USA is just too locked-in. The rest of the world will double its number of personal computing devices in the next few years and will have ample opportunity and motivation to port apps or to create apps for GNU/Linux. Already we have Android-x86 able to move into x86 spaces and ARM is ready to compete with x86 in desktops and notebooks and thin clients. It won’t be long. I give the ISVs a year or two to realize that GNU/Linux is “there” and a few more years to expand the availability of apps for GNU/Linux.
A consequence of this is that distros will have to be more accepting of external repositories because most distros have not the resources to host a million apps. Debian GNU/Linux is huge with 30K+ packages and takes a year or more to package one release. The load on developers and servers grows faster than the population of packages. A standard package and file-layout will be needed. APT is a good choice but has not been tested at these scales. Obviously mirrors cannot afford to grow to such sizes so distros will have to move to some server farms to distribute packages unless ISVs could manage to integrate somehow with the distros. A search engine specializing in finding packages may be the best option. Doing more things with web applications may be the best solution since it is easier to obtain OS-independence. ISVs may find web apps work better for them with GNU/Linux because there is no packaging problem and they can bill customers directly by usage.
These are problems that need a solution in the next few years. Will Google step in to provide service? Will a distributed system of mirrors and servers scale better? It will be interesting to see how the good features of current distros can be spread to a more diverse environment.