The link below is to an article at silicon.com predicting winners and losers in 2004. Linux is in the first list for servers and the second for client side, partly because of the unnecessary, confusing and destructive battle between KDE and Gnome. While I don't necessarily agree completely with the article I think it has some fair points.
http://www.silicon.com/management/ itdirector/0,39024673,39117578,00.htm?nl=e20040106
I'd also like to predict some more losers:
1 The public wi-fi access point. As long as prices stay where they are, that is. 2 Java as a client-side solution, given the increasing bloat with every version release. (That may well apply to .NET too, for the same reason, but I wouldn't know.) Why in hell can't Sun or IBM modularise the thing so if you don't want it all you don't have to download it (or force your customers to)? 3 Sun - already in the loser list, but for the additional reason that they don't understand end-users. Some fine ideas but seemingly targetted at a world that won't exist for another five or more years. 4 Anyone who expected broadband to be taken up by every Mom and Pop in the land. 5 Mom and Pop, who can't download the service patches from Microsoft and thus catch every virus and worm going. Nice to think Linux could gallop along on its white horse and save the day, but it's too fragmented and difficult for said Mom and Pop to set up. After all, ain't nobody else gonna do it for them.
And just to show I'm not all negative, a winner:
Computer users everywhere, able to use an increasing base of good-quality free software as an alternative to overpriced commercial stuff or nag'n'bitch shareware. When this trend gets far enough along, coupled with much better standards than at present and making source downloads a LOT easier to use, Linux may emerge as a viable desktop candidate for ordinary people.
-- GT