> > [...] I.e., imagine MS-mess but on Internet scale, not just
> > desktop-scale. If you can.
> As you said, it's hard enough managing a network with simple and open
> protocols. For M$ to keep OSS on the back foot, they would have to
> constantly "upgrade" (read: change in some subtle and incompatible
> way) their protocols. These upgrades would have to come at a faster
> rate than the OSS community could cope with.
That's exactly what they do on the desktop: Office'97 files unreadable by
Office'95 ring a bell? The stupid crud MSIE uses, so there are websites
that you _can't_ see with other browsers? I fear it is not just theory,
they _can_ (and _will_) use those tactics, and (given the overwhelming
desktop position they have) they even might have a chance. Sure, they tried
to assimilate the Internet, and failed (for now...).
> This is a two-edged sword. If M$ pushes a new protocol or "extension"
> onto the Internet (or even into a corporate LAN/WAN), they will
> *break* that network. It will break because it is simply not possible
> to upgrade all components of a network at once, and older components
> will not understand the new extensions. The result is that the new
> servers will have to be put back the way they were.
Nope. The way is _force_ them forward. More $$$ for M$, as everybody
upgrades the server to version 107 since 106sp15 has glaring security
problems, but then the clients don't work, and... People _do_ complain
loudly than old versions are broken, and that new ones are too. But buy
anyway... delicate balance, that one: Broken enough so that people get fed
up and upgrade, but not broken enough that people get fed up for good ;-)
-- Dr. Horst H. von Brand mailto:vonbrand@inf.utfsm.cl Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/