<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body text="#000000" bgcolor="#ffffff"> On 02/16/2010 04:47 PM, Zoran Popović wrote: <blockquote cite="mid:96b40cab1002161547v6ee2b667t1f9d0fb1a2c4c9b4@mail.gmail.com" type="cite"> <div class="gmail_quote">2010/2/16 Tom Sightler <<a moz-do-not-send="true" href="mailto:ttsig@tuxyturvy.com">ttsig@tuxyturvy.com</a>> <blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">... Still, comparisons based solely on "revenue" does fairly represent market share from Redhat's perspective. Even if revenue was somehow equal, $300 million would be what, probably a few thousand EPIC servers, but that same amount would equate to 10's of thousands of x86 servers </blockquote> <div> or 10's of thousands of blade systems (even bl870s which are high-end in that range at the moment only for shortly, while smaller HP Integrity blades cost not much more than Xeon-based Proliants - I don't know if people around this are aware of that, and that new Itaniums will use same motherboard chipsets as new Xeons) ... ... or a couple of unknown IBM mainframes, whatever. </div> </div> </blockquote> The problem here is that RedHat/Fedora is not a Debian like organization, it's a real "For-Profit Corporation" that requires cash for salaries and to provide a Return-On-Investment for it's stock holders. While the majority of Redhat/Fedora Linux users expect FREE (as in Beer) Linux, that doesn't pay the bills. For Redhat/Fedora to support the massive cost of development, regression testing, and distribution of IA64 means it need a justified revenue stream for that works. That requires Intel, HP, SGI, and other IA64 equipment vendors to fork over significant amounts of money each year to pay for that work. What we saw, is that those vendors cut their budgets to cover the Redhat/Fedora direct labor and facilities costs to support IA64 Fedora. What we also saw, is that Redhat closely manages Fedora, requiring that all key Fedora positions are funded RedHat staff, and that Fedora build facilities have to be on-site and directly managed by Redhat/Fedora staff. A real Debian like community development model doesn't actually exist, the the pretense of a community development model, and because of that, bits not produced directly by Redhat/Fedora must be relabeled and not represented as a Redhat/Fedora product. At this point there are two current FREE (as in Beer) Linux releases that I'm aware of: Debian, and to some degree Ubuntu. It appears that at least one vendor, HP, is actively releasing patches up stream to support these releases. Efforts that were repackaging Redhat/Fedora IA64 efforts seem to have ceased with Fedora IA64 work stopping, which strongly suggests there is not ANY community based IA64 effort to support Redhat/Fedora bits, and produce regular up to date releases that track the Fedora product. There are however, some significant efforts to keep Debian IA64 on track, which are being picked up by Ubuntu, and released. LaMont Jones continues to coordinate Ubuntu IA64 volunteers interested in helping. The latest Ubuntu bits install fine on SR870 Intel reference platforms, but fail with console problems when booted after install. LaMont suggests there are other (HP) platforms that install and operate as expected, and I plan to get an SR870 in laMont's hands for testing soon. In the long term FREE (as in Beer) IA64 Linux needs a community based support program that is not a sham, which means either supporting the Debian/Ubuntu efforts, or finding and building a team and facilities to take on the uncoordinated task of cloning Fedora releases on a regular basis. It also means, that limited vendor funds (both direct cash and donated labor) should be pooled behind a single FREE (as in Beer) IA64 Linux effort if they expect Linux vendor supported releases like Fedora or Ubuntu to continue. Ubuntu appears better funded than Fedora at this point, and is at least interested in trying to release current IA64 bits, at least from my vantage point ... So I suggest we join that effort, since Redhat appears to say it will not pay for Fedora IA64 releases. I'm certainly open to others ideas and suggestions otherwise. have fun! John </body> </html>