I thought it might be a good idea to bring this together into a thread targeted at discussing what we can do towards a path of resolution. Many of us have the same complaints, and have had them for years. We've watched our installations go from <300M to >700M for a "minimal install". The primary resolution offered has always been "do it all manually in your kickstart". I, and I can tell I'm not alone, think that this is not a good answer. But we have let it be. I'm going to try and keep my thought flow as concise as possible, but if it strays a bit I apologize. I am really looking for collaboration here, to start a dialog. My thoughts are by no means the only way things should be, so if it sounds that way, it was not intended to.
Before I start on what I feel needs to be discussed, let me point out what we should remember is on our side: RedHat One of the biggest contributors to open source. A strong proponent of LSB. A strong supporter of the community. IMO they would also benefit heavily for this type of work. I'm sure myself and others on this list would love to have them join in a dialog about why we think this should be better and what RedHat's opinion of it is and how we can help each other to make RHEL even stronger for its target audience. Fedora Fedora is a VERY active community that takes input from anyone who wants to work towards those goals. They have project goals for each distribution, which may be a good way to approach a technical solution to this issue. They also have Revisor[1] for creating re-spins (of which there is a minimal install respin of f8, I don't know if they have done one for f9 yet), which is also being made to work well with RHEL. Now lets step back and look at what we are complaining about: Linux Standard Base (LSB)[2] Handled by the Linux Foundation[3]. A wonderful concept. Truly it is. Per their website they aim to be the "highest common denominator" across Linux distributions[2]. This is something we do NEED to have to help ISV's support us. The problem, as I see it, is that it is to broad. There is no single "linux install". Our installations are not just servers. Our installations are not just desktops. So why is it that when I install a server, I get more desktop applications and features than server features? In general it is because of the LSB. The install needs to have bluetooth, it needs to have ISDN, it needs to have printer support. Why? because a desktop user should have these (well, that's debatable in many cases). The LSB already has some basic workgroups targeting various needs. Carrier Grade, Desktop, Mobile, etc.[4] What boggles my mind is that there is not a Server work group, that I see. Maybe this would be a good point to raise, and even a possible split in the actual LSB. As I mentioned in the original thread, how about lsb-server, lsb-desktop, and possibly even lsb-mobile? Would not such a split help to solidify both the standards and the support of the community for those standards? The services could/would even be completely complementary. You want a your base server to have base desktop components, install both lsb-desktop and lsb-server packages and their dependancies. The LSB does have methods established for communication. I do not know how often this has been addressed with them, but re-iteration by a portion of the community would in theory be something they should listen to. There is a User Advisory Council[4], which is supposed to be a method for us the users to make our suggestions heard. Listed on the page is also a mailing list, lf_uac. To go to this group and present something, it would be wise to have an idea of what we really want to see accomplished. A semblance of a unified goal. If we want to see resolution we can't just send a random message that says "hey fix lsb, we don't like its requirements". I've listed one of my thoughts. Your turn. What resolutions are there to this? How can it be made better? What are your thoughts? Hopefully that all came out clearly and helps get something started. -Greg Swift [1] http://revisor.fedoraunity.org/ [2] http://www.linuxbase.org [3] http://www.linuxfoundation.org [4] http://www.linuxfoundation.org/en/Workgroups [5] http://www.linuxfoundation.org/en/User_Advisory_Council _______________________________________________ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
