Andrew J Guay wrote: > Hi; First let me identify myself as a non Solaris User as of right > now. I have for the last 5-6 years been experimenting in Linux, > installing many versions from many different distributions. To this > day I have yet to master the command line sufficient to be able to do > very much. I have installed some jar, tar.gz, & some sh run. apt-get > & rpm based programs. I can move in and out of root. change > directories, init to different run levels and buld Nvidia Drivers > using their Program. Thats about it. Now to my point. I spent 3 hours > trying to get Solaris 10 into a single ISO image from the 5 different > segments. My problem is at least two fold one the first immage was > downloaded and unziped and goes into a text format. Not sure if it > should but, when I open it un konqueror I see the same image that I > expect as it is identical to the other 4 with the exception of the > "a" designation. Without this being a hobby and interesting to me I > wouldn't give it the time of day. There are very brief instructions > for doing the concatonating in Windows and Unix. No instructions for > Linux I assumed they would be the same as Unix. Anyway I made the > above statement to say that I'm although not very well versed I am > certainly not totally clueless and I found getting into Solaris so > far near immpossible. Which brings me to one of the Unix/Linux type > difficulties. When the Read Me files are created the istructions > inevitably are written with idea of whoever is reding them knows > certain basics. The level of detail that would be required to bring > the novice to the point of actually being able to install the given > application is minimal but increadibly crucial. >
Would it be fair to summarize the above as you're suggesting that the instructions for Linux users should perhaps assume a lower level of skill than they do today? I've excerpted the rest and reformatted a little since your posting got a little run together by the forum posting software: > > My Sugestions 1. Aim Higher, Your aiming to be as good as linux, NOT > GOOD ENOUGH. Apparently, the paper doesn't give the proper impression. We have two goals, on page 20, but neither says anything specifically about Linux. Where goal #1 says "competitive" that means against the competition in the markets we're targeting - Linux is one of those competitors, but so are Windows and OS/X and the other Unix systems. Could we aim higher? Perhaps, but it's not clear that we would get sufficient payback for the resources required to achieve that goal. The question, succinctly, is this: does Solaris, given it's other strengths, need to be the absolute easiest to install for a novice in order to make headway in this market? Our position right now is that being on par with the competition in this aspect will be sufficient for Solaris for now. One thing that does, though, is to leave an opening for other OpenSolaris distributions to perhaps specialize in that particular segment. I'm realizing, though, based on this and a couple of other comments that the Caiman footnote on page 20 perhaps is giving people some specific ideas which aren't meant to be the case. > 2. One aiming point might > be to create or port software through a tool that could do a scan of > the written in language, a scan of the individuals computer hardware > and make it happen for not only Solaris but the multiboot systems > OSes as well Lets face it JAVA is on just about everything Java is > cross platform Sun Knows Windows, Linux, Unix, Apple all of wich have > thier roots in UNIX. I've read this several times and I'm still not sure I understand what you're suggesting. Can you explain further? > 3. Make sure that everything works. I'ts very > hard to find a Linux version that all of the installable software for > that sytem will even function. That tends to rarely be a problem on Solaris, due to our emphasis on binary compatibility. > 4. Instead of backing off from the > idea of a single drive running multiple desktops expand on it make it > easy. I have a total of 8 hard drives one for both Operating Systems > on 4 computers. I have recently been thinking of putting the > Operating systems on a single drive and accessing them externally > from the computer via a network. We're not proposing to eliminate diskless support, instead rethink it for the current and future technologies and try to simplify so that we have fewer distinct models to maintain. Storage networks or iSCSI seem to offer much the same capabilities that our current diskless-over-NFS architecture supplies, so we're not sure there's a need for essentially duplicate functionality. > 5. Utilize in a legal way the know > and available absolutes such as drivers for anyting I have ever > purchsed. I'm not sure I completely understand your suggestion, but if it's to leverage drivers from other platforms, we'd like to. The problem is that the architecture of each platform is very different in this respect. We're always looking at solutions here, because we recognize driver support is a huge impediment to using Solaris on lots of systems. > 6. Their almost none of the linux stystems now that cannot > automatically detect my network and at least get me on line > immediately get me on line. By your own admission yours cannot. (No > Excuse) if you want to be what I believe you can be based on the > soundness of Sun This should be a no brainer. We can, in some cases, but we don't do it well for the common, general cases seen today. The Network Automagic project (http://www.opensolaris.org/os/project/nwam/) is going to fix this issue in what we hope will be a very flexible, powerful way. > 7. Stop with the > grandure of your Solaris system for now. I'm sure it's great for the > server/Corporate community but right now its near useless for me. If > I tried to install your system with all the problems you mention in > your article I would do with it what I have many other Linux stystem > I would pull the disk out of my drive and throw it away. I'm not sure I understand what the suggestion is here. > 7 Aim for > more users. Linux is not aiming just for the server market. Their > aiming at everyone. You need to do this if you want to get where I > think you want to. We're aiming as widely as we can, but we cannot solve every problem immediately, so we have to have some focus in order to make real improvements in a reasonable time period. I don't think we're overly server-centric in what's proposed - if we were worried only about that market, the Live CD and coexistence with other OS's wouldn't be things we'd necessarily focus on, for example, because they are not particularly of interest for server buyers. Thanks for your interest and thoughts! Dave
