Re: Porteus, Sounds great, suggestion
Mark Guertin <[email protected]> Thu, 10 Jul 2003 15:39:21 -0400
| Newsgroups | gmane.linux.zynot.devel |
|---|---|
| Message-ID | <BB333929.894%[email protected]> |
On 7/10/03 3:17 PM, "Zach Welch" <[email protected]> wrote: > Okay, so I forgot to mention that I have been designing and implementing > client/server systems for ten years, from scratch, in languages ranging > from C/C++ to Perl. > > I am very cognizent of the problems we face, and I continue to claim > that these features would be both easy for me to fully specify and > design - addressing the needs for scalability, security, reliability, > etc.. > I am still at a loss as to how we can make remote machines do ./configure and link up to libs that may or may not exist on remote machines... This is the boon of a from source solution. Unless things are dictated from the head end of the system (i.e. The server) on the exact pkgs and versions that are installed on all clients at any given time, it will be very difficult to make happen aside from having full cloned copies of each client file systems on the servers...This is the point I'm trying to illustrate here. Doing this as a full binary setup is super simple :) > First, I must solve the problem that *I* do not scale, and then > we can move forward with a next generation package manager. Certainly, > these features are not likely to appear any time soon, but they should > go on the product definition or risk constraining ourselves too soon. > We need that cloning machine :) > That project will probably take a more than a year to complete, but does > that mean we should not try? I merely present criteria for an "A+" > package management architecture, and I will settle for no less than our > best efforts to make that grade (even if we have to make it > incrementally and iteratively). :) I'm not saying we shouldn't try, in fact quite the opposite. If this is a goal we decide to pursue, then now is the time to make some decisions on how to lay this stuff all out as opposed to 6 months down the road realizing that something in our fundamental design leaves us short. Building any one of several versions from source catered to the local machine doesn't scale in this area at all, so we need to possibly take a different approach if this is a definite goal for the distro. Not trying to be negative, just trying to figure out the scope of what we are trying to accomplish :) Mark ===================================== Mark Guertin Technology Manager, Bruce Mau Design 197 Spadina Ave, Suite 501 Toronto, ON, Canada M5T 2C8 [email protected] 416-260-5777 ext. 291 ===================================== This email and its attachments are confidential: they are for use of the named recipient(s) only. Information contained in this email and attachments to it are protected by privilege, work product immunity, copyright and other applicable law. If you are not the intended recipient please notify us immediately by telephone at 416.260.5777 or by e-mail at [email protected] and delete this email from your files.