Dear Fink community: Since my computer skills and diplomatic skills are both quite rudimentary (an assessment based on reality, not modesty), I've been reluctant to say much (apart from accepting the blame for suggesting the latest scipy-core be updated, which I based on some mistaken assumptions). Just as there are core members of fink, there are people like me who are very peripheral in the sense that all of the packages I maintain are at most dependencies for other packages I maintain. So if I mess something up, I am probably the only one who suffers the consequences (along with the end-users). This gives me something of an "outsider's perspective." Hence I hope you might permit me a few observations from that perspective (with my full advanced acknowledgment of its limitations).
1. X11.app Like many end-users, I have been updating to the latest X11.app from Xquartz, not because I crave the bleeding edge, but simply because I understood this to be the best way to have as few bugs in X11 as possible, and also because I (wrongly) assumed that the latest (2.4.0) would be what came out with OS X 10.6. I got burned, and got into a typically unpleasant discussion on the X11 mailing list. The net result is I learned I should have stopped at or before X11 2.3.3.2, but more importantly, that the X11 in /etc/X11 that came with 10.6 would remain untouched, except for Apple bug-fixes, and future Xquartz releases would install elsewhere (presumably /opt/X11) to avoid clobbering the officially distributed Apple X11. I think this is a Good Thing, assuming all fink packages work with the SL X11. We should just assume it is there, and not concern ourselves at all with what gets installed in /opt, or support it. What the end-user does is out of our control, but in principle there is no reason we would have to worry about this one way or the other, as long as we can rely on the current contents of /etc/X11 being in place. 2. "Stable" vs. "Unstable" I tell people to translate the word "unstable" as "current" when they express shyness about using it. With the SL upgrade, we have, at least initially, an inversion, where packages in Unstable may work when packages in Stable do not, or are obsolete, etc. I wonder if it might be worth changing this demarcation to "Core" and "Peripheral" or some such designation, the main point being to have the talents and energy of the "core" fink team focused on the most important packages, and to keep people with limited abilities, like me, out. 2.1 Core branch Put packages including all the core fink packages, gcc44, fftw, tcltk and friends, gtk+2, gnome libraries, python, guile, and all the other packages that tend to get built primarily as dependencies for other packages into this branch. These are the ones that are by far the most critical for end-users, would be of the most use to have available as a stable, binary distribution that mainly got updated only for bug fixes, and would benefit the most from having the fink core maintainers maintaining them. Then allow the core maintainers themselves to decide on whatever policy they think is best for keeping the core packages of fink up and running smoothly. Because maintaining these packages would be the highest priority, it would be best to focus the talents and energy of the best and most capable people on these packages to ensure they all work together to form a seamless infrastructure. 2.2 Peripheral branch Put packages that are of use to a subset of end-users (e.g.: my crystallography packages come to mind) here. Mistakes in these packages are more easily tolerated, because they do not cripple the core infrastructure of fink, and so these can be entrusted to peripheral maintainers who possess more limited skills like me. Distributing these as binaries is a much less compelling concern. To decide where a package goes, just ask what would happen if the package were to be deleted from fink. If the answer is that the presence isn't critical, then assign it to the Peripheral branch. 3. Let's all take a deep breath I had the good fortune to have a former Apple VP (who was responsible for much of Xcode) as a graduate student in a class last spring. I asked him what it was like, and he described his erstwhile job as helping highly creative autistic people to have productive careers. I know it is a bit of a stereotype, but computer people often aren't qualified to be Japanese diplomats, and it is also very hard for even the most socially adept individuals to infer intent from a pile of ascii text characters. We all need to give each other the benefit of the doubt when interpreting each other's words and actions. We all want the same thing, after all: a viable, fully-functional and up-to- date fink. Peace and joy, Bill ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Fink-devel mailing list Fink-devel@lists.sourceforge.net http://news.gmane.org/gmane.os.apple.fink.devel Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-devel
