Brock: You can hit me up personally to help with migration - anytime.
Before someone says 'wait I need brand new hardware now?' just wanted to point out that you don't always need extra servers, as Raj stated it's a matter of downtime preference.
Examples:
-Old hardware > new hardware (with very minimal downtime)
-Old os > vm/temp server for interim > new os on old hardware (designed to get back on the same hardware with moderate downtime)
-Old os > storage somewhere > new os (a little more downtime if you have a window)
Either of these using any migration method.
Ok, so we have these same discussions whenever a platform is discontinued. This one boils down to the fact that we didn't give far enough warning (for long term planning time) and feedback input when we were looking to minimize the list. And for that we apologize profusely.
The code is out there for anyone to build from. Is it a death sentence for a platform when we stop producing binaries for it? No. Because if deep down it's "a rock star operating system" that everyone adopts in the field, it'll be back as a official build. Loosely it's the law of averages: If an OS has certain features that make it stand above other offerings, it will either be in demand by enough people to warrant it's return, or that feature will be incorporated into other leading distros. Infact that's the best part about the linux kernel being open source; if a platform starts ignoring requested enhancements another rises to fill the void. The others can choose to survive by adapting, or fall by the wayside. Either way you win.
No that's not an attack on OpenSUSE - I've run it & love it - but using that concept relative to Zimbra: When a platform is perceived as valued on one level or another people will make Community builds of ZCS for it. When more popular still there will FOSS builds, and if enough customers say "hey we want this" it turns into an NE release. That last hurdle may seem extremely daunting because it's hard to see just how many requests we weigh against other factors, and we can't always put the bottom line numbers out there publicly.
I'd like to say download and install reporting stats be damned, but no one truly is able to in today's market. There are entire companies devoted to ROI & analysis: television ad effectiveness, the amount of sugar in soft drinks for best enjoyment, comparisons of peanut butter vs Nutella sales in Italy. Of course the reply is "Zimbra should help kick start all of these distros into popularity with official ZCS releases" and trust me I always argue for more platforms, as I think it's worthwhile to say 'we can' and have a wide range of choices provided directly by us to help spread adoption of Zimbra. The important opposing argument of course is the time and money to maintain them, in both technical aspects and human overhead for both engineering, QA, and sometimes support. Reducing variables is also one of those things that leads into the quality discussion on the other thread. As we push the boundaries of collaboration sometimes 'more can be better' while other times 'less is more'. Everyone has to make decisions daily about allocation of resources, whether it be an NCO in the military to a CFO in a corporate environment. That doesn't come easily, which it why 'compromise' is never listed near anyone's definition for 'fun'! |