On Tue, Jan 15, 2013 at 4:40 AM, John Crisp <[email protected]>wrote:
> On 14/01/13 18:15, Charlie Brady wrote: > > > > > > I don't think those timelines are necessary, or realistic. > > Better to have a plan than no plan. Fail to plan, plan to fail ? It can > always be amended. > It is pointless without people committing time. As before, I can no longer commit this time. Those who are able to do so need to stop waiting for someone else to do it, and get involved. This was true years ago, and is true today. Either way, as a first step I would agree with Hsing-Foo on moving to > 6.x asap. > This needs people committed to building releases. Much of that work is "Change version <a> to <b>, adjust config, lather, rinse, repeat". Once you have documented, reproducible steps for making version <b> work from the current working version <a>, someone needs to commit those changes and build new packages. Once the packages are built, someone needs to be build an ISO and then the lather, rinse, repeat continues for the installation and upgrades. The packaging steps are arcane, but are also easy/mechanical, especially if they are package tweaks. The bulk of the work - making it work - is something anyone who administers an SME Server should be able to do. Setting milestones is great, but let's assume there are a hundred packages that need tweaks to work with CentOS6 (I have no idea how many actually need these) and that each of the cycles I mentioned above takes eight hours of lather, rinse, repeat. That's eight hundred hours people need to put in. It's easily divisible by the number of people involved. Thanks, Gordon > > > > >> a. What do we need > >> b. Who do we need > > > > I think that depends on how it is done. I suspect the existing > > contribs.org infrastructure is over-heavy for what we do. The tools in > use > > are designed for building Fedora. I think we can use something more > > lightweight. > > > > If we do things as we currently do, then we need buy in from Shad and his > > various deputies. He probably already has a 64bit CentOS 6 build farm, > but > > we'd need to know that for sure. > > > > If we do things differently, we need to plan that from scratch. > > You said "I know from work I have done at Mitel that moving the code to > a CentOS6 base is not a hugely difficult task." > > I don't have the knowledge or experience to quantify this, but it is > something that needs quantifying. > > I presume that if we use the current system, it will be quicker in the > short term, but possibly more costly, whereas 'lightening' the system > means building a new 'development environment' first, which would take > time, but possibly is cheaper and easier/faster to use ? > > For us mortals, would you care to outline things a little more and what > sort of time frames might be involved for the options mentioned ? > > B. Rgds, > John > > The man who can't even SPELL Future !!!!! > _______________________________________________ > Discussion about project organisation and overall direction > To unsubscribe, e-mail [email protected] > Searchable archive at http://lists.contribs.org/mailman/public/discussion/ >
_______________________________________________ Discussion about project organisation and overall direction To unsubscribe, e-mail [email protected] Searchable archive at http://lists.contribs.org/mailman/public/discussion/
