+1 Projects are not only source code anymore. Most of my projects have a gazillion of images, javascript, CSS files, etc. Storing all those external files in .mcz packages is not scalable (and not even elegant).
FileTree+Git, while not ideal, solve the issue quite nicely. And you don't need Github anyway... On Sun, Nov 3, 2013 at 4:13 PM, Camillo Bruni <camillobr...@gmail.com>wrote: > > On 2013-11-03, at 15:52, Stephan Eggermont <step...@stack.nl> wrote: > > > Kilon wrote > >> I take a look at previous experiments like squeaksource and I find > little justification to not support Github. But then I am not against > Smalltalkhub or other >repos being available to Pharo. The more the merrier. > > > > I see some very strong arguments against depending on github: > > - it is centralized infrastructure, essentially unsuitable for use with > a distributed version control system; > > - it doesn’t support working at the right granularity; > > - the smalltalk community is too small to have any influence on the > directions github is taking. > > It is a commercial organization that can decide to do something we > don’t like at any time. > > It is free, so we are the product. Just take a look at sourceforge; > > - we can do much better than github (but don’t have enough time). We > should be using a P2P, > > bittorrent like system for version control. > > github != git and whether we use github or now does not matter at all. > What matters is that we use technology that is robust and that we have a > versioning > system that works decentralized. All of that is solved by git. > > With filetree we have the proper granularity (methods) > With github we have an awesome website, such as we have an aweseome > website with smalltalkhub, execpt that monticello should be modernized. > Sadly the community is too small to achieve that, and inventing yet another > versioning tool/system won't help on the short run. Maybe, yes someday in > the future we can have our own fancy fully object-oriented versioning > system, but IMO that is wasted effort, as git/mercurial come very close to > something ideal. > > I am happy to give more insight into git, because I think you have quite a > wrong picture about it :) > >
