I polled my local RoR UG and the lack of webserver choice regarding your proposed components was the main comment/complaint.
Depending upon what the costs are to do it up front, I feel that it may be worth pushing back the stack delivery by a few weeks rather than postponing 6 months+. (6 months is a really long time these days). -Brian On 9/15/07, Matt Ingenthron <Matt.Ingenthron at sun.com> wrote: > Prashant Srinivasan wrote: > > (snip...) > > > > > > Question is, how much choice do we want to build into the stack? Are > > we focussing exclusively on development environments? > > > > Personally, for the majority of users (and that's not to say large or > enterprise deployments), they'll want to deploy on the same bits they > develop on if the capability is there. In our long-ago past we did > 'desktop' and 'server' versions of OS consolidations-- it doesn't make > sense to revisit those days. One stack that has the appropriate > attributes for both development/deployment is possible and preferable in > my opinion. Just as OpenSolaris has stuff for desktops and servers, the > web stack should be for development and production, if you ask me. > > > How do others on this list think about this? > > > > Brandorr wrote: > > > >> It seems like a good list. Any thoughts on Nginx and lighttpd? > > In the long term (which may be just 6 months out), I think we should > consider either or both of them in the stack personally. > > Like I've expressed before, most users are just looking for packaging in > their OS distribution, so from that perspective, there will be those who > decide upon lighttpd or nginx and we'd love to have as many web stack > users as possible on OpenSolaris, whether it's Solaris + AMP, Solaris + > lighttpd/mongrel/RoR, Solaris + nginx/apache2/mod_php/postgresql, etc. > In fact, it was this kind of thought exactly that prompted me to post a > comment regarding software clusters to Stephen Hahn's latest (excellent) > blog on packaging [1]. This is definitely a little different than other > consolidations in OpenSolaris, where there's a lot of thought put into > adding something since it may need to be there for a very long time. > > Let's walk before we run though-- get a good, solid, out of the box > experience with a complete set of common extensions into Nevada. Then > we can think about other components. > > Shanti has already blogged on building lighttpd[2]. Once PHP and Ruby > hit Nevada, perhaps we we can assemble some blogs to help those who want > to build on top of the web stack in Nevada. It's not the same OOB > experience, but it'll make things easier for those kicking the tires on > OpenSolaris. > > Longer term, if someone wants to pick up the work of integrating > lighttpd/nginx (either a Sun engineer or a community member), that would > be great in my opinion. > > - Matt > > [1] http://blogs.sun.com/sch/entry/pkg_1_a_no_scripting#comments > [2] http://blogs.sun.com/shanti/entry/lighttpd_on_solaris > > -- > Matt Ingenthron - Web Infrastructure Solutions Architect > Sun Microsystems, Inc. - Global Systems Practice > http://blogs.sun.com/mingenthron/ > email: matt.ingenthron at sun.com Phone: 310-242-6439 > > -- - Brian Gupta http://opensolaris.org/os/project/nycosug/
