On 12/20/05, Kevin Dangoor <[EMAIL PROTECTED]> wrote: > On 19 Dec 2005 22:49:23 -0200, Jorge Godoy <[EMAIL PROTECTED]> wrote: > > > Then your application egg will only depend on the turbogears egg. > > > > +1 on that. Modularizing parts such as identity, i18n, catwalk, toolbox is > > interesting for applications that won't need that, but making it easy to add > > -- either for when those applications grow up or for new applications -- is > > very interesting. > > > > The smaller the core, the better, IMHO. > > setuptools makes some of this possible. Ronald's admi18n tool can be > separated out, but i18n is somewhat pervasive. > > Personally, I don't see shrinking the TurboGears egg as a huge > priority right now.
Grr... Google broke my session when I clicked the send button, losing most of the message. Personally, I don't see shrinking the TurboGears egg as a huge priority right now. In 1990, I had an 80MB hard drive. Now, my little PowerBook has a 60GB hard drive. Space is not a huge issue. that said, setuptools includes a lesser-known feature called "EntryPoints" that TurboGears already makes use of. TurboGears defines EntryPoints as a way for people to plug in new functionality. If an egg includes pieces that fulfill an EntryPoint, then just installing the egg is enough for TurboGears to spot them. You can already add tg-admin commands this way! (This will be used for widgets, Toolbox tools, etc.) My thought is that the core package will include the functionality needed for a large portion of webapps. i18n, identity and widgets all fit this description. The feed generation code that Elvelind wrote is probably the most borderline, but I think that feed generation is an increasingly important part of all kinds of different apps today and belongs in the core. Kevin -- Kevin Dangoor Author of the Zesty News RSS newsreader email: [EMAIL PROTECTED] company: http://www.BlazingThings.com blog: http://www.BlueSkyOnMars.com
