On 11/21/05, Levi <[EMAIL PROTECTED]> wrote: > > > > Is STAN useful to Turbogears? > > Mabye. > > But I feel like the real question is: Should Turbogears support STAN > "out of the box"? > > I'm hoping for "no" on this one. One of TG's most valuable assets is > its roadmap. New adopters can skim the docs and walk away thinking > "learn slqobject, kid and a little cherrypy." and be on track.
You are correct. People come to TurboGears because it answers a lot of questions in reasonably good, flexible ways. Adding choices that are "six of one, half-dozen of another" just makes things more difficult for people. I am strongly in favor of MVC separation, but I am also willing to bend the rules in places where the benefit is great enough. (Widgets are an example of that... there's some rule bending going on, but the benefits make it worthwhile.) STAN is definitely a better way to generate HTML/XML than string concatentation. Arguably, though, it's not better than Kid. It's just different. Kid provides a whole bunch of useful (and yet Pythonic) ways to get things done. I also think that creating a Kid template is not much of a burden. If there was a realistic example where STAN would be considerably more concise or easy to work with than Kid in TurboGears, we can discuss this more. (BTW, as pointed out earlier in the thread, you can use alternatives at will with TurboGears. This is more just a question of what's in the core. If *many* people want to use STAN, that likely means one of a couple of things: there are use cases not handled well by Kid, or there are use cases that need to be documented better...) > As a side note, I really like STAN. In fact, I'm sure I'll end up using > it because a) kid just hasn't clicked for me and b) I've never worked > with a designer that could produce production ready html. But I'm not a > new adopter. The great news to me is that I can make this choice > because TG doesn't stand in my way. Yep, you certainly can. > I do think MochiKit's use of STAN is very interesting. Maybe the real > question is "Should STAN replace kid?". So I guess the real bottom line > is "There can be only one!" is more important than "STAN Rocks!". If I > had the freedom to pick a replacement for kid, it would be STAN. "There can be only one!" is the general philosophy (and, I guess, Kid has lopped off STAN's head http://www.imdb.com/title/tt0091203 :) That said, TurboGears also aims to be a sufficiently complete toolbox to deal with a wide variety of real-world needs. If STAN offers significantly more efficient ways of doing some things, it's worth considering. Kevin
