Re: Better start with Wonder than WO
Re: Better start with Wonder than WO
- Subject: Re: Better start with Wonder than WO
- From: Mike Schrag <email@hidden>
- Date: Thu, 20 Mar 2008 17:03:35 -0400
I would totally disagree with that. Wonder introduce another layer
of complexity and patching on top of WO. I think you should only use
Wonder if you have a good reason for it.
This is just untrue, and misleading. First of all, I know you don't
use Wonder, but I also know you had your own frameworks when you lived
in the real world with the rest of us, and I suspect you wouldn't
start a new project without using those frameworks, because it would
be silly and wasteful. To normal developers, Wonder introduces a VERY
thin API layer (extends ERXApp, ERXSession, use ERXEC factory), and in
exchange you get thousands of man-hours of development and enhancements.
This ridiculousness that it "adds a layer complexity" is completely
upside down. It REMOVES complexity -- huge complexity. You know when
the last time I ran into an editing context locking problem? I can't
tell you, because I don't run into them. You know what new users do?
They run into editing context locking problems. The only ones that
don't are the ones that take queues off of Apple's laughably bad
examples and build entire apps using
session().defaultEditingContext(). I now no longer care about inverse
relationships -- Wonder does that for me. People who use D2W don't
have to care about terrible locking issues, because Wonder solves
those problems. We offer drop in remote object synchronization. We
have a stateful component-based Ajax framework. We do migrations,
prototypes, component-based emailing, attachments, tagging, captas,
natural language date parsing, and lots more. We're talking thousands
and thousands of man hours of other people's sweat here.
As far as the layer of patching we add on top of WO? That's because
WO has bugs that have historically taken YEARS to fix. We fix things
that don't work, and we provide those fixes to the community so that
people can get real work done in spite of those bugs, and we make it
completely transparent to the users of Wonder. When those fixes are
no longer necessary, we roll them out as we move up our suggested
deployment platform. When a version of 5.4 comes out that we feel
comfortable recommending to people, we will move our baseline to 5.4
and remove fixes that are no longer necessary. Currently we support
both 5.3 and 5.4, though 5.3 is our recommended platform. Believe me,
none of the Wonder committers want to be fixing bugs in the core
frameworks. We'd much rather be spending our time writing things that
are actually enjoyable.
So the answer is _USE WONDER_. Take advantage of other people's
work. Let them solve your problems for you. There are an enormous
wealth of features available to you in exchange for a very small API
commitment. If you're starting a new project where you are free to
make the decision (and you don't already have your own frameworks to
build on), there is just no reason not to use it ... You're only
making your life more difficult. Just look at http://projectwonder.blogspot.com/
and browse the "This Week in Wonder"'s. Those are just the new
features since January 1 of this year. There's another 5 or 6 years
worth of cool stuff in there.
ms
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden