Em Sábado 22 Abril 2006 14:52, jvanasco escreveu:
> That's weird because Jeff Watkins -- the author of Identity -- don't know
> anything about my apps and don't even know how many I've written. Other
> people seem to have no problems with it as well.
googling identity in the groups, i get a ton of threads of people
trying to alter the identity schema like here:
http://groups.google.com/group/turbogears/browse_thread/thread/ded13da1472386dd/1f43b7e35ba38305?q=identity&rnum=1#1f43b7e35ba38305
> Could you show the representativeness of these numbers? Where did you get
> "9/10" from? This list has more than 1500 users and I haven't seen, for
> example, 9/10 of them disagreeing with the current implementation.
> (Your quotation is very messy, please take care not to attribute something to
> the wrong people...)
Well, I pulled that number out of my ass based on 3 things:
a - The source to identity says "Reasonably basic User definition.
Probably would want additional attributes."
b - Every post I've seen in this group regarding identity , including
from the author of the package , has mentioned altering the class
definition
c - Every webapp I've seen ties far more info into the user identity
> You don't need to do that if you don't like it. There are docs explaining how
> to create your own classes and there are options at the configuration files
> to use them. Just do it (do I have to attribute it to Nike? ;-))
> The current design is a suggestion that the author thought would be useful.
> There are step by step docs showing how to create a new model. You found that
> you needed to go further to use them? What did you missed from them?
> The current module is very flexible and allows alternative providers -- for
> LDAP, SSO, SQL, etc. -- so that it can be used together with other projects
> or tied together with other technologies.
> I still believe that you're lacking reading more about how TG work and how
> each of its parts work. Unfortunately, to get where we are now, there were
> lots of changes and the development pace was very fast so docs are starting
> to become available now. The source is your best bet to get up to date docs.
>
> This will be changed soon, fortunately.
I've been reading TG docs and sourceode nonstop for 4 days trying to
get this project to work. And I think most of my problems tie into
this: I picked an AWFUL time to try TurboGears - the docs are all out
of sync with the source ( they seem to be .9a2 <> .9a4 , the source
itself is constantly changing ( new releases seem to break old ones and
come out often) , and there's a new release on the horizion. In
regards to the Identity stuff, the identity information in the wiki is
all out of date.
I'm not a 'great' python person , but I'm pretty decent after using it
for a few years (you're all definately better than me). I *really*
want to do a current project in TurboGears, because while this 1/2 of
it is http, the other 1/2 is all Twisted networking - and I'd rather
keep this project in 1 language if possible.
But after a few days , my experience has led me to question whether or
not its worth using TurboGears -- and I'm starting to consider just
doing this in my mod_perl/TAL framework.
The reason is that TurboGears is so Pythonic, probably too much so.
And while everyone in Python is touting how great that is, from the
outside its annoying and a struggle to work with and adapt to.
People in Python LOVE large files with all their classes in it. Python
people also love tons of files in the same directory. People in web
development hate that stuff - it makes it a struggle to divide up work
and assign tasks to different people -- which is critical when you're
breaking something down into components for a few people to work on
simultaneously in the same room, or cross country, independant of one
another. Its even harder if you're trying to limit permissions to
certain components via SVN.
The tg-big template solves a lot of that in regards to controllers, but
its largley/wholly undocumented (will this change in .95?). If you
want to break up the templates into directories - its entirely
undocumented, and lots of conflicting info exist on the newsgroups (i'm
guessing from API changes over the past few versions) In .9a4
,touching __init__.py and importing master.kid using a relative path
seems to do the trick. ( BTW I have a documentation patch half done
for that . I'll add it to trac later on )
TG is a great product - its super powerful and intelligently designed.
The DB modeling is way better than Django , and the templating
selection/options blow it away. TG is not written in ruby, so it
automatically blows rails away in that respect. But rails got that
default division of components/directory mapping done right , and
Django also fared well on that. Also, Rails lowered the barrier for
entry - everything is dead simple, and you extend from there. With TG,
some defaults (like Identity) are just a bit more than is necessary,
and require knowlege to dumb-down. Thats great if you're doing the
kind of application they're designed for, or you have a good
understanding of the TG internals -- but its just a nightmare
otherwise. Everyone says that Rails took off because of the hype - but
the more I've played with TG / Django , I think its because everything
in rails seems designed from the approach of 'Rails For Dummies'.
While I'd really love to see TG come out on top - so much of it makes
sense and is so intuitive to use - right now too much of it requires an
advanced knowledge of the TG internals to work on projects in a more
web-shop / less python way. I don't know what .9a5 and .9a6 will look
like - but I hope they will be easier for people who don't 'think in
python' .
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"TurboGears" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/turbogears
-~----------~----~----~----~------~----~------~--~---
