The lift list will be moderated for a while

[David Pollak]

Folks,

There was a totally inappropriate post to the lift list.  I deleted the post, but it still made it into peoples' mailboxes and I sincerely apologize for that.

I changed the flag to "moderated" and you'll see a delay between posting and the time your post hits the list.

This list is a place for people to exchange ideas, thoughts and opinions about lift, other web frameworks and computer languages in general.  Everybody is welcome and encouraged to voice their opinions about these topics and topics related to these topics... and yes, team dynamics related pretty directly to the language and tool-set folks choose.

I actively (and generally privately) discourage people from making personal attacks, name calling, and judgments about the personal aspects of given poster.  *EVERYONE* is entitled to their opinion.  Disagreement with the substance of a post is different from judgment about a person's worth.

Topics and subjects that mock other's beliefs are unrelated to computing and are entirely unwelcome on this list.

I take the quality of the lift community very seriously.

Thanks,

David
--
lift, the secure, simple, powerful web framework
http://liftweb.net

[Viktor Klang]

David,

 

I fully understand you point of view.

In this day and age, our freedom to voice our opinions is getting smaller and smaller by the day.

 

Mocking is not mature behaviour, but who is to judge what mocks what?

I for one want us lift-people to be peers and treated as equals.

 

Cheers,

 

-Viktor

 

[Derek Chen-Becker]

David Pollak wrote:
> Folks,
>
> There was a totally inappropriate post to the lift list. I deleted
> the post, but it still made it into peoples' mailboxes and I sincerely
> apologize for that.

All I'm going to say is that somehow we did (myself included) get
off-topic. In that light, I'd like to discuss some ideas I have about
forms and validation/behavior. In some of the forms I'm designing I have
a select that controls what kind of querying will be performed (top N,
raw logs, etc). Right now I can control the form by adding id attributes
to the generated form fields like:

val startDateField = S.text(startDate, startDate = _) % new
UnprefixedAttribute("id", "startdate", Null)

And then reference the id with some javascript in the xhtml template. I
feel like doing it this way may be somewhat brittle (and dirty) since
now I technically have code in the template; I believe ithat's something
we're trying to avoid here. I'm trying to figure out if there may be
some way to unobtrusively add support directly in lift for controlling
form behavior. Another thing that might be useful if we're going to add
in javascrip support is having client-side form validation. From my
brief reading of the Mapped* classes, it appears that they do a little
validation, but I think it might be useful to have validation support a
little lower level. I'm just starting to think about this, so if there
is already support for this and I've overlooked it then please let me
know. I'm also not entirely convinced that it would be good to add this
support to lift, so I'll listen to arguments on either side.

Cheers,

Derek

[David Bernard]

Hi,

I don't know the current architecture. (I work on clientside validation for JSF (2 years ago), and currently implements a solution for Wicket 1.3 and jquery validation framework).
My basic suggestion is to have a "Collection[Validator]".
* The collection could be filled handly (ex: validators.add(myValidator)) or by a AutoValidators (helper that add Validators to the collection from binded field meta data (Mapper info, hibernate
annotations, ...)
* Validator must provide a read-only api :
* to provide information for end-user message (if validator failed)
* to allow ClientSideValidator to be created and mimic server side configuration => hook to parse collection[Validator], and to insert in the generated html the configuration, js lib, css,...
* to allow AutoValidators to avoid put twice the same validator :)

Feedback messages should be filled by server or by client

there are 2 groups of validators : before convertion (only work on String) and after convertion (work on other type : int,Date,...)

My current skills (scala and liftweb) don't allow me to suggest some code.

My 2cents (requirements).

/david B.

[David Pollak]

Folks,

I put up a page on the wiki to allow us to design this feature:
http://liftweb.net/index.php/Unified_Validation

Can folks post requirements, suggested implementations and pointers to
how other frameworks handle this on that page?

Thanks,

David

--

[David Pollak]

Viktor & others,

After some thought and some private email exchanges, I've changed my
mind about moderating every post.

A huge thanks to the folks who are helping guide me and a big thanks
to the community as a whole.

David

[David Bernard]

David P.,

Sometimes a graph is better than text, could you install graphviz extension[1] (or similar) into the wiki ?

[1]:
http://www.wickle.com/wiki/index.php/Graphviz_extension

Thanks,

/david B

[David Pollak]

Will Do.

--
David Pollak
http://blog.lostlake.org

[David Pollak]

On Nov 13, 2007 7:58 AM, David Pollak <d...@athena.com> wrote:

Will Do.

Done
 

[Julien Wetterwald]

By the way folks, Graphviz is an awesome tool. In college, while
designing a project for a software engineering class, I programmed a
compiler from an UMLish language to the Graphviz DOT format. I used it
to generate all class and dependency diagrams. It took me a week-end
to program the compiler, but then I could design the software in a
single text file and could generate all diagrams from it (dependencies
were inferred) whenever I wanted. On the other hand, other groups
spent weeks redrawing diagrams using crappy and expensive UML modeling
tools. Why loose time laying out boxes while an algorithm can do it
for you?

I really recommend using Graphviz whenever you have to output some
kind of graph. The DOT format is extremely simple and perfectly
adequate for a logging facility related to some kind of graph data
structure.

Cheers,
Julien