[TYPO3-dev] Advanced Forms

[Elmar Hinz]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Edwin,

it's nice to have this beginning of a concept. But it will be probably forgotten
if it is only documented in the mailing list. I propose to open an own project
page in the wiki.

In this case I can add it as "project searching for a team" to

http://wiki.typo3.org/index.php/Extension_coordination_team

* So it will no be forgotten.
* Other people can add own ideas.
* Discussions in the mailing list can point to the porject page.

Regards

Elmar


Edwin de Jong schrieb:
> Hello developers of Typo3,
> 
> Does anyone knows if there is an extention that is 'like' mailForm but 
> actually is able to group the form fields into mulitple sections and even 
> pages? Have conditions and such? If not, I will write such an extention. 
> Therefor, I really appreciate any ideas, suggestions for implementation, etc.
> 
> Basic functionality proposals:
> 
> 1. per field validation based on conditions (each field can have multiple 
> conditions, like regular expressions or simple arithmetic comparisons)
> 2. fields being grouped into sections
> 3. sections being grouped into pages
> 4. hints (question mark with popup on the right)
> 5. conditional field display and section display (for example, when a user 
> marks a checkbox a javascript will display or hide a certain other section 
> based on changing the style)
> 6. Progress indication (for example, when an extended form is placed over 
> three pages, a big 1, 2 and 3 display above which change color from grey to 
> red to show where you are)
> 7. ability to change between 'in-line' form display and 'block' form display 
> (basically, sometimes you want a layout that has the form inputs layout 
> horizontally eg. when entering a birthday or a postal code and a city name)
> 8. Different and extendible types of entry fields
> 
> Some of the rendering of the above mentioned points can be done with 
> templates.
> 
> I'd rather design this from the ground up. I'm experienced in making these 
> kind of applications. As a developer for a local company I've designed such a 
> system being fed with information from an XML messaging server. Here part of 
> the validation happened on the XML server, part happened on the webserver 
> (Cocoon) and part in the users browser. *Very* complex system, for the rest 
> of this mail I use a lean version of this.
> 
> While designing I just ran into the same problems we were having before... for 
> example: each field, section and form has two predicates, one which - if true 
> - will display the item, and one which - if true - will validate the item. 
> Each predicate is associated with other predicates by inheritance, there are 
> predicates like quantors, logical operators, arithmetic comparitors, regular 
> expressions etc. etc. Some of these predicates have operants which are not 
> predicates but parameters (eg. the value that was entered by the user, an SQL 
> query that returns one value or even the results from typoscript). The 
> problem however is that: a user should be able to navigate through the 
> different pages of the form and there is a small chance that these predicates 
> are bypassed when using the back button of the browser. Basically: when you 
> make the visibility of a page [n] (or subelement thereof) dependent on a 
> value on page [m], where [n]<[m], then you can still reach page n by using 
> the back button. Basically, the predicates need to be 'ordered'.
> 
> For this I think what I need is a small functional language with list 
> comprehension to do the predicate checking, might be overkill though. Eg. to 
> do:
> 
> FOREACH f: this (f.type = 'date' => f.value > date_born_field && f.value < 
> NOW())
> 
> if you want to check if the dates entered by a user are actually after he was 
> born. Since these languages are all easily parsable, and can be built up in a 
> hierarchical way, its quite easy to make a user-interface to change these 
> expressions like the one above.
> 
> Ok, I've also made a simple entity relationship diagram. Basically, what I 
> want to put in the database is:
> Form (relates to 1 Predicate/NULL)
> FormInstance (relates each to a Form and zero or more FieldInstances)
> FieldInstance (contains a value in XML format (to allow Date's and such)
> Page (relates to one Form, relates in 'next' to Page/NULL and relates in 
> 'previous' to Page/NULL, relates to 1 Predicate/NULL in 'visible' and relates 
> to 1 Predicate/NULL in 'Validation')
> Section (relates to one Page, relates in 'next' to Section/NULL and relates in 
> 'previous' to Page/NULL, relates to 1 Predicate/NULL in 'visible' and relates 
> to 1 Predicate/NULL in 'Validation')
> Field (relates to Section, relates in 'next' to Field/NULL and relates in 
> 'previous' to Field/NULL, relates to 1 Predicate/NULL in 'visible' and 
> relates to 1 Predicate/NULL in 'Validation)
> Type (relates to 1 Predicate/NULL in Validation)
> 
> Well, at first I also wanted to model the predicates in the database, but now 
> I feel more in making either a seperate language ala Lisp, or use XML to 
> describe them.
> 
> Basically it contains the following entities (I only describe one, the rest is 
> obvious):
> 
> BinaryLogicalPredicate (like, the 'AND' operator), contains a predicate A and 
> B and resolves to true if A and B are both true. In the same sense: OR and 
> FOLLOWS.
> 
> UnaryLogicalPredicate (like, the 'NOT' operator)
> 
> QuantorPredicate (like the 'FORALL' operator which has three arguments: one 
> unique identifier, one reference to a Form, Page or Section and one 
> Predicate. Is true when the Predicate is true for each 'identifier' 
> substituted by the reference to each Field contained in beforementioned Form, 
> Page or Section).
> 
> UnaryScalarPredicate (eg. 'IsSet' scalar predicate which has one argument 
> reference to a field F and returns true when F is set by the user)
> 
> BinaryScalarPredicate (eg. '>' operator which has two arguments A and B which 
> are of the same (sortable) type and returns true when A > B). Also in this 
> league: a regular expression matcher.
> 
> I guess we need also a set of operators on the Scalars, but maybe we can use 
> the help of PHP for this (eval... better not).
> 
> Well, this is my first quick specification draft. I would really appreciate 
> the input of others on this subject.
> 
> Regards,
> 
> Edwin de Jnog


- --
Climate change 2006 is killing people: floods in California, drought and fires
in Australia, Texas, Sahel, Oklahoma, South Africa. The Bush administration is
responsible for corruption of the Kyoto Protocol. The US majority is responsible
to the world for reelection of a convictable [...censored by Echelon...].
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFD11T+O976RNoy/18RApKiAKDo0LQhk9M8gzavTXXtaYJ+ULHtgACgjUjm
dhdySOhTfD7SnbQeXdMuSfQ=
=CDB1
-----END PGP SIGNATURE-----