In fact, I really dislike WACT. I checked out the newest committed stuff and patched a file again that wasn't patched in the CVS yet. Then I went to grab some food and when I arrived back to my home, I saw that Jeff committed the patch :) Lesson learned: checkout only before you actually begin to work.

Until now, I was fine with the simplest rules only (REQUIRED_RULE, SIZE_RANGE_RULE). Now I need to compare two fields, specifically passwd & passwd2, where passwd2 is for confirmation in the registration procedure. So I started to browse through the rules to find out how to write one. For my greatest surprise, I found MatchRule, so I don't have to write anything on my own :)

The reason to hate WACT is that it's fcking good to work with :) Awesome architecture, a lot of pre-written classes. Thanks to everyone who participated in the project until now.

Harry already reported about this but I missed it somehow; an interesting article of a Development Infrastructure for PHP. Scheduled to be read later on today.

Looks like there's a couple of new commits again in WACT, Jeff changed the way tags are registered in the template engine. This doesn't really affect average WACT users (like me now). But hey, let's start testing the new stuff :) Just checked out of the CVS

In the last days Jeff committed the new controller architecture which makes things much more simple. I found a very simple bug in Controller::getRealPath(), that is also patched now. WACT 0.2 planned to be released on 1st December.

These weeks I get more and more involved in WACT. I'm very impressed, it made development exciting again, I didn't feel this for more than a year now. Besides WACT I still use Savant of course, DB_DataObject still speeds up my development, and HTTP_Upload takes its parts in my daily job. I'm a bit sad because I didn't found the time for improving FormBuilder and I couldn't attend The Conference. Everyone's there except me and Helgi, but we promised to attend the next one, if possible. That's the perfect place to discuss TGCP. Wondering what's that? :) It'll be announced during next year :)

Thanks for everyone for the comments and the mails regarding my previous entry. In the last two weeks I came to the conclusion that what I look for is more like an IDE than a run-time form generator engine. I have very big plans in my mind about a new IDE that speeds up PHP development to the max--it's a web app designer desktop application that generates whole websites/portals in PHP. And yes, it can be written to be flexible enough(*). It should be able to generate "plain" PHP code or code using PEAR packages and/or other external libraries (ADODB, Smarty...).

The process could look like this:
- database structure creation (+ additional info like field labels, validation etc)
- form generation from the database tables (add/remove fields, modify field labels/validation, add upload fields etc)
- template creation (skeleton from db tables for lists, forms; they can be modified in any ways)
- sitemap definiton (defining "actions" for web requests)
- fill in the actions with code: this should work using the drag & drop technique, pre-written code snippets like "get record with id #x from table #y", "load template" or "send mail to $address" (which generates a call to the mail() function or puts the mail to the Mail_Queue, depending on the project configuration)
- code generation

The main idea behind this is to get an IDE where you deal with the webapp itself and not the code and/or the libraries you use. You define what you want to achieve, decide on which libraries to use and the system should generate everything for you. It's a must to avoid any possible limitations.

Did anyone see any project like this on the web? Let me know! I'm strongly thinking about starting such a project by myself. Anyone willing to invest in me? ;-)

--
*enough: for developing small-to-middle sized web projects.

I always hesitate which method is easier, faster and more efficient (in terms of development time and costs): to create web forms manually (e.g. copy/paste crappy html code in the form templates) or to use some form generator engine. I know this sounds funny because I'm one of the developers of such a package. Why hesitating then? Because there's no one best solution that fits everything. I had the same problem with template engines in the last month. The conclusion was to use an engine that gives the greatest flexibility and has no limitations.

When I work without the designer friends of mine (respect, 1, 2, in no particular order) I have to write html forms on my own. It's very frustrating to see yourself typing the same code every week over and over again when you know you have more important tasks to do. Additionally, if you care about usability, you have to flood your form templates with so much crappy code to display error messages belonging to specific fields. Sometimes it comes handy to pass back a form freezed to the user for confirmation. If you create your templates manually, it takes many time to achieve this.

I'm looking for a solution which is flexible enough to always fulfill my clients needs (I never say no because of limitations in the implementation) and provides usable web forms, helps me getting rid of those validation and freezing problems, auto-fills form defaults with both static values and those coming from database fields, auto-populates linked fields from databases in a more flexible way than FormBuilder does now. I'd like to get efficient results and I'm going to consider implementing it in FormBuilder, or at least talk about it with Markus and Justin.

The process

1. build a form (e.g. call functions to add inputs, set <FORM ACTION>)
2. fill in the defaults (with one call or walking through the form inputs by inputs)
3. assign the form template (either a static template which is created for a specific form thus giving the greatest flexibility or a dynamic template which describes how to display different field types thus can be applied to any form)
4. parse (and display) the form template

The user fills in the form and posts it to the server. Any time the user posts the form, the server should start with step 5 (so it goes like 5-6; 5-6-7; 5-6-7-8 not 5; 6; 7; 8). This way it is guaranteed that the user gets back the form any time he/she posts invalid values, gets to the confirmation after then (only with valid values that are ready to be processed) and the form is processed only if both the form was valid and it was confirmed (if needed).

5. form validation
6. if the form was not valid, pass it back to the user for correction (go back to step 1 but instead of filling the defaults in step 2, fill it with the values the user just entered)
7. if this form should be confirmed before processing and it wasn't confirmed after the user edited it the last time, pass it back to the user "freezed" (=read-only) for confirmation
8. process the form (this means storing it in a database in most cases)

This flow is what I want to achieve. Additionally, I don't want FormBuilder to generate anything for me by default, I want to call it multiple times, once for each field that I want to include in the form. This way I can a) render fields in arbitrary order b) add arbitrary fields even to the middle of the form if needed. I didn't talk about this with anyone yet, this is what I feel I need. Let me know what you need.

I'm looking forward to see your comments on this, it's greatly appreciated.

More reading for those of you who are interested:

• Apache Cocoon control flow: Continuations (java)


• Forms that work


• Forms, usability, and the W3C DOM

21th August, 2004, Tiszaújváros - the second event of the PHP Roadshow 2004 series. I'll have a PEAR talk, probably between 2:10 p.m. - 2.40 p.m.

Update:
I'll have a 55-minutes session (1.10pm - 2.05pm). The entry is free for the whole event but you have to pre-register yourself.

Newest update: 45 minutes :)

For those of you just starting out with PHP 5, I've found a great article on why it rocks.

Well, I've tried a couple of template engines. I used HTML_Template_IT for a while but it's a replacing system rather than a real template engine and I didn't like this. Then I started to work with Smarty. Then I lost $1500 because I didn't RTFM Smarty docs and I couldn't finish the project in time. Lesson learned.

Well, this sounds a bit lol, but I didn't wanted to learn another language. You cannot just figure out things. You have to learn Smarty. I couldn't use associative arrays (you have to assign the key for a variable first and you can use that variable to index the array - but you can't use a string index directly). Smarty is too complex for me. The other thing I didn't like is that it's a compiling engine and it has multiple directories which should be placed outside htdocs.

Then I met the Savant stuff written by Paul. He proposed it as a PEAR package but the community didn't accept it because there are already 5 template engines in PEAR and (besides I think Savant is a great thing) I feel o.k. that it didn't get accepted into PEAR because of the same reasons that the community said. Then Alan came up with Flexy and showed that it's pretty easy to make wrapper classes for Flexy so that it can act like it was another template engine - Savant for example. He had a blog entry that said Smarty is not the best template engine around there and I hoped Alan had the same in his mind when writing Flexy as me. Then I took a closer look at Flexy and there were three things I really liked. The first one is that you can use special HTML properties in your HTML tags and flexy can parse that out. You can repeat a table row like this:

<tr flexy:foreach="a,k,v">

   <td>k is {k}, and v is {v}</td>

</tr>

This is great. I've never seen any other template engines capable of this. The another great thing is that it can parse your forms so you can assign field values directly from within your PHP scripts. The third thing was a small one: by default, it calls htmlspecialchars() on each output, unless you specify to not to do so.

Then I started to work with Flexy, and damn, again, I couldn't use associative arrays... I wanted to use Flexy together with my favourite PEAR package, DB_DataObject and I couldn't figure out how to do things the right way. I wanted to call PHP functions from my templates but I couldn't.

Conclusion: whatever template engine I've tried (it probably helped me a lot in developing applications faster BUT) all the time, I mean every time there were things that I couldn't do with them.

(Update: "limitations" is a good word here.)

So I switched back to Savant. Simple. Nothing to learn. Nothing that you can't do. Pure PHP. I recommend it for y'all. Give it a try. (it has drawbacks also.. if you want your users to be able to edit the templates, it's probably not a good idea to let them enter arbitrary php function calls.. there may be other things.. feel free to comment).

29th Mai, 2004, Dunaújváros - the first event of the PHP roadshows

I chose Flexy.

I've found a listing of the well-known php template engines and some links to easily compare them. The next template engine I'll try out is Alan's HTML_Template_Flexy because it's similar to Smarty and I like the Smarty concept but not Smarty itself :/ I really tend to like the idea of basing webapps on existing frameworks. It has to be based on PEAR for my purposes. There are two frameworks like this afaik atm, HTML_FlexyFramework and the Seagull project.

I also have to make some improvements on FormBuilder to fit my needs. They'll get contributed of course.

Some reading on Zend: PEAR/PECL Weekly News #1, #2, #3

Zend started PEAR/PECL Weekly Summaries. Issue #1

Time to play with phpoutsourcing's ideabox a bit.