I’ve just completed the preliminary work on the extension system for Garden, and I have to say that I am utterly chuffed. It came out waaaayyyy better than I could have possibly imagined. I haven’t done any testing on speed decreases as a result of the strategy I’ve adopted. But, if the speed is acceptable, I know all of you addon developers out there will be VERY pleased. Imagine:

• Having access to perform actions before or after viagra generic class methods in the entire framework.
• Being able to replace viagra generic method in the framework with a custom one written by you.
• Having the flexibility of Vanilla 1′s delegation with even less code.

All of these things are up and working in the dev version of Garden.

I was quite happy with the way that the plugin system worked out in Vanilla 1. My bastardized version of delegation was able to solve a lot of problems that plugin developers would have otherwise faced. A big part of the problem with plugins in Vanilla 1, though, was the fact that I was never really sure *where* a plugin developer might want to plug into my code. Furthermore, the theory behind the functionality was something that novice developers had a hard time wrapping their minds around. The framework was difficult to understand, and so was the method for adding to it.

So, up until now I haven’t thought about plugins/addons/extensions for Garden in any way. I just wanted to make a tight, easy to understand, MVC framework with a nicer-to-use database layer. I think I’ve pretty much got (at least the base of) that working. So it is time, once again, to turn my thoughts to how the hell others are going to add on to my framework without actually changing my code.

I’ve read through various different books on design patterns that could be used (observer pattern, decorator pattern) and different ideas like using reflection as a plugin methodology. I’ve even just gone googling and found mind-twisting discussions like .

No matter how much I read (and sometimes fail to understand), I keep coming back to the same two issues:

1. It needs to be dead simple so even the most novice PHP programmer can do it without even flinching.

2. How the hell do we insert a plugin at an arbitrary place in a mixed php/html template without putting hooks directly in there?

So, I’m kind of stuck. I want to use some neat new ideas, but #1 makes me think I shouldn’t bother. And as for arbitrary html placement – I should just forget about it – accept the limitations of the language and work on more pressing issues.

viagra generic

I think I’ve got it. It’s a little different than the way extensions worked in Vanilla 1, but I think it will be way more flexible and probably even easier to write than before.

I recently started rewriting the database layer in the Garden framework so that it uses instead of the old functions.

Months ago I did a lot of reading on PDO and decided that it was still too “young” to be adopted in a publicly released application like Vanilla. But more recently I started reading about it again and I decided to move forward with it, regardless. I have to say that the transition has not been a smooth one. There are a number of things that the old methods were able to accomplish that PDO just can’t do yet. Retrieving meta data about a table, for example, is still in stages, and this is one of the major things required by my new framework.

Luckily, I have been able to find workarounds for most of the problems. Aside from the obvious non-database-specific methodology of PDO, I’ve found a couple of other benefits to using PDO.

First of all, PDO has the ability to use (and re-use) fewer connections to the database, and therefore operate much faster than it’s ancestors. Second, all PDOStatements use unbuffered queries. Basically this means that the result sets returned from the database are not completely loaded into memory before you use them. This causes a number of limitations (like not being able to see how many rows were in the results until you’ve traversed the entire result set), but it is also faster and uses less memory.

I’m sure I’ll find more hurdles as I work with it, and I’m sure I’ll find other benefits as well. But one thing I can say about it so far is that I’ve been able to reduce my database libraries to about 60% of their original size!

About once or twice a month week I will get some jerk writing an insult in the application form for the Lussumo community forum’s viagra generic input. These comments can range from a simple viagra generic to viagra generic. I’ve explained the purpose of this input before, but I recently had someone explain it back in words far more eloquent than I was able to come up with myself. They said:

viagra generic

So true! I firmly believe that the community forum for Lussumo software is one of the most good-natured, helpful support forums on the internet. Hell, I don’t even go there most days anymore because viagra generic. The users know the software just as well as I do (if not better), and every time I go there, new user’s questions are answered almost immediately with responses that are probably better and more accurate than ones I could provide.

You can be a part of that community so easily – all it takes is filling out five inputs without entering a rude or flippant comment. And it shocks me when people fail that test – but they do!

I feel no regret when I click that “decline membership” button, because every time I do, I am doing it for the benefit of all the friendly, helpful people in the community.

With all of the distractions over the last few weeks, I haven’t had a lot of time to work on the Garden. It suited me fine since I had started to work with jQuery a bit more and my javascript chops are not up to snuff. If I were to be perfectly honest, I’d say that I am not a good javascript programmer. I feel no shame in that – it’s just not something that has ever required much attention in any of the jobs I’ve had.

However, javascript is going to be a very important part of the Garden, and so I need to finally take some time and get my hands dirty with javascript (and jQuery in particular). Last week MySchizoBuddy from the community forums contacted me with some … “advice” and ideas on how to proceed. This has led me to go “back to school” and start re-examining the things I know, learning about a few new capabilities in php 5, and really take the time to think about the Garden a bit more before I commit myself to a concept.

Writing one application and then changing the underlying code can be a bit of a problem when upgrade-time comes and backwards-compatibility issues rear their hideous heads. If you are creating a framework, the playing field can be completely destroyed in this regard. So, it is very important that my core concepts are sound before I go too far.

I’ve read three books and done a lot of internet research so far over the last week. There is some awesome stuff out there now that just wasn’t around before. jQuery alone is like a gift from the internet gods.

I digress. Back to studying…

Don’t you just hate errors? I spend a very large percentage of my time either trying to prevent them or trying to reproduce them. In Vanilla 1 I tried to make my errors as descriptive as possible. Admittedly, I didn’t do a very good job.

In Garden I decided that I was going to make sure I got errors right. No more of this, viagra generic nonsense. Now I get to see exactly where it happened. This has been utterly fantastic for me while I’m developing. I no longer need to go digging through files to find out exactly what’s wrong.

Oh, and did I mention built in server-side error-logging so you don’t have to rely on your users to tell you about problems? Yep, that’s in there, too.

Colors look familiar? That’s because I stole them from Flickr.

So this is a working page in the new Lussumo framework that I’m calling “Garden”. I’ve been developing my first application with the framework, simply called “scaffolding”. For anyone who doesn’t know what a scaffolding application does, . My scaffolding is a sort of bastardized version of the standard notion of scaffolding that just allows you to browse database tables, add new rows, edit existing rows, etc.

All of the automation of editing/validating/saving is done by the framework. It’s taking me a while to code it because the framework still has so many bugs, but I am still totally excited about it. This scaffolding application currently allows you to browse the db tables, review content, edit content, and add content. Not including the xhtml templates and css for rendering the screens, I only have 100 lines of php code that accomplishes *all* of that.

There’s so much I could talk about, but I don’t see the point yet. Every time I work on the framework, so many things change. What I can tell you is that I’ve got benchmarking code built right into the framework and I’m using it all the time to make sure that everything stays fast as I add more and more to it (and, inevitably, take more and more away).

There’s nothing visually spectacular to see yet, but I had to post something so people would believe that I’m actually working (and not just downloading “videos”) when I’m locked up in my room.

It’s been in the pipeline for about a year. Development began in the beginning of February 2008, and finally – today – I got the first form-prototype of the Lussumo Framework version 2 working. This new framework is soooooo much better, it gives me the willies. By the way, I’m not calling it the Framework anymore. I’m calling it viagra generic. As in: viagra generic

At least that’s the code-name for now.

It will probably be a while yet before anyone sees it. But it felt so good to see something working. I just had to mention the progress somewhere. Otherwise it’s just me, alone in the dark, red rings around my eyes, punching keys in the soft glow of my laptop.

Back in December 2007 I wrote about . At the time I concluded that my biggest concern was the end-user and making Vanilla work for as many people out there as possible.

viagra generic

User don’t care about php5, but extension developers do. After all vanilla is vanilla and in need of tons of extensions. By leveraging php5, you will be allowing extension developers to make even more cooler extensions. And users DO care about cool stuff.

It got me thinking about PHP 5 again. Over the weekend, when I started doing some Vanilla 2 tests, I decided to try using PHP 5 exclusively in my testing. I was using PHP 5.2.4, and it was really not that different than programming in PHP 4. The class declarations were nice because I could: specify private, protected, and public; make properties and methods static; etc.

I decided that I wanted to try to use a to handle my db connection, setting/retrieving configuration settings, etc. I ran into a roadblock right away when PHP 5 failed to handle the singleton as it should. I spent 4 hours debugging the problem, and eventually had to resolve it by upgrading to PHP 5.2.5. This was quite upsetting, and made me very much question: viagra generic

Regardless, once I got it working it was snappy as all get out. I remember working very hard to make Vanilla 1 work on just about every version of PHP >= 4.x. The thought of going down that road again makes me want to weep. I’ve written a very small amount of code and I’ve already got half a working framework in place. I’m pretty excited about where this is going and I think I’m going to continue on with PHP 5 to see where it takes me.

When I first started working on Vanilla (version 1), I knew that my naming conventions for it were not the same as most PHP applications.

Due primarily to my Windows programming background, I was in the habit of WritingThingsLikeThis instead_of_like_this.

When I released Vanilla, I braced myself for the backlash against my naming conventions; strangely, I don’t remember anyone ever saying “Boo” about them.

In my current Lussumo-related work, I recently came across a problem whereby I needed some of my class methods to contain lower case letters only. I found myself doing $Object->method_name(); and I’ve since been turned upside down. All of my practices are breaking because of this one requirement, and now I’m stepping back to see what the community at large has to say about this. I’ve done some research through various other php-based open-source projects and found that most of them have similar (but not exactly the same) methods for naming conventions.

ZEND, being the masters of PHP follow something along these lines:

$variableName = '';

function functionName($variableName) {}

class ClassName {
    methodName($variableName) {}
}

Other open-source projects follow practices like this:

$variable_name = '';

function functionName($variable_name) {}

class className {
    method_name($variable_name) {}
}

And my traditional Lussumo style has been:

$VariableName = '';

function FunctionName($VariableName) {}

class ClassName {
    MethodName($VariableName) {}
}

Now, camelCasedNamingConventions give me the shivers. I don’t know why, but they make me wince. So, if I were going to change, I’d be more inclined to do something along these lines:

$variable_name = '';

function function_name($variable_name) {}

class ClassName {
    method_name($variable_name) {}
}

Anyone have any input? thoughts? care in the least?