Since Twitter itself is way restricted into the 140 chars (which I like most usually - less offtopic and media) I will better blog this off here as a couple of things I don't like in Cake.

First of all, I've been doing Cake for a year and a half and have several projects up and running with different web services, sync mechanisms and so on. It's usually one of my preferred platforms (right after WP and Croogo which is actually Cake based) but it doesn't mean that I adore all of the features in it.

Auto recursive models 

By default linking the models in Cake sets a recursive level of 1 - so you get a direct access to level 1 of all corelated models. it is usually nice as you don't have to join or bind models every few requests. The bad part however is that every serious project (and even non-serious ones with more than 10 tables!) gets bloated with so much insignificant data not being used anywhere in the site. When we have a product which has categories and part of an order, which has a user and so on there are lots of queries back there and tons of useless data which leads to reduced performance and page load time. 

Yet again - very useful, especially for non-technicians who have hard time joining and so, but I get different requests from oDesk or local clients with old Cake sites that need optimization and fine tuning cause the site used to work at the beginning but the previous developer uses the standard recursive=1 settings and therefore the more the database records, the more hardcore the end results.

ACL

ACL is... well, it sucks big time in my opinion. It has a wrong concepts at the beginning, it is also hard to implement (lots of years technical background and lots of hours, if not days, trying to setup something that needs to be in the core). @jose_zap - I like Croogo's way of setting the jquery matrix of roles and controller actions and predefining the actions for each role. The UI plugin for the standard ACL is too complex in a usability manner and doesn't do the work.

Another thing is the role based auth. Cake does a pretty good job restricting different roles, but the autogeneration of MVC implies that no user-based authentication would be done or so. Another few projects of mine used to fix actions accessible via URL (no controller backend checks) and protecting every single add/edit and listing as index/view from unauthorized data listing. Or in other words - user number 2 is usually able to change the URL and see the listings of user number 3 or click the edit/5 link and edit the records of another user. It is not hard to implement it manually, but it takes time and having the logic predefined and working and generating tens of MVCs from tables opens lots of vulnerabilities out there.

i18n

What I don't like here is basically using one table for all translations by default. Once I tried setting different tables for i18n for the different DB tables but it was kinda tough for me to set up the models. Also the multilingual content with the localized data (so to speak i18n with the l10n together) had to be implemented by some third party tutorials with lots of app_controller magic in between.

---------

I had been able to fix all of the concerns above back then. It just happened for them not to be straight forward or workarounds are painful which leads to discomfort while coding or revamping an application.

These are my top concerns for Cake so far. Pretty sure I've had many more back there, but now as I do explore a framework, I usually look for several things first:

• multilingual support
• user management
• security
• design adaptiveness and plugins capabilities

I think I have some hard time with AJAX as well, but can't recall the specific projects with it. 

Already had two trainings so far and few more to go. There is going to be another batch on November for another colleague of mine and I'm trying to sync my materials to serve as a good reference during the next batch of trainings. 

Last time I was doing Croogo work was in December, but recently I had to do some development on small Croogo-based projects and due to the chance that some of the features are going to be reused later, decided to do some plugin work. This is the first plugin released and I will consider contributing another one or two small thanks to Fahad's work on Croogo.

As an employer I prefer to charge on a project basis. It's easier to plan my budgets and the costs of the final results. 

As a contractor I prefer hourly payments. It helps dropping the pressure from incorrect specifications and further negotiations after the project and the budget have already been set.

However I am well acquainted with the pros and cons of both methods. But there are three terms that I believe most clients use in an inappropriate context without gaining the maximum productivity and optimal costs for their projects. We're speaking about the hourly rates, the amount of work and availability.

Hourly rate

As I said I got tired of arguing with clients with dummy specifications and clonings of the "Clients from Hell" samples. I don't expect all of them to be technicians. But usually the time dedicated for meetings, specifications and other meetings during the process exceeds the actual development time. And also the initial meetings and negotiations are still 'Awaiting' and projects are not actually accepted so there is a serious risk of "This price doesn't fit to my budget, I cancel" so... yeah.

So I prefer the hourly rates. I estimate single tasks and modules based on what I know. In case of any misunderstanding we have a short chat/call with detailed data what is the extra work all about, what else needs to be done and how will this reflect the previous estimate. On approval we work, on rejection - I don't waste my time. Simple as that. 

What annoys me most is that clients are looking for specific hour rates. For instance: "I am able to pay $10 per hour" while I am estimating some projects on $20/h. And I get rejected because of my _hourly rate_ without even being able to estimate the amount of work to be done.

This is completely wrong for one simple and basic reason. A task to be solved is based on the following formula:

Solution = amount of time * hourly cost * quality coefficient

Lets start backwards.

1. Quality coefficient - a decimal variable between 0 and 1 that values the final quality of the solution. For instance, talking about a website creation, a QC of 0 would be completely useless end product. A coef of 1 would be a multilingual website (if needed) which is W3C valid, cross-browser compatible, reusable and well documented source code, stable and secure. 
2. Hourly cost - simply the price for working for 60 minutes.
3. Amount of time - here is the tricky moment. This is the time that an expert needs to complete a given task. You know what? My $5-coders I work with are pretty slow and their end work quality often needs refactoring. While they solve a simple problem in 4+ hours with a coefficient of 0.7, my core developers for $20/h complete their assignment in less than an hour with 0.95+ quality. And asking for more.

So, what I am about to tell is: don't judge the coder by the hourly rate. It's irrelevant until you know the end estimate and the quality of the work.

Amount of work

This is another standard requirement which is particularly based on the hourly rate.

In my experience there are 2 types of projects - maximum work for a short term or flowing work which is long term. Most employers require long-term non-stop working. This usually affects the productivity of the worker and his/her motivation as well. 

If you work with an experienced coder or designer, or whatever who could do proficient and high quality work for a short time, you need to give him/her the freedom of the 9-5 working office schedule. Sometimes this is not possible due to some company requirements about the availability and meetings and so on, especially when people do work in different time zones. So here it is step 3 - the availability.

Availability

The availability is the time range while the worker is available to contact the project responsible person. However most clients expect that the person is available only during the time of the actual work and also, that he/she is working all the time during the availability period. Which is completely not mandatory.

For projects that need some quick reaction and support the contractor must guarantee availability hours and _not_ work hours. Its just like the support positions - 2 people should be _available_ 12h for any possible exceptional situation. But it doesn't necessary mean that they need to do actual work during this period. 

My practice is working on 2 or 3 projects a week, as the projects vary in their specifications I dedicate 40-50 hours a week at total. I am available 70+ hours a week for a client if needed to get some consulting or just status updates and I do quick and clean work for the 15-25 hours that I have agreed on working for the given project. This usually leads to best results at a minimal cost.

What I could have done for the conservative clients?

I don't like cheating and I usually cancel projects and reject client invitations for people that wouldn't understand this article. But, you know what? If I am confident in my skills (as I am in few technologies I've been polishing my skills in during the last 7 years) I could have done the following:

• estimate projects completion time multiplied by three or four 
• giving half of my availability

As a result I would charge for 4 times slower work and 2 times the price I would charge normally with a flexible client. I would mark in my work diary 120 hours of work (3 projects * 40 hours) while it is hardly possible while I actually work on my normal rates and schedule. No one would even doubt if I am a fast and clean worker. But why should I cheat if we could agree on my preferred rate and a flexible working time?

Check Sonne here.

There we are - https://dev.twitter.com/ . I would really like to see a feedback from the Twitter dev team for the usability of the platform. It's a well known fact that Drupal is a great platform for developers but has an non-intuitive interface for end clients which is one of the reasons WordPress is so popular right now. 

So I decided to add a new setting configurable through the JavaScript call in your call-menu file where you could use php/java/python/whateva to retrieve the correct path dynamically and just pass it. I'm using v.1.5 and I did 2 corrections:

1. in the ddsmoothmenu.js file replaced all smoothmenu.arrowimages strings with setting.arrowimages (just changing 'smoothmenu' to 'setting' in these three lines). There are 5 strings to be replaced at lines 75 to 77. What I do is add a new setting for them in the next section.
2. Adding the paths to the images as a setting. So at the end my call includes the dynamic path to be used in the menus. Since I'm using WordPress, here there is my code for calling the menu:

 
ddsmoothmenu.init({
				 mainmenuid: "header_top_menu",
				 orientation: 'h',
				 classname: 'ddsmoothmenu',
				 arrowimages: {down:['downarrowclass', '<?php bloginfo("template_url"); ?>/img/down.gif', 23], right:['rightarrowclass', '<?php bloginfo("template_url"); ?>/img/right.gif']},
				 contentsource: "markup"
			});
 

Few hours of clicking around I was happy with the light version of network - the standard minimalistic design (total lack of any design according to some users) and lightweight platform unlike Wave. The invitation procedure that we know since GMail works as well - plenty of mediocre users wanted them to be _in_ only for the being special case. However.

A great review on the subject is the comparison of Plus with Wave and Buzz. I'm hitting the +1 on that one.

However, people that comment on a blog post normally have no way to get feedback eventually if anyone comments back in the same post. The post author (and administrator) receives notification for the comment but the comment author, on the other hand, has no natural way to be pinged back for a reply. This is a serious leak in the WP standard functionality. It is at least unethical not to inform someone for the reply (which might occur in a day, month, year even more).

WordPress comes with a standard feed for latest posts and feed for recent comments as well. However subscribing for all comments in a blog or finding a specific thread to subscribe for is not usable and not practical as well. The solution is the Subscribe To Comments plugin that adds a checkbox to the comment form which allows one to subscribe for further comments in the same thread. This is completely optional and up to ones preferences, but instead of breaking the whole conversation because of the 'echoing' this provides the instrumentation for a real communication.

WordPress.com uses Subscribe To Comments for 2 years or something, it's integrated in their web service. So why is it not included in the platform yet?

Hello everyone, today we will discuss the hot topics on the Internet and specifically the WordPress platform and some premium hints from the ThemeForest market presented by one of the authors there - Slobodan Kustrimovic (a.k.a. wpcanyon). He is an expert in the WP world and has published few themes on ThemeForest and currently working on a new framework in his passion to constantly improve the quality and usability of his products. 

----------------

M: Hi Slobodan and thanks for your time. How are the things going on your side?

S: Hi Mario.

Can't complain, everything is great at the moment. Currently working a new framework that'll give the user even more flexibility and also be a great base on which we'll develop themes faster.

M: Great - can't wait to see your first demo. Where are you from?

S: I'm from a little country in Europe called Serbia, right next to your country.

M: What do you do for a living and in your spare time? 

S: I make premium WordPress themes which are sold on ThemeForest and in my spare time pretty much what a typical 22 years old guy does, watching movies, hanging out with friends, getting drunk occassionaly...

M: Sounds like fun then - glad that you are able to manage your spare time and still produce such a high quality of templates. I have been exploring the TF market for months and I have noticed your products there and the activity online. How many themes have you released on ThemeForest?

S: We have published 4 themes by now (Lifeline, Brainstorm, Eptonic, Nuance), the latest one was published few days ago. I also developed one theme as a freelancer some time ago that was for ThemeForest.

M: Do you work alone or in partnership with other authors?

S: It's a partnership, Justin is the designer and I am the developer. But there's more to it then just designing and developing, there's testing, making the demo, handling support, writing documentation...

I admire those who do everything alone, it's simply way too much to handle for one person.

M: Indeed. How long does it take for a theme to get "up and running"?

S: About 25 days from the start of designing to submitting the theme to ThemeForest. But we plan to cut that time to about 15 days so we can release 2 themes per month.

M: Pretty ambitious. Which is the most time consuming element in the whole cycle - planning, design, development, testing, release, support, upgrade?

S: Well I can't say for sure about designing since Justin is the designer, but i think development and designing take about the same time and they take the most of the time from the whole process.

Support is also time consuming.

M: Name the most frequently asked questions from your clients via the TF support? What do clients need more or don't understand by default?

S:

1. "The theme doesn't work, it says the stylesheet is missing"

This one is by far the most frequent question all the authors get. The reason is that people are used to just install the theme with the zip they download, but on Themeforest that zip contains more then just the theme so WordPress gets confused. So what the customers needs to do iz unzip it and inside of it they'll find the actual theme zip.

2. "The theme breaks a plugin"

But in 99% of the cases the theme doesn't brake the plugin, it's the plugin that brakes the theme.

M: Where do you go for marketing - are there any specific channels where you popularize your theme or warm up the buyers?

S: The best thing about Themeforest is that you don't need any special marketing, it has a huge traffic and it's very popular. But we do use our twitter account to promote our items a bit. We do plan to start a little website to share design freebies, write tutorials and publish free themes so that'll be a nice way of promoting our themes.

M: What is your weekly load in a time manner for ThemeForest - i.e. how many hours a week do you dedicate on this?

S: Can't say for sure, sometimes i work more sometimes less, depends on the mood i'm in at the time. That's the best thing about premium WordPress themes, you work how much you want. If i have to give a number then i guess it's about 50 hours a week.

M: What was the toughest thing at the beginning? Have you changed the way you work for TF based on the leaks from your first theme?

S: The toughest thing was falling asleep the night we submitted the first theme.

I was simply way too excited and a bit terrified so i simply couldn't. I kept hitting F5 to refresh the page every 20 seconds to see if there are any news. And when it was approved and published, well let's just say F5 on the keyboard doesn't work as good as the other keys since then.

I didn't really change the way i work but the way the themes work has certainly changed a lot thanks to tons of suggestions from the customers.

M: I know that feeling of excitement - hopefully you will continue to develop yourself and your themes and have larger challenges to check out!

Where do you see yourself in 1 year?

S: I'm pretty much sure i'll still be making WordPress themes with Justin, no need to change something that works well. And i hope we'll be in the first 20 top selling authors by then.

M: Do you foresee any major change in the way developers and designers create WordPress themes and the ThemeForest market? Probably new fancy changes in WP 3.3/3.4 or another popular product to be themed such as Tumblr? We see the OpenCart category drastically growing up right now.

S: The way developers and designers create WordPress themes for ThemeForest is certainly going to be changing quite a lot in terms of flexibility and usability, not because of WordPress versions but because the auhors try to make every theme better then the one before.

WordPress has alredy evolved into a wonderful CMS and i don't think any major changes will be made. I would like if they improved the current features instead of adding new ones, the shortcodes system definetely needs to be rewritten.

About OpenCart it seems to be getting quite popular, sales are nice, but i don't know much about it.

M: Could you spoil a part of your new framework for the readers - some useful feature or anything for the developers to build against?

S: Well as you know we have that nice content composer, it'll be rewritten so it's more flexible and one awesome addittion will be "live content composer" which will allow the user to also make changes right there on the post/page instead of doing it in the admin.

M: Name your favorite ThemeForest authors and top themes (except yours, of course )

S:

Authors

1. Orman Clark

2. Kriesi

3. DMThemes

Themes

1. Gridlocked

2. Periodical

3. inFocus (you got to admire the sales, 6K+)

M: Are there any regular resources that you follow (online media, blogs, tutorials)?

S: Tons of them Here are a few of them:

justintadlock.com

yoast.com

wpengineer.com

wpcandy.com

digwp.com

nettuts.com

speckyboy.com

smashingmagazine.com ...

M: OK, that was great - thank you for your time!

S: Thank you for having me