Davey Shafik's Pixelated Dreams

Well, after my tease the other day, let me announce the official proposal for Zend_Service_Server:

Zend_Service_Server Proposal

Please feel free to comment

- Davey

I have been so very busy hacking away on this latest version of Tooltip.js and I think that it offers major enhancements over previous versions.

The changelog is as follows:

• Major Feature Enhancements!
• Add ability to hide current Tooltip when clicking anywhere in the document, or opening another Tooltip (on by default, Tooltip.autoHideClick)
• When autoHideClick is on, don't hide current tooltip when clicking on an element when the element is a child of the Tooltip and it does not have a class of 'close'
• Add ability to have the Tooltip move with the mouse (off by default, Tooltip.autoFollowMouse)
• Add ability to automatically hide the Tooltip after a timeout. (40 seconds default, Tooltip.autoHideTimeout)
• Add ability to have the Tooltip displayed where the mouse clicked and stay there for the duration of that show event. The Tooltip will automatically find the edge of the screen and *will* stay within the bounds of the right edge (on by default, Tooltip.autoMoveToCursor)
• Add ability to set the "Close" link text that is auto-added ("Close" by default, Tooltip.closeText)

You can see how to use these new options in the Advanced Tricks page of the Wiki.

There is also a minor BC break from previous versions. Before, I was using the display CSS property for the show/hide effects, however this conflicts with the Script.aculo.us effects. So now I have moved to using the (more proper) visibility CSS property.

However, this means that you must set the CSS for the Tooltips to use position: absolute; to make them not distort the page layout. This is fine if you're using the autoFollowMouse or autoMoveToCursor options, however when you want them to display in the page you start to have issues. You either have the Tooltip content going over the other content of the page (if you have a transparent background, this is just not acceptable!) or you have it in the wrong place.

So, the solution I am using on the Tooltip.js Wiki is the following in <script> tags:

// Turn off autoMoveToCursor - bad for the huge comment blocks!
Tooltip.autoMoveToCursor = false;
// Overwrite the showMethod with a simple one that just changes the position property back to static and calls the scriptaculous effect
Tooltip.showMethod = function (tooltip, options)
{
    tooltip.style.position = "static";
    Effect.Appear(tooltip, options);
}
// Overwrite the hideMethod with a similar one, but make the element absolute again.
Tooltip.hideMethod = function (tooltip, options)
{
    Effect.Fade(tooltip, options);
    tooltip.style.position = "absolute";
}

Just like Last Year I'm jumping on the bandwagon again.

Despite the fact that it is hard to top some of the feats I managed last year (I for one, am not planning to marry anybody else!) it has been a very productive and exciting year for me.

This year, I got my Work Visa in just 9 days, and because of this blog, I was notified about a job being offered locally. My card arrived on a Saturday, by Tuesday I had my first interview, Thursday my second and I started work the following Monday. I am still happily employed. To everyone involved there, thanks so much for providing such interesting challenges for me to meet on a daily basis.

February began my first and thankfully, second look into PDO, my move to Serendipity and I spent most of my time settling in my new job.

With March came the first stable release of PHP_CompatInfo and the first release of Crtx_PEAR_Channel_Frontend.

April had its share of fun with the Jason Macer/OSI Hosting debacle coming to light.

May is my Birthday month, and it was my time to move to the darkside and try out Ubuntu Linux (which I am still using) and finally, PHAR/PHP_Archive came to the public eye.

With June also came the little cries of a new kitten, In July, we finally saw some public releases of Cerebral Cortex. I had my first interview in August, for the Pro PHP Podcast

September had myself and my wife travelling to and back from Toronto for php|works, my first conference where I had a great time. I also attained my Zend Certifications.

October was somewhat more quite, we added 2 more kittens to our household (with the loss of two of our cats, this left us with a total of 6), whilst November heralded the first (and subsequent) releases of Tooltip.js.

December entailed many good things, and some not so good. Firstly, I recieved my full green card (and a 19" LCD) and unfortunately when I shutdown Cerebral Cortex after a mere 5 months in the public eye.

So what does next year entail? Well, I'm poised to release a new version (vastly improved) of Tooltip.js - this will be beta and hopefully the last release before my initial 1.0.0 Release Candidate.

I am also now working for Zend on the Zend Framework.

Finally, I was also accepted to speak twice at php|tek which is almost in my own backyard!

There are other things I wish to achieve this year; I want to hone my project management (and people management) skills. I want to write a book and I want to get back in to the swing of writing articles.

Statistics

• 178 Blog Posts (142 more than last year!)
• 7 Months of Ubuntu Linux
• 3 Kittens
• 1 Work Visa
• 1 full time job
• 1 Green card
• 1 Anniversary
• 1 New language learnt (properly - !JavaScript)

I hope everybodies year was as successful and entertaining as mine, lets see what 2006 brings shall we?

- Davey

As a follow-up to my full review of Komodo 3.5 Beta I decided to update it for Komodo 3.5.1, after the help of the Komodo team to help me work out why I had some of the issues I brought up.

Installation

This time around, I had no problems installing Komodo on Linux or !WinXP, I installed the license after the application as instructed and it just worked. Great

Editing Files

!JavaScript

Although I was told the ability to resolve !JavaScript symbols within the current file is available through ctrl+space (after typing the start of the symbol), I was unable to get to it to work.

However, just having the object tree, code folding, syntax highlighting and error detection is a major plus.

PHP 5

In my last review, I stated that support for PHP 5 was missing, I was wrong (there, I said it! ).

The actual problem is that Komodo isn't able to work by default with PHP 5.1. The reason for this, is that it is designed to look for and load $installdir/lib/support/php/phpfunctips-$version.txt. Where $version is e.g. 5.0 not 5.0.x, the x is disregarded.

The simple solution to this, is to copy phpfunctips-5.0.txt to phpfunctips-5.1.txt - though it is only current with 5.0 I believe. Its much better than being stuck with 4.x

Furthermore, and more importantly IMO, Code Completion works, the reason it did not previously, as was finally found out through a bunch of Q&A e-mail sessions with Jeff Griffiths from !ActiveState, was my character set was broken. I had a Copyright symbol (©) in a UTF-8 document, and I was using the ISO-8859-1 character for it (apparently, I had copied the comment it was in, from an ISO-8859-1 document to a UTF-8 document in Zend Studio).

I found this problem out by trying to copy the contents of the file into a new file and then saving - at which point it fussed about the bad characters and requested to replace the bad characters with '?'.

I personally would rather have this check done on file load, and be prompted to replace the bad characters replaced with a ? or some other symbol, or to change the charset. Oddly enough I double checked the file in Zend Studio, the character is just not there, which is even worse behavior!

Conclusions

I am definately considering using Komodo for my non-PHP editing tasks, though I wish the CVS/SVN support was in the personal edition

I will never touch Eclipse again, I'm sure. The support for languages in Komodo is more than I need, and the resources footprint is far friendlier.

I would highly recommend checking this out if you're doing more than PHP on a regular basis and haven't yet purchased Zend Studio, Komodo 3.5.1 is now out for Linux, Windows and OS X. For those who do have Zend Studio, be sure to check it out as an editor for non-PHP tasks, though I still feel Zend Studio has the edge where PHP itself is concerned.

- Davey

There seems to be a recent fascination going around about re-creating the desktop online.

Unfortunately, most people are trying to model these desktop recreations after Windows circa 1995 UIs. They claim that this makes their application usable, and friendly.

Familiarity is a very weak form of accessibility. It requires pre-existing knowledge, which may or may not be available to the user.

One major problem - and I'm sure most of you will agree - with this; Microsoft isn't all that good at accessibility.

Unfortunately, the alternatives (OSX, Gnome, KDE) aren't much better, though OSX is IMO the one with the best usability, its still not perfect.

A common theme in these desktop-recreations is to place the "Start" menu at the top. I'm sure a reason for this is that its a lot easier to constrain a page at the top, and there are various IE bugs which make it hard to force an element to stay at the bottom. Especially, if you play with the font sizes.

This completely defeats the familiarity aspect, all of a sudden, its backwards. OK, so OSX users and often Gnome users like myself will have their "Start" menu at the top (yes, I'm aware thats not the name are fine with it but you've just made it awkward for upto 80-90% of your audience.

Please, people, lets try to move on, we have so much more flexibility with the web UI, lets try to better past attempts and create something which whilst it doesn't have that "familiarity" element, is immediately intuitive.

A common excuse for sticking with the "Windows" interface, is "Its what the (client|boss) wants". Well, it is your job to educate your client, or your boss. They will appreciate it at the end of the end of the day - I promise.

Yes, its some times hard to get through to these people, thats why you need to bring up these arguments I just made, and you need to provide alternative concepts. If you're using CSS-P properly, making the changes won't require too much work. And if it does take some work, think about your final outcome - a great Resume, or a very happy boss (which could lead to a raise, promotion and/or bonus

Don't let those with less knowledge dictate where the web goes, they are hiring you because you know what you're doing, don't blindly follow just to please, it will always backfire ("Oh, that was made by the previous guy, yes, its crap, we need you to change it")

- Davey