Samuel Birch is the developper of some nice scripts and webdesign ideas like the imagemenu, published on his website phatfusion.net.
Author Archives: Marco Barnig
Google Website Optimizer
Google recently launched the Website Optimizer, a tool that can help you improve the effectiveness of your website by allowing you to test different versions of your site content to determine what will best attract users.
Website Optimizer uses two types of testing: A/B testing and multivariate testing.
An A/B experiment allows you to test the performance of two (or more!) entirely different versions of a page. You can change the content of a page, alter the look and feel, or move around the layout of your alternate pages; there’s plenty of design freedom with A/B testing.
Multivariate tests, on the other hand, allow you to test multiple variables — in this case, sections of a page — simultaneously. For example, you could identify the headline, image, and promo text as parts of your page you’d like to improve, and try out three different versions of each one.
A/B experiments are the simpler version of testing with Website Optimizer. Experiments can run on many webpages of a website.
Website Optimizer needs a Google Analytics Account, because it uses the same powerful tracking technology that is used in Google Analytics to collect experiment data. The performance of experiment pages won’t be impacted, with the possible exception of the very first page-load after you have added the tracking code.
A quick start guide is available at the Google site. Different test strategies are possible (time on page, conversions, clicking rates, landing pages, …).
Website Optimizer scripts will fail W3C validation because they meet browser requirements which do not strictly adhere to the W3C guidelines. There is no workaround available for these errors.
Understanding Google Analytics
During the creation of a new website, I was able to have a closer look on the internal working of Google Analytics. Four computers were used to test the first layout of the new website connected to three different IP adresses. The setup was the following :
- PC 1 : Windows XP, screen resolution 1024 x 768, three browsers : IE 8.0.6001, Firefox 3.08, Chrome 1.0.154.153
- PC 2 : Windows XP, screen resolution 1024 x 768, three browsers : IE, Firefox, Chrome
- PC 3 : Windows XP, screen resolution 1280 x 1024, browser IE
- PC 4 : Windows Vista, screen resolution 1280 x 800, browser IE
Google Analytics reported 18 visits from 8 unique visitors during two days, which is correct if every used browser is considered as a unique visitor (use of cookies). The reported screen resolutions, operating systems, browser versions and flash versions were correct.
The unique pageviews can not be higher than the total number of unique visitors, which was the case. The 18 visits were segmented in 8 visits from new visitors and 10 visits from returning customers.

Customer loyalty
In the visitor loyalty report, the visits were indicated as :
- 1 times : 8
- 2 times : 5
- 3 times : 3
- 4 times : 1
- 5 times : 1
Web Design : block and inline elements
Block elements take up the full width of the page and are naturally displayed on a new line. Block elements can not be nested within inline elements.
Examples of block elements :
- <div>
- <h1> … <h6>
- <p>
- <table>
- <ul>, <ol>, <dl>
- <li>, <dt>, <dd>
- <form>
Inline elements only take up as much width as is required to display the element and will not force a new line. A width or height will not displayed on an inline element.
Examples of inline elements :
- <span>
- <img>
- <a>
- <strong>
- <em>
With the CSS command “display: inline” block elements can be altered to behave as inline elements. By floating these elements, it will give the appearance of an inline element but will also allow you to style them as if they were block elements. Therefore width and height can be applied to float’s successfully irrespective of whether the element has a default display of block or inline which is the reason why float’s are used so much in web design.
Target audience for an artgallery website
In marketing and advertising, a target audience, or target group is the primary group of people that something, usually an advertising campaign, is aimed at appealing to. The same rules apply for a website. The target audience for the website of an artgallery are art collectors, artists and art lovers.
To design a new website tailored to the preferences of this target audience, you have to identify the profile of these people, which is not an easy task. A first approach is to analyse the profile of the visitors of the old website, assuming that it reflects the typical profile of the target audience.
The following data sources can be explored to get useful informations :
- Google Analytics (if installed on the website)
- Alexa (web traffic metrics)
- Quantcast
The Google Analytics report for an artgallery website, operational since more than 3 years, provides data from more than 4 million pageviews, 3 million visits and 2.3 million unique visitors, which is quite impressive and representative.
The results from the analysis for the browser capabilities are the following:
Browser types:
- IE : 75,6%
- Firefox : 18,2%
- Safari : 5,1%
- others: 1,1%
Operating Systems:
- Windows : 91,8%
- MacIntosh: 7,5%
- Linux: 0,5%
- others: 0,2%
Screen Resolutions:
- lower or equal 800 pixel horizontal: 7,2%
- 1024 x 768 pixel: 43,5%
- higher or equal 1152 pixel horizontal: 49,3%
Flash support:
- Flash 10: 20% (augmenté à 40% les trois derniers mois)
- Flash 9: 64,4%
- Flash 8 or less: 15,6%
Java support:
- yes: 98,3%
- no: 1,7%
Traffic sources :
- Direct traffic : 58,2%
- Refering sites : 40,1%
- Search Engines : 1,1%
Key search terms : women in art, artgallery, digital art
Number of recurring visitors (more than 9 visits in 3 years): 31.326
Number of very loyal visitors (more than 200 visits in 3 years) : 2.891
The Web Information Company Alexa provides the following metrics (Demographics) for the website:
- Age : Relative to the general internet population, 55-64 year olds are greatly over-represente
- Gender : Relative to the general internet population, Females are over-represente
- Education : Relative to the general internet population, people who went to graduate school are over-represented
- Location : Relative to the general internet population, people browsing from home are over-represented
- Key search terms : video, alice, women in art, digital art
Quantcast provides following informations : This site reaches approximately 12,614 U.S. monthly people. The site attracts a 50+, somewhat wealthy, rather female, more educated group.
Web Design : layout’s with rounded corners
last update : 18 August 2011
In the past it was impossible to create any kind of curved shapes using pure HTML. There were three methods to generate layouts with rounded corners :
- CSS
- Javascript
- Images
There were a lot of tutorials available on the net how to create rounded corners with CSS. A simple way to generate the CSS and HTML you need to create anti-aliased corners without using images or javascript was the online tool Spiffy Corners from Greg Johnson.
One of the first creators of CSS layouts with rounded corners was Alessandro Fulciniti. He improved his Nifty Corners technique by adding Javascript.
Both methods didn’t work as expected on elements with fixed height, even if a external wrapper was applied. For this reason I preferred the image method with four corner images. Douglas Livingston‘s demo-page show a simple way how to do it (creative commons license). Another useful tutorial about rounded corners has been published by Trenton Moss, complete overviews of techniques are available at CSSjuice and Smileycat.
With HTML 5 and CSS3, the described techniques have become obsolete. CSS3 supports native rounded corners.
Flexible Web Design: web-page width
Last update : June 24, 2014
Web-page layouts can be grouped into four categories based on how their width is set: fixed-width, fluid (or liquid), elastic and hybrid. Creating web pages without fixed-width is considered as flexible web design.
The width of the overall layout of a fixed-width design is set to a value in pixels that’s decided by the designer, based on common screen resolutions, such as 800 by 600 or 1024 by 768. A recommended page width for 1024 by 768 pixel screens is 960, which is divisible by many numbers (2, 3, 4, 5, 6 and multiples). Fixed-width designs are rigid: they don’t change size based on any variations in the user’s setup. Because some people don’t browse with their browser window maximized, the screen resolution doesn’t match the browser window width all the time. Fixed-width designs are always going to result in some segment of the audience seeing a design that is either too wide for their windows (necessitating the dreaded horizontal scrolling) or too narrow (leaving oceans of space on one or both sides of the layout). Many people get almost as distressed about “wasted space” in their browser as they do about horizontal scrolling.
Fluid layouts, also known as liquid layouts, change in width based on the size of the user’s viewport (the viewable area of a page in the user’s device). If fluid layouts don’t have a width assigned with CSS, they will fill up the user’s viewport no matter how big or small it is. If a designer does assign a width to a fluid layout, it will be measured in percentages, not in pixels. The percentage refers to the portion of the viewport it takes up. When a fluid layout changes in size, all of the content within it has to shift around on the page to fill up the space. Horizontal scrollbars are the sworn enemy of readability. After all, scrolling continually back and forth to read across several lines of text does not make for the most enjoyable reading experience. With a liquid layout, horizontal scrollbars almost never happen, because users can size the window however they like to produce lines of text that they find the most comfortable to read and understand.
Elastic layouts change in width based on the text size set in the user’s browsing device. A user who has set a larger default text size will see a page where not only is the text bigger, but the entire layout is bigger proportionally than that seen by people with the default text size. Like fixed-width layouts, elastic layouts always have a width assigned to them, but that width is set in a unit of measurement called an em. One em is equal to the font height, which in turn equals roughly two characters in width, since most characters aren’t nearly as wide as they are tall. Elastic layouts are the rarest type of layout because before CSS became usable for page layout, they were simply impossible to create with tables. In many browsers, you can use the zoom feature to make all pages act like elastic layouts, which reduce the interest to use real elastic layouts.
Most web layouts are built using the idea of columns, whether or not the columns are explicitly visible in the design. Each column can have its own unit of measurement and be thought of individually as fixed-width, fluid or elastic. Mix them together, and you’ve got a hybrid layout. For example, a common type of hybrid layout has a fixed-width sidebar with a liquid main content area.
A resolution-dependent layout is a page that uses JavaScript to switch the CSS and thus the layout of the elements on the page by detecting browser window size. The CSS properties min-width and max-width allow you to set limits on how far a flexible layout will expand or contract.
More informations are available at the Peachpit website and at baekdal.com.
Newsletter analysis with spam check tools
To be sure that a newsletter with innocent words and formatting get not trapped in people’s or ISP’s spam filters, it’s useful to check it with an appropriated tool. Site Build It offers a free SpamCheck tool by form on the web or by e-mail.
I use the free MailingCheck software (version 104) from eDisplay to analyse my newsletters before sending them to the recipients. The test-email file has to be in .eml or .htm format. MailingCheck, as most spam check tools, use the open-source Apache SpamAssassin Project.
Another tool used by ISP’s to prevent spam are blacklists. A useful guide about this topic is available at the SpamLinks website. Blacklists or blocklists are lists of IP addresses, domain names, email addresses or content of the headers or the body, or some combination of these different types, that can be used to help identify spam. A special subset of IP address and domain name lists exist which can be queried using DNS, which are called DNS Blackhole Lists or DNSBLs. Blacklists can be unverified and cause “collateral damage”; their criteria for listing may not be clear.
I use the free Anti-Spam Blacklist Thing software from BlacklistPatrol/SpamButcher.

Antispam
Blacklist Patrol offers also a paid service to proactively monitor major anti-spam blacklists to see if an email server has been listed on any of them.
More informations about spam and blacklists can be found on wikipedia.
GWT-Ext
GWT-Ext is a powerful widget library that provides rich widgets like Grid with sort, paging and filtering, Tree’s with Drag & Drop support, highly customizable ComboBoxes, Tab Panels, Menus & Toolbars, Dialogs, Forms and a lot more right out of the box with a powerful and easy to use API. It uses GWT and Ext. GWT-Ext has a LGPL license option.
NetBeans : another IDE
“What IDE is the best : Eclipse or NetBeans ?” is an often discussed topic among developpers.
NetBeans is a free, open-source Integrated Development Environment for software developers with all the tools you need to create professional desktop, enterprise, web, and mobile applications with the Java language, C/C++, and even dynamic languages such as PHP, JavaScript, Groovy, and Ruby. The actual version of NetBeans is 6.5.