The core message of Andy’s presentation was that other industries have long understood the importance of a positive user experience, and the Web can learn a lot from this.

His presentation gave lots of tips, with plenty of rich examples:

1. First impressions count: Hoteliers understand this, hence good hotels offer smiling greeters in the lobby, chocolates and written notes on your pillow.
2. Attention to detail gets noticed: Apple are a company that realize the difference that attention to detail can make — consider the effort that has been put into the packaging of their products, which results in customers actually photographing the unboxing of their iPod. From the sound of the door closing on a new BMW to the rubbish bins at Disney theme parks being themed, people notice this stuff.
3. Personalisation and customisation matters: Starbucks allows its customers to customise their coffees, and the Nintendo Wii and other games let users create their own characters. Second life takes this to a new level to accommodate people’s needs to assert their individuality. For web developers, something as simple as welcoming people by using their first name can have an enormously positive impact on their experience on your site.
4. Dripfeed your information: Andy reference computer games as a huge source of inspiration for him, and showed a video of a combat game that taught players how to perform tasks incrementally. By gradually releasing information about how to perform tasks in digestable chunks, you can avoid overwhelming them. As an example, Basecamp displays optional videos for new features, and the Yahoo! site utilised a lightbox to explain new features when they launched their redesign last year.
5. Take care of the mundane: Andy referenced the hospitality industry heavily when making this point — the doorman hailing you a cab, the waiter filling your glass when you don’t notice, and the barmaid remembering your drink are all great examples of what makes a positive user experience in one moment, which can carry the user through to a positive experience throughout their entire visit. Web analogies include taking care of unnecessary tasks for users by prefilling them with intelligent defaults.
6. Make it fun: Andy showed an example of a drink package that had the text “Stop looking at my bottom” on the underside of the container. Little touches like this obviously require the right context and a bit of thought to do properly, but when they work they can make your users smile for the rest of the day. Another example was a personalized email from MOO, which read “Hi, I’m little MOO, the software that will be managing your order…” A banking site might not get away with being this cute, but it’s definitely possibly to be playful without being unprofessional.
7. Feedback is important: Poker machines are feedback-heavy — they emit a “bling” sound, and the cleverly crafted “ding” of coin as it hits the metal tray causes other punters to take notice, sit down and try using a machine themselves. Likewise, a button always lights up in an elevator — if it doesn’t then you might wonder whether the lift is broken or not! Status bars and other indicators to let a user know where in the process they are presently located help manage expectations and keep the user engaged.
8. Recommendations are powerful: last.fm is a good example of how recommending music to a user is a terrific service that really adds value. If you can provide contextual information to your users that taps into the knowledge that the entire user base has created, then your users will pay attention.
9. Users love competitions: By offering a competitive element — whether that be through scores and a leaderboard, or just by letting people “collect” things — badges, blog posts etc — your userbase will have an additional reason to return to your site and engage with your data. Leaderboards can have a negative effect too, of course — the digg leaderboard stopped being a useful indicator when those users at the top started posting quality over quantity, in order to maintain their position on the leaderboard.

These experiences can be plotted as a user experience curve — an actual graph that, while subjective, shows a user’s positive and negative experience over time. Andy mentioned Maslow’s hierarchy of needs as applied to the Web. The items in the hierarchy were:

1. functional
2. reliable
3. usable
4. convenient
5. pleasurable
6. meaningful

I agree with Andy’s comment that it is very rare that a user experience on the Web meets those levels of experience at the top of the hierarchy (pleasurable, meaningful experiences). He encouraged attendees to look to other industries, and other areas in the offline world for inspiration to bring positive user experiences to the Web.

Three members of Team SitePoint are wandering around at the conference, so feel free to come up and say hi!

Photo credit: Dave Shea

This week saw the beginning of the Adobe Community Summit, a briefing event for Adobe Community Experts and user group managers. According to Aaron West’s Day 1 summary post, there are over 150 attendees this year, catching up on the latest in Adobe tech and giving feedback on the various community programs they are involved in. Aaron also revealed that Rachel Luxemburg was introduced as the new user group manager, replacing Ed Sullivan who many people involved in their local user groups would have had contact with over the past few years.

Adam Lehman, Platform Evangelist at Adobe, released his set of demo ColdFusion applications, covering the Ajax, Flex, PDF and presentation features of ColdFusion 8 (amongst others). The quick and dirty demos are designed to work with the standard cfartgallery sample database, an Apache Derby embedded database that gets installed with CF if you choose to include the samples, so if you’ve got a development server handy and you haven’t had a chance to try all the new features, check Adam’s demos out.

Also from the Adobe camp, briefly:

• Russell Brown has a controversial request for ColdFusion 9: drop the built-in Ajax and JavaScript libraries. Although it’s unlikely that Adobe would take action, his comments are an interesting read.
• Jim Priest is curious as to whether there are developers who are waiting for Adobe to release a ColdFusion IDE. Is that you? Leave a comment for Jim.
• Adobe have released a hotfix for the image tags and functions in ColdFusion 8.0.1, so if you were affected by any bugs in image manipulation, you’ll want to download this (hat tip to cfsearching).
• Stephen Moretti has posted his review of ColdFusion 8, nine months on… with a little help from some friends.

Open Source

OpenBlueDraon is still a hot topic. Neil Middleton thinks that Open BlueDragon makes CFML development easy again. Do you agree? While we’re talking about it, the official blog has been launched, and a VMWare “appliance” - built on CentOS with Tomcat, MySQL and Apache - has been released. Matt Woodward has posted about his new development setup running Open BlueDragon and ColdFusion 8 side-by-side and Per Kleven has written detailed instructions on how to get Open BlueDragon and Tomcat running in a manner which allows easy upgrades (hat tip to Mr Corfield). The MXUnit team has posted on how to setup and deploy OpenBD on JBoss and Tomcat for Windows.

SourceForge have opened nominations for their 2008 Community Choice Awards, and for the first time it is not limited to projects hosted on SourceForge. Got a favourite open source project? Go nominate it! I think it would be amazing to see a CFML-based project short-listed for this award.

The Daemon team have announced that they are considering changing the licence under which FarCry 5.0 is released. While nothing has been decided yet,  the debate rages on the FarCry Google Group. Jeff Coughlin has written an interesting article about the issue on Fusion Authority: What’s in a License Anyway? : FarCry Considers Moving to a Dual License Model.

Oscar Arevalo has announced the release of ColdBricks, a free and open source content management system written in CFML for portals and dashboards (hat tip to a very excited Serge Jaspers).

“ColdFusion Conference Season” has started

cf.Objective() is over - read my summary of what the blogosphere had to say about this event - next week is WebManiacs in Washington DC (May 19-23), Scotch on the Rocks in Edinborough, Scotland is on from June 4-6, then CFUnited is gearing up for their 10th anniversary event, June 18-21 in Washington DC. Sneaking in first however will be WebDU, held in Sydney, Australia on June 12-13, and I’m pleased to say I will be there covering all the happenings. Then, later this year, CFDevCon08 has been announced in Brighton, England in September, and Sean Corfield would like to know what you think he should present on.

Community

The ColdFusion book scene has been rather sparse of late - other than Ben Forta’s staple CFWACK volumes, as far as I can tell no new books have been released for ColdFusion 8. Niche publisher Pakt are looking to change all that, announcing that John Farrar’s ColdFusion 8 Developer Tutorial will be shipping in August this year. Described as an “intense guide”, the book covers the basics as well as some of the juicier CF8 features, so should be a useful addition to the CF developer’s bookshelf. Good on you Pakt and well done John!

• Ben Forta has a list of international web hosts that support ColdFusion, and he’s put out a call for people to update the list where it’s gotten out of date.
• Charlie Arehart has a long way to go to cover all the categories in his Tools and Resources directory, but this week he’s highlighted CFMAIL Repalcements/Enhancements, CFML Documentation and Help Tools and CFML Caching Tools.

Code cuts

• Jedi master Raymond Camden shows how to easily get the current directory from a CF script, and Dan Switzer has followed up with how to get the URL path, without using cgi-scoped variables.
• Chris Phillips was impressed at the cf.Objective conference by how Transfer ORM handles caching, using the Java SoftReference class, and has shared some information on how it works.
• Dan Vega runs through an example of how to use multiple filters in CFDIRECTORY as a follow up to Ben Nadel’s two posts on the same issue
• Doug Hughes explains how search engine friendly URLs work in Model-Glue 3
• Barney Boisvert has updated his Amazon S3 CFC
• Gary Gilbert talks about speed and multiple CFQUERY inserts with some interesting test results

So that’s all, folks. Got a tip for me? Email kay at smoljak dot com, tag a link for:kay.smoljak in del.icio.us, or leave a comment.

Updated timezone support

Timezones will finally become a first class citizen in Rails. You will be able to set the timezone, and all subsequent time calls will be with in that zone.

Time.zone = "Australia/Perth"
Time.zone.now # will return something like Wed, 24 May 2008 22:56:00 WST +08:00

Better Gem dependencies

If you rely on gems in any of your projects (and why wouldn’t you? Code re-use and all that), you may have come across the pain of gem versioning. Rails 2.1 will allow you to stipulate what versions of each gem you need.

config.gem "chronic", :version => '0.2.3'

It also adds a rake task that will automatically install the right gems for you.

rake gems:unpack

Improved caching

Previously, the only caching options Rails developers had was based around file fragments, which is fine for single server setups - but scaling that up to multiple servers could cause synchronisation issues, causing your cache to go stale. To combat this, you can now use a number of other shared caching systems including a memory store, drb store and a mem-cache store.

All very exciting stuff, and well worth a checking out. If you can download the beta gems by running the following from your trusty terminal window:

gem install rails –source http://gems.rubyonrails.com/

That’s right, the entire 278 page book, yours to keep, forever!

The Photoshop Anthology: 101 Web Design Tips, Tricks & Techniques is the ultimate Photoshop compendium for web designers.

It’s brimming with tried and tested real-world Photoshop solutions that will add impact to your next web design project. If you’ve ever been stuck for inspiration, have puzzled over just how to create a shiny aqua-style button, or wanted to create that seamlessly tiling background image you saw on a site recently, you need to download this book.

Download it now!

This book is free to download thanks to the generous support of 99designs. If you’re looking for a place to give your new-found Photoshop skills a run, why not earn some extra cash along the way?

The team at 99designs are giving away a shiny new MacBook for the best logo design or web design, so be sure to check them out.

When is a table not a table?

When its data isn’t really tabular; in other words, when the data it represents isn’t really two-dimensional. What we have here is visually two-dimensional, hence at first glance it might seem reasonable to represent it as tabular data; but the data itself isn’t really two-dimensional at all, it’s linear.

A two-dimensional data set has two axes, each representing a different range or set of values, so we should be able to plot axes against the data and be able to make meaningful cross-references. But as the illustration below shows, we can’t:

Sure we can plot those lines and create that reference point, but they don’t mean anything, because the x and y axes both represent the same scale (skill level).

Before we can implement the design we need to decide how the data should be marked up, and that’s why it’s so crucial to determine the inate structure of the data first, in order to know what kind of element semantics to use. And when considering semantics, the bottom line I always use is to think about how the data will be comprehended by those using a screenreader, in which there are no visual cues, only semantics.

Plotting that data in a table, it seemed immediately obvious to me that this was the wrong way of structuring it:

How coherent is the structure — do we gain a progression of data by traversing across the columns? Does the ordering of rows and columns mean something explicit such that a single point within the table can be cross-referenced? Or is it merely that the data happens to fall into columns because we’re still thinking of it visually? I think it’s the latter, and that presenting the data in this way would be innacurate and confusing.

So now lets represent that same data linearly:

And immediately it seems far simpler and more coherent, and much much easier to understand, whether you’re looking at, it or hearing it read linearly.

The conclusion I came to then was that the data is a simple linear structure, and is only presented in a visually two-dimensional way because this makes it easier to see patterns at a glance.

Why is the CSS for this so hairy?

Well it isn’t particularly difficult or complicated, but neither is it desparately simple. The major difficulty with this, as with all precise layouts, is creating it so that it scales with font and window size. This requirement means that all sizes and positions must be relative, and all boxes must be flexible, so that the layout remains consistent and accurate regardless of the user’s settings:

Another challenge that crept in is the fact that the colored scale across the top is of visual interest only — this information has no meaning when the content is linearised, therefore it needed to be created in a way that only impacts on the visual layout and isn’t present as content.

The explantory text down the right-hand side was similarly tricky to position. In terms of the markup, this content comes immediately after each book title; but visually it’s way off to the right. The CSS for this had not only to create this layout, but do so in a way that ensured solid flexibility — so that the both height of the book box and the height of the text would affect the overall item height co-dependently (ie. altering either of them affects them both).

Have a look at the stylesheet if you’d like to poke around.

And of course, the endless gradients and rounded-corners our trendy designer insists on using makes everything far more complicated … I guess I’m just a glass is twice as big as it needs to be kinda guy!

But, what is the correct type value for JavaScript? The other reason Mr Crockford provides for dropping the attribute is that that the value most people use, text/javascript, is wrong! It’s obsolete, according to rfc4329. This is also confirmed by Anne van Kesteren who has covered this issue already (way back in May of 2006 - yes Anne is one of the gurus and I am not), as well as the SitePoint HTML Reference.

The correct type value for JavaScript is actually application/javascript. But wouldn’t you know it, Internet Explorer will not execute the code if you use that type value. So here’s a hell of a situation, to produce a valid HTML or XHTML document with script tags, we are required to use a type attribute, but to get it to function across all browsers we have to use an incorrect value.

Just goes to show, the web is really still only an alpha version. Can’t wait for the first beta!

1. Client call during work hours

Pick up the phone. Using voice mail has its place after hours, but during work hours: pick up the phone. An answering machine says to a client – I’m not here and that (in a client’s mind) can translated to, you being unreliable and you don’t care about their business. Think of it this way: a client takes an extra (perceived) risk using you — a freelancer — as you are not as ‘safe’ to use as a company. Therefore, it is vital that the client ‘feels the love’ and can speak to you during business hours. This helps to reassure them that their business is safe in your hands and you are not some ‘fly by night’ amateur.

2. Personal calls during Work hours

It is pretty standard for your mates or family to call you during work hours at home, and it is an issue you will need to deal with sooner, rather than later. Friends and family need a bit of ‘education’ on your work. During normal working hours you are working — full stop. A polite can’t talk now, flat out busy working – can I call you later tonight?, will usually do the trick. It doesn’t hurt to drop a subtle reminder about, how busy you are during work hours and that you can talk more in the evenings.

3. Client call out of work hours

It is not unusual to get client calls outside of normal business hours (I’ve had a late night call on a bank holiday). This is generally unacceptable, unless you provide 24/7 support.

The only acceptable time to call outside of normal working hours (IMO) is if there is a critical business issue i.e. Help. My site has been replaced by a pill pushing porn site or Help. My site is down -– those are critical issues. I can’t remember how to create a link or I’ve forgotten my password, are not critical issues.

So how do you deal with it?

1. My strategy is a simple one: if it isn’t a critical issue, then explain that they have called outside your normal business hours and that you will look into it the next working day. Be polite, but firm. Explain that if they want support outside of business hours, then there is an extra charge. In 99% of cases, the client will respect your privacy won’t do it again.
2. The next working day, address whatever the problem was, and then let the client know by email what you’ve done and restate your support hours.

As you might already have gathered it is all about communication. Let people know when and what it is acceptable to call you about. TTFN.

It seems most delegates have arrived back home and updated their blogs, so I figure it’s time to see what they all learnt over the three days…

Brian Rinaldi wrote a comprehensive wrap-up for Fusion Authority. He noted that as the CF community continues to mature, developers are looking to the more advanced tools available to them to solve problems, namely Java, and this was evident in a number of session topics. He also noted that Flex was a hot topic, with so many sessions that a developer could spend the entire conference focused solely on Flex. Brian also reviewed several sessions in detail: Selling Professional Development in a Hostile Shop by Terrence Ryan; Leveraging Code Generation by Brian Kotek; Mate Flex Framework by Laura Arguello; Transfer Caching Mechanisms by Mark Mandel; and Flex: No Frameworks Required by Maxim Porges.

Mark Mandel wrote up his highlights, which included workshops on Model-Glue, ColdSpring, and Terrence Ryan’s session on Selling Professional Development at a Hostile Shop. While he was there, Mark presented two sessions on Transfer ORM (now at version 1.0) which were well-received - watch out for an interview with Mark on SitePoint in the near future.  He noted that the ColdFusion 9 information that was revealed was nothing overly surprising, but welcome nonetheless. Oh, and Mark was the lucky winner of the Nintendo Wii!

Brian Meloche posted three daily summaries - Day 1, Day 2, and Day 3 - as well as an overall review, giving the conference an ‘A’ grade. Brian goes into quite a bit of detail about each of the sessions he attended, and makes note in particular of the Open BlueDragon initiative (a beta version was released during the conference) and covers what was discussed at the BF9 “Birds of a Feather” session. He was disappointed that more emphasis was not placed on the need for a dedicated ColdFusion IDE from Adobe.

A lot of material was published post-conference! Some more highlights include:

• Brian Kotek took away from the conference an impression that ColdFusion is at a crossroads. His thoughts on the future direction of the language are an interesting read.
• Terrence Ryan posted his conference postscript in easy to read bullet points - and created a cf.Objective() 2009 to-do list.
• Peter Bell summed up a whole stack of the presentations, noting the cf.Objective() “rocked”.
• Adam Lehman posted about the suggestion that cf.Objective() might expand to Europe and/or Australia.
• Free copies of the latest issue of the Fusion Authority Quarterly Update were distributed to attendees.
• Andy Powell posted his cf.Objective() Flickr stream in a funky slideshow.

A number of presenters have provided session materials online - slides and code samples:

• Jason Delmore’s Developing ColdFusion Java Hybrid Applications
• Brian Meloche’s ColdFusion as a Different Type of “Glue”
• Phil Nacelli’s Leveraging Basic Design Patterns in ColdFusion
• Peter J Farrell’s What’s New in Mach-II 1.6 and Using ANT
• Raymond Camden’s Spry and Model-Glue sessions
• Peter Bell’s Software Product Line Presentation
• Brian Kotek’s Leveraging Code Generation
• John Mason’s Unit Testing and PCI-DSS
• Dan Wilson’s From Procedural to OO (slides linked from bottom of Chris Phillips’ session review)
• Terrence Ryan’s Selling Professional Development Techniques
• Andrew Powell’s Enterprise MVC with ColdFusion and Java

One factor that nearly every post mentioned was that it’s the people that make cf.Objective() so great. Steve Bryant resisted the urge to do some name-dropping, while Ben Nadel likened the experience to walking among giants. It’s certainly a recurring theme - as was the wish of many attendees to be back for cf.Objective() 2009!

Tracers and error handlers

Both of these functions produce a rather crude output though, so naturally people have written wrappers around them to remedy this. I think Harry’s pretty bluescreen was one of the first dedicated libraries I’ve seen. Xdebug spouts a similar output on error, although arguably not as pretty. Or blue.

What bluescreen is for debug_backtrace, krumo is for var_dump. Recently, FirePHP — building on Firebug — does a similar thing. FirePHP uses HTTP-headers to send data from server to client, which turns out to be very handy when dealing with non-HTML output (Eg. Ajax stuff). Because it builds on Firebug, it only works on Firefox, and in particular only on Firefox 2 (Another reason for Ubuntu-users to downgrade from Firefox 3).

Frameworks

Apart for these general general tracing tools, a couple of frameworks have their own, more or less specific, tools. Symfony’s Debug Toolbar is probably the most impressive (And in contrast, Zend Framework the least), even providing information about database-queries, which is very helpful for profiling.

Another framework-specific tool, which I’m quite impressed by, is the Theme developer module. It’s advertised as “Firebug for Drupal theming”, which is fairly descriptive. In case you haven’t used Drupal, it has a complex theming system for rendering pages and this tool makes it a lot easier to figure out which part is doing what. Even if you don’t intend on using Drupal, it’s fascinating to see how well polished it looks.

I think that most frameworks with an integrated presentation layer, could benefit hugely from something similar, although the specifics would obviously vary. One of my own projects, Konstrukt, has a tool for displaying debug-information about which controllers renders in which order. This can be helpful, because a given page will often have several controllers working in concert to produce the final response.

Another take on this, is Cake PHP’s debug messages. This extension adds very specific and descriptive error messages in various places, which almost works out like a tutorial for new users; Seems like a good idea.

Webgrind

I already mentioned Xdebug, which is indispensable for profiling PHP applications. Last week, Webgrind had its first released. It’s a HTML-based frontend for displaying profiling output, generated by Xdebug. There are already tools for the purpose; If you’re on Windows, WinCacheGrind and for Linux-based systems, kCacheGrind. However, I can see a few good cases for Webgrind;

First of all, a zero-install client is a bit easier to reach out for. For Mac users, it’s also the only option. The featureset is quite limited, but this turns out to be somewhat a benefit. kCacheGrind can be quite intimidating and for the purpose of profiling, Webgrind’s features seems fairly adequate.
The most important though, is that the logfile generated by Xdebug (Which is the data-format for the various grind’s) tend to get rather huge. Naturally, this makes it awkward to have server and client on separate machines — This is not a problem for Webgrind, since it runs at the server side.

As is the case with many markets, online and off, there exists a healthy competition between the owners of these two competing sites. And as is often the case with individuals who enjoy talking about the money they make, this translates into a bunch of trash talk, teasing and a fair amount of back-and-forthing. For the readers of the blogs, this is no doubt highly entertaining, and probably drives traffic to both sites in question.

This latest stunt caught my attention — a competition to build a blog from scratch in a designated niche in 30 days, and see who can sell it for the highest price on the SitePoint Marketplace:

Instead of seeing who can make the most money with the blogs, lets invest 30 days into developing, marketing and building the blog and then list them for sale in SitePoint and see who can get the highest successful winning bid on selling it? The person who can sell their niche blog for the highest amount in a NO RESERVE Sitepoint auction wins the game. Both sites will end up selling and the runner up who sells for less, transfers their earnings to the winner of the game.

Details surrounding the rules of the competition are still being fleshed out, but I’m watching this one with a keen eye! In fact, I think I’ll start taking bets around the office on who will be victorious.

Pick a team and follow the action (either at JohnCow.com or GarryConn.com).