Tiny Pixel Blog

RadiantCMS as a Hosted Service

Wed, 12 Nov 2008 03:41:14 GMT

As of late I’ve been working on Bigsteepl, a simple light-weight web development service for small churches. It has really got me thinking about what it would take to run RadiantCMS as a hosted service.

Multiple sites, One Radiant Instance

There are a couple extensions that accomplish multi-site functionality. Unfortunately they were built with single client usage in mind (with the small exception of the vhost extension). In other words, they offer multiple websites under a single ownership. What about one owner per website?

Accomplishing one owner per website in a single radiant install would mean scoping all created content to a specific site. That also includes content created by extensions such as paperclipped or page_attachments. At the moment there just isn’t an easy way for extension developers to do this.

A Simple Solution

Attaching content to specific site is very simple in concept and would only affect form-related view files and add an additional column to a database table. We can’t forget to add belongs_to :site in whatever model your content belongs to. Pretty simple ya?

Here is how I see it going down.



# Your Extension Model

YourExtensionClass &gt; Page
  belongs_to :site

  # Your methods below...
end



# Your view code
-if multisite_enabled?
  %input{:type =&gt; "hidden" :name =&gt; "site_id" :id =&gt; "site_id" /&gt;

Then you run some migrations to add the site_id database column and boom, you’ve got scoped content. The above code is rather spares and barely says anything at all. But I think the idea behind it is clear. I will definitely be hacking on some extensions this weekend to see how the mileage varies on this approach.

Users are not complicated

Tue, 11 Nov 2008 15:11:22 GMT

Building usable web sites and applications doesn’t have to be rocket surgery. No matter how many features you offer, users will always look to the simplest way to get things done.

More Simple Features vs. Simplicity

The fact that a feature is simple doesn’t open the door to implement more simple features. More features always begets increased complexity no matter how simple those features may be. A perfect example of this is Microsoft Word. It has (God only knows) enough features to build 10 other word processing products. I’m willing to bet their users only see and use maybe 40% of those features.

Let’s pull out Pages by Apple Inc. that comes with their iWork productivity bundle. Pages has a lot of the same features as MS Word, minus the other 60% that users don’t or never will see or use. Often I hear users say Pages will never compare to MS Word on feature comparison alone. Personally I think it’s great. Your features should be solid and get things done.

Give em’ the keys to car

In web applications I often see features that only lead to another feature. Sometimes this is necessary, but never should it be the first resort. If you implement a feature and it’s important, don’t waste a user’s time by hiding it behind another seemingly “simpler” guiding feature.

An interface should drive the user to get things done. From the get go, users are pretty good at this already. And “trying” to make it “easier” by adding more simple features will only ever make getting things done in your website or app all the more complex.

Conclusion

Assuming the user is an idiot is the worst assumption. These assumptions usually lead to feature bloat and a serious obsession to over-complicate something that should not be a feature at all or should have been kept simple.

Supporting :hover in IE6

Mon, 03 Nov 2008 14:48:36 GMT

If you’ve ever built websites professionally, you understand a certain pain known as Internet Explorer. One specific pain is supporting :hover in IE6.

The best way to do this is using the code below.



input, button {
  _azimuth: expression(
    this.onmousedown = this.onmousedown || new Function("this.className += ' active'"),
    this.onmouseup = this.onmouseup || new Function("this.className = this.className.replace('active','')"), 
    this.onmouseenter = this.onmouseenter || new Function("this.className += ' hover'"), 
    this.onmouseleave = this.onmouseleave || new Function("this.className = this.className.replace('hover','')"), 'inherit');
}

Essentially the code above adds class selectors to an element on the fly using IE-specific javascript. For some reason, Microsoft thought it would be awesome if CSS supported its own subset of the javascript langauge. This great cause I can now support hover on any element in IE6. But the only catch is that you now have to add .hover, .active, etc ... styles in your CSS as opposed to using :hover, :active, etc...

Pretty friggin` sweet, eh?

Drinking Your Own Design Koolaid

Sun, 02 Nov 2008 15:49:39 GMT

Design requires a great deal of attention to quality, detail and most importantly, execution. A designer should never lose sight of quality for the sake of quantity. However, reality doesn’t really work that way.

In web design, you deal heavily with deadlines and balancing quantity with quality. It really does take a bracing shot of ignorance in order to not lock yourself into ruts of perfectionism just to get things done.

Koolaid, not for everyone

If there isn’t one thing I’ve learned, it’s knowing that not everyone will like what you design. It doesn’t matter how small or large a design might be or even how good or bad it is. There is always some person or group that will dislike what you’ve presented. As a result, you have two decisions to make. You either deal or leave.

Up until now, I always just dealt and adjusted as needed. Design for me has now evolved into a more opinionated creature that wants to do everything well the first time. For me that means beautiful code and design from start to finish.

Getting over the ones that were lost

Not everyone cares about beautiful design or even the code that goes with it (especially clients). Sometimes folks just want something up and running yesterday, without a care of how its done. Time and quality are a constant balancing act and sometimes quality just needs to take the driver seat. Actually, it ALWAYS needs to take the driver seat.

Conclusion

Do it right the first time, even if it takes a little longer. Always re-align areas that can be better and never use time as an excuse not to do something right.

Brain Malfunction. Thanks, TV!

Sat, 18 Oct 2008 22:34:57 GMT

Lately I’ve been watching ridiculous amounts of television. There are a few shows I follow regularly and then there are the shows I watch when no networks are playing the shows I want. The later is what makes me stupid. It causes serious issues in my verbal communication and overall mental capacity to be intelligent. Ugh. I need brain food!

Inline Hacking for IE

Tue, 14 Oct 2008 10:09:31 GMT

Hacking Internet Explorer in CSS is a common practice when it is required that appearances remain virtually the same across mainstream browsers. And by mainstream I mean, IE6 and IE7. Let me show you how to retain the most flexibility using inline hacks.

If you do a google search on inline hacks that target specific browsers you will most likely find posts about targeting IE6 only and IE6+7 only. However, there seems to be a lack of help in the area of IE7 only. There are various comment-based hacks, but those are dirty. While working for Headmix I discovered (I’m sure someone else has as well) a new exploit in IE7.

IE7’s rendering engine isn’t all that different from version 6. IE7 was in a lot of ways, a bug fix release. Because this is the case, we see some things that render the same between version 6 & 7, while others render with only a slight difference. So in order to fix this I use inline hacks that target only one version.


#content {
  >width: 120px; /* Targets IE7 Only */
  _width: 160px; /* Targets IE6 Only */
  *width: 200px; /* Targets IE6 and IE7 */
}

As you can see, we have a very flexible, clean and unobtrusive approach to IE hacking. I find inline hacking more maintainable than the separate stylesheets w/ conditional comments approach. You rid yourself of having to memorize various CSS style attributes and declarations. Try it out. You’ll love it, I promise.

Moving to a New Publishing Format

Sat, 11 Oct 2008 21:46:59 GMT

Just wanted to let folks know that I am in the process of moving tiny pixel to a new publishing process. After upgrading to v0.6.9 of RadiantCMS, I have the ability to do more with my content. Furthermore, this current blog will lose the ability to display images in posts for support of a new asset management tool for RadiantCMS. I will probably keep the feed url in the same place, so no worries there.

Sample the Benefits

Tue, 07 Oct 2008 09:56:00 GMT

I’ve been thinking a lot about Bigsteepl as of late. All this time I’ve been trying to figure out to offer a big solution for a very low price. It dawned on me that this will never work. I have to sample the benefits first.

Starting out I was convinced that offering the full enchilada for a less-than-feasible price was the way to go. Then I factored in how much work it would actually take to not only design and deploy a website, but also support it on a daily basis. The result is quite daunting and edging on unrealistic for the price I’d originally set.

Get them in the door

The thing about small churches is that they’re usually not very web-savvy. Their focus is more on maintaining growth and keeping people in their congregation inspired to give and commune with others. Building a website is probably the last on their list of priorities.

Bigsteepl should offer a low-commitment solution that is easy to maintain on both ends (developer and client). The cost of something like this should be in the range of 50 – 80 dollars a month for the church. If a church’s needs grow to large then they can upgrade and build something custom that fits their growing needs.

Conclusion

Small churches want something that’s easy to sign-up for, easy to maintain on a regular basis and doesn’t make their church look less-than-quality. That’s really the bottom line. I don’t think, as a small church, it’s about trying to look bigger than you are. I think it’s really about getting information to people who need it. Small churches generally grow really fast in the first 5 years and keeping the congregation abreast of all the things and events going on is critical.

User Assumptions

Sat, 04 Oct 2008 18:25:00 GMT

It doesn’t matter what kind of company you work for, there are always heavy assumptions made about how a user, consumer or client will react to change and these assumptions often decide what goes in or out for a product. We all assume things every now and again, but the result never turns out as expected. We’re often wrong as a matter of fact.

Usual Expectations

Every product has a solid list of unchanging expectations regarding their use. Using cars as an example; we expect that drivers will use the brake pedal to stop and the accelerator pedal to move forward. During the QA process we document all other cases that would vary a user’s expectations. In our car example, the driver could be disabled or have physical discrepancy that prevents them from using something the right way.

Web-based products also retain a (long) list of user expectations. However, the web is different in that we’re dealing with an enormous amount of meta details such as operating system platform, browser platform, computer hardware limitations and of course the illusive world of user behavior in a web-based context.

We have all these details, variant use cases and ever-changing behaviors to work against. So how do we solve what seems to be an unsolvable platform of problems and barriers? Enter the 3 rules of UX problem solving.

1st: Stop assuming

Every product you use is based on a number of assumptions that have either been derived from studies, tests and or actual real world usage. Not once will you ever see a product succeed based on pure assumption.

2nd: Evidence

Sometimes an assumption will be based on preconceived notions established on nothing or a minority of weightless reactions. If you’re going to assume, bring evidence of what problems you’re trying to solve. Sometimes we (as in developers) create more problems than we’re discovering.

3rd: Solve problems

The biggest problem with solving problems is that sometimes it changes things. Changing things displaces a user’s expectation and frankly there is nothing anyone can do about it. Maybe you can just not change it, but where is that getting you?

Conclusion

User experience is NOT a science by any means. Users aren’t complicated and solving their problems isn’t difficult. Although when you assume, it then becomes a science, user behavior appears complicated and their problems are very very difficult to solve. Save yourself the headaches and stop assuming, bring evidence of real problems and then solve them.

I Need a Hobby. Xbox? Halo 3? Call of Duty 4?

Tue, 16 Sep 2008 19:47:00 GMT

For 8 hours a day, I hack on user interfaces for the web. Before web design became a career, it was actually a hobby. Like clock work, when you merge a hobby into a laborious career, you loose all sense of curiosity, wonderment and discovery. The likelihood of getting that back is very low at this point. Which begs the question. What else can I do as a hobby?

*Gaming, an unlikely obsession* Ever since I was little, owning video game consoles were as much a dream as ruling the earth. Nor my mom or grandmother had the means to afford such luxuries during my childhood. Eventually I'd get hand-me down systems such as the Super Nintendo or Sega Genesis. However, by the time these systems made it my way, consoles like Sony Playstation and Nintendo 64 were all the rage. Needless to say, my attraction to video games is comparable to that of a volcano getting ready to burst. As the holiday season approaches, I think I owe it to myself to afford at least one luxury. With the Microsoft Xbox pouring in at $200 bucks a pop, I think this is the perfect time to get one. Over the next couple days I'm going to put together some proposals to present to Mrs. Jones and see what happens. _Wish me luck!_

Ray: An Extension for Managing Extensions

Sun, 14 Sep 2008 16:08:00 GMT

Ray is an extension for RadiantCMS, a Rails-based content management system. Ray lets you manage extensions quite easily. If managing a Radiant installation gets any easier, I might start a business based on RadiantCMS. At any rate, Ray is one of the more powerful extensions to hit the streets.

Click here to visit website

Rails Discovery: URL Routing

Sun, 14 Sep 2008 15:32:00 GMT

I recently picked up a book called, “The Rails Way”. I’m about 80 pages in and it has been a joy so far. I’m learning a lot about how rails works under the hood.

I recently finished up a 20+ page chapter called, “Routes”. Routes are how rails applications translate incoming and outgoing HTTP requests. Needless to say, I was never clear on how the URL routing system actually worked. Obie Fernendez explains it very clearly.

Before this book, Routing seemed like a useless bit of the entire ROR stack. In the past it had never been clear just how critical routing can be to a given rails application. Without routes, your app would simply be a body without ears or lacking any way to translate incoming and outgoing communication. I was absolutely shocked that I never realized this.

Routes also provide a philosophy of web application design, a philosophy of which I was not aware of before reading “The Rails Way”. In the next chapter I will be learning about REST and Resources which is exciting in my opinion.

I’m going to be writing a lot about the things I discover about rails. Who knows? Maybe I’ll build an app in honor of my new found knowledge.

Internet Explorer is Coming Back

Sun, 14 Sep 2008 15:15:00 GMT

Microsoft has long stood as the browser of poor choice. Mozilla, Apple and Opera are much like the cool kids that preach their superiority over all the nerds, or in this case, Microsoft. The hay-day is coming to an end.

Call me a doom sayer, but Microsoft has been far too quiet since the release of Vista. That quietness can only lead to a few things. My primary hunch is the fact that MS will be competitive again. For years they dragged their feet out of arrogance as they are the software company of choice on most if not all user levels. Since it has completely ruined their reputation and destroyed their competitive edge in the software market.

In the case of Internet Explorer, we see that competitive edge making a comeback for MS. Ever since Internet Explorer 5.5, the development schedule had come to a complete crawl. Every year or so, we were lucky to see so much as a hotfix release. Lately, Microsoft has stepped it up tremendously. We saw Internet Explorer 7 hit the streets in October of 2006, then less than 2 years later Microsoft releases the first public beta of Internet Explorer 8.

It’s still early to say whether or not IE will regain all of it’s market share. However, we are going to see a lot of windows users growing more comfortable with a Microsoft-driven internet experience.

Rails Start-up Process: Pwnd

Tue, 09 Sep 2008 10:29:00 GMT

Starting a rails app that includes other daemon process which also require help from the command line can get quite hairy and annoying. The solution is to drop all these commands into a nice little shell script. Go here to see an example of what it would look like. Also don’t forget to drop that file in your /usr/local/bin or /opt/local/bin/. Then do a chmod +x your_file_name. Win.

My Wordle.

Tue, 09 Sep 2008 10:15:00 GMT

I stumbled upon this reading a friends blog. Thought it was neat so I tried it. (hat tip: Social Media Girl)