The new aesthetics of your site's URLs
A fashionista I'm certainly not, but it's probably a testament to how much of a geek I truly am that I would attach an aesthetic value to a URL. And apparently, I'm not alone.
Web development frameworks like Django and Ruby on Rails, motivated by the community recent head-over-heels love affair with the almighty permalink, heavily promote the use of elegant URLs. And enterprise-level platforms like Microsoft's ASP.NET and JavaServerPages also provide URL munging capabilities to morph a site's page addressing scheme to something more pleasing to the eye. And of course, there's the ultimate sin: including DLLs in your URLs.
It's now seen as passe to include file extensions in your URLs, thus making it immediately visible what platform underlies your application. (I'm sure the vendors are loving that.) Likewise, it's a development faux pas to use query string-resident values. Liberal uses of HTTP POST with sparing uses of GET where applicable is the preferred trend. My most popular blog post to date was on in which I proclaimed the arrival of the new in-thing for web development: killing the query string.
But there's still something to be said about proper directory structure within a web application, lending to the psychological impact a URL has on those users who bother to even notice them (assumedly, the vast minority). I applaud the efforts of Microsoft in years past and Google today to structure their sites with logical URLs (e.g., microsoft.com/word, mail.google.com). When laying out any site, this is a chief concern for me, and I did this with KUAM.COM (e.g., kuam.com/rss, kuam.com/decision2004, kuam.com/search, kuam.com/mobile). Still, when applying .NET path rewriting, I changed the structure of my news stories from http://www.kuam.com/news/story.asp?id=13090 to http://www.kuam.com/news/13090.aspx.
MSNBC.com hides the fact that it's running ASP.NET from it's revamped site, of which a new page addressing strategy was put into place. The URLs are all numerically-based, and don't display .ASPX, .ASHX or other file extensions that would give anything away. It's low-level security, at best. Still, I don't hold MSNBC.com as a premiere example of how to properly manage URLs because the format is completely uniform across it's numerous content areas. Looking at a random URL I just copied - http://www.msnbc.msn.com/id/3032076/ - even people that know the site really well one can't immediately discern whether the destination is a major area, an individual story, or an affiliate homepage (it's the site's Health subsection).
So let's bridge the two schools of thought and really capitalize on this trend with proper implementation, in the hopes of enjoying the best of both worlds. Let's leverage the programmatic abilities of major frameworks to create nice-looking URLs, but also create folder hierarchies that make sense.
Web development frameworks like Django and Ruby on Rails, motivated by the community recent head-over-heels love affair with the almighty permalink, heavily promote the use of elegant URLs. And enterprise-level platforms like Microsoft's ASP.NET and JavaServerPages also provide URL munging capabilities to morph a site's page addressing scheme to something more pleasing to the eye. And of course, there's the ultimate sin: including DLLs in your URLs.
It's now seen as passe to include file extensions in your URLs, thus making it immediately visible what platform underlies your application. (I'm sure the vendors are loving that.) Likewise, it's a development faux pas to use query string-resident values. Liberal uses of HTTP POST with sparing uses of GET where applicable is the preferred trend. My most popular blog post to date was on in which I proclaimed the arrival of the new in-thing for web development: killing the query string.
But there's still something to be said about proper directory structure within a web application, lending to the psychological impact a URL has on those users who bother to even notice them (assumedly, the vast minority). I applaud the efforts of Microsoft in years past and Google today to structure their sites with logical URLs (e.g., microsoft.com/word, mail.google.com). When laying out any site, this is a chief concern for me, and I did this with KUAM.COM (e.g., kuam.com/rss, kuam.com/decision2004, kuam.com/search, kuam.com/mobile). Still, when applying .NET path rewriting, I changed the structure of my news stories from http://www.kuam.com/news/story.asp?id=13090 to http://www.kuam.com/news/13090.aspx.
MSNBC.com hides the fact that it's running ASP.NET from it's revamped site, of which a new page addressing strategy was put into place. The URLs are all numerically-based, and don't display .ASPX, .ASHX or other file extensions that would give anything away. It's low-level security, at best. Still, I don't hold MSNBC.com as a premiere example of how to properly manage URLs because the format is completely uniform across it's numerous content areas. Looking at a random URL I just copied - http://www.msnbc.msn.com/id/3032076/ - even people that know the site really well one can't immediately discern whether the destination is a major area, an individual story, or an affiliate homepage (it's the site's Health subsection).
So let's bridge the two schools of thought and really capitalize on this trend with proper implementation, in the hopes of enjoying the best of both worlds. Let's leverage the programmatic abilities of major frameworks to create nice-looking URLs, but also create folder hierarchies that make sense.
2 Comments:
At January 8, 2006 2:04 PM,
Anonymous said…
Apple -- naturally -- is the, like, champion of normal URLs that you can actually verbally tell to other people. It's like ingredients you can actually pronounce.
At January 8, 2006 3:33 PM,
Jason Salas said…
I agree...good call.
Post a Comment
Links to this post:
Create a Link
<< Home