1) Employees in tightly-controlled IT / enterprise environments
2) People with absolutely no concept of how to upgrade.

Group 1 will go away with time, but Group 2 will only go away when members of that group trash their machines for new ones. I think the only way to get members of this second group to upgrade is just to deny service: if XBox owners could play any game from XBox360 they wanted, even at reduced performance, they would probably be loathe to upgrade. But if upgrading is REQUIRED for any sort of functionality, they’ll probably go through the hassle of buying a new machine, plugging it in, and figuring out the new interface. These users won’t go out and download FF3… they’ll probably just go buy a new machine.

That said, simply cutting off IE6 use is a dangerous proposition, even for companies like 37Signals. Check out this complaint for an interesting take on the issue.

@jez - IE 5… wow… I can’t say I have a good solution. My suggestion would to lovingly nudge your users/clients to upgrade. 5.0 is ancient. :)

Notice the odd margin occuring on the right side of the buttons.

Working at a small to medium web dev company, I personally have given up on wrestling with IE6’s melodramatics. I spend that time on educating clients as to IE6’s limited support of features. I understand the stalemate of larger corporations, but a revolution needs to be started. I agree wholeheartedly with Jacob Souva. Functionality of sites should still be supported on IE6 (for a little while longer), but they won’t have all the bells and whistles of their big brother’s.

Rey - jQuery Project

People using office computers often cannot upgrade per IT restrictions. The problem is that these companies have internal sites developed to work exclusively on IE6, so until they upgrade, we will have to continue to support it. Anyone that has ever worked in a large corporation can attest that is never a timely process.

My guess would be all the intranets and internal applications that have been developed specifically for IE6.

If they forced an upgrade, they would be “breaking” a lot of web platforms that corporations have spent big money on.

It’s a catch-22 for MS, on one hand the longer they hold off fully supporting web standards the more market share they lose; on the other they have to support the legacy code to appease their corporate base that has invested big dollars in it.

I mention it because I work at an enormous company, on a laptop with Windows 2000 (which cannot support an upgrade of IE6!)

Our Vista upgrade project is scheduled to be complete in 2009.

@Rob: Go ahead & snicker.

@Michael Montgomery concerning corporate computers: This argument comes up a lot. I think the further we get from the launch of IE7 the more these companies will get complaints on why a site looks great at home, but at work they’re getting a second-rate experience. I also think that these companies will think twice before putting version specific internal systems on their machines, realizing how out of touch they’ve become. “Egg on the face” is still a great motivator.

This group doesn’t stop at the 26% of users, it expands deeply into the IE7 and Safari crowd (IE7/Safari shipping by default with practically all computers now). Most users don’t consider their browser a replaceable, upgradeable program, but as a standard piece of the computer. They think about their browser the same way I think of the DVD player or calculator on my computer: it’s just there, it came with the computer, it does what it’s supposed to. If you told me I should go through the confusing process of installing and learning Calculator 2, you’d need to give me some serious incentives.

The trouble is that the incentives just aren’t there in most cases. Users have spent years in their current browsers, they know them well, and the browsers just work. Firefox’s abstract feature list just isn’t enough to get some users to potentially lose their bookmarks and install a dubious-sounding piece of software. We’re not going to establish upgrade behavior in these users without giving them strong incentives.

In this regard, I think Facebook might be on the right track. Earlier this year, they unrolled a great new chat/notification feature that works for all major browsers, but works especially well for non-IE6 users. And every time an IE6 user logs in, Facebook gently reminds them that they’re using a reduced version of the functionality. However, there’s still no real incentive (they don’t explain to IE6 users that they’re missing something desirable), so we’ll see how that goes.

It’s going to take IE6 a long time to step out of the picture, but once it’s gone, we’ll have the same problem in some other form. Unless there’s a real model for incentivizing and educating the upgrade process, some users will always be lagging behind and we’ll always be building for weird browser issues.

EDIT: Yeah, what Elliott said!

Unfortunately, many people at the top find find it necessary to bend over backwards for the client even when it means having to cater to outdated technology, which frustrates you and me who know the benefits of a modern browser. There’s a level of communication that needs to occur between developers in educating those in non-technical positions as to the benefits of upgrading. This knowledge can then be passed on to the client and eventually trickle down to their users.

The fact of the matter is that as long as the casual user has no reason to upgrade (their sites look/work just fine), they won’t. Why would they?

However, there is still part of me that says “I can’t just leave the site looking totally broken, even if the browser is extremely crappy.” So, I’m writing up a little script in jQuery to disable all stylesheets for IE6 users.  Maybe I should also throw in a nice JavaScript alert on EVERY page load that tells them to upgrade ;-)

Thanks for writing this and validating my perplexed feelings over their use of the word ‘weblog’ and the martian landscape of the admin interface.

For all the reasons you stated I too have added it to my toolbox of development platforms (tagged: fast, flexible)

I don’t know if this holds true for the examples listed but this is why I believe there are beautiful examples of layout with such a defeating technical liability.

The EE “core” package (https://secure.expressionengine.com/download.php?ACT=agreement&id=34) would be worth downloading. I installed it on a demo site a few days ago and from what I can tell, it’s got enough functionality to run a basic blog or web site with custom templates. I haven’t built an entire site on the core package though - does anyone else have input?

One thing to note, the install was a bit of a pain. The new version should be better, but in the current version your URLs will all have “index.php” in the structure unless you take these steps: http://expressionengine.com/wiki/Remove_index.php_From_URLs/

It took me a few extra minutes to sort that out but I was able to get the core package installed and working properly on a personal Dreamhost account in about an hour.

I wouldn’t be able to implement a licensed version for now, but it sounds like you don’t get nearly as much without paying. What are your thoughts?

I’m working on a plugin for OpenID for EE. It’s coming along quite well. We’re hoping to release soon!

thanks for the explanations, I have a better idea now about it all :)

I expect that version 2.0 will work out most of its remaining shortcomings and crown EE king of the CMSs for a lot of people.

A CMS doesn’t always make a site easier to update, especially if you are a web developer who knows how to write and maintain code. The best strategy for choosing and implementing a CMS depends greatly on what type of site you’re working on. If a client is going to want/need to update it, chances are you’ll want it on a CMS.

Even though I know web code, I built my personal portfolio (http://mindywagner.net) on WordPress so that I had a standard format (description, thumbnail image, bigger image, category, etc) for every design piece I put on my site. I use WP categories to automatically organize things. When I post a new design I categorize it web, print, or identity - or some combination of these three. Then it’s sorted on the front-end without me manually adding each new “post” to the necessary pages.

Viget.com (and all 4 blogs) are built on EE. I’m sure Doug can elaborate on the many reasons why, but one obvious reason is that it allows everyone in the company to edit and add to the site without knowing code.

A good way to help less tech-savvy clients get used to the CMS is to create a detailed how-to manual for them with specific instructions and lots of screenshots. A manual can be time consuming to put together, but it will cut down on confusion and give them something to refer to when they are stuck and you aren’t available. Some people go weeks or months between updates so even if you give them a great face-to-face overview it can be forgotten by the next log-in.

So when creating a site and using a CMS like wordpress or EE, what makes them easier to update and maintain control over your website?

and how do you make it easy for a client to edit a news post, or update things on their own?

Thanks to Michael Boyink for the instuctionals, Ellis Labs for their hosting, the team at EE for their support, and Viget for another insightful article praising EE.

Geof, I have no doubts I will look back at my first site and be horrified by how unsophisticated my templates are… looking forward to learning new and elegant ways to use EE.

PXLated - good point. I was primarily thinking about the custom fields for data, which let you create and control things on a much more granular level. There’s no need to chunk out your design templates if you don’t want/need to. And actually you could break your design templates up in WordPress by using lots of includes, but all of it feels much easier to implement in EE.

To clarify #4, one doesn’t necessarily have to chunk up their master html/css in EE, you can include it all in one template, but it does make it much more manageable to use what I call template fragments - header, footer, sidebar, etc. and embed them in other main templates. A change to the master then ripples through all other templates using that fragment.

Great post!