In his recent blog entry, fellow blogger David Berlind is all over Microsoft's "Office bloat" and is convinced Google is in a better position to understand users' needs for an Office application. Despite the fact that Microsoft collected 1.2 billion data sessions to design its new Office 2007 user interface, David still feels that Google knows users better. But I have to wonder what all this commotion about "Office bloat" is. [Update 7:00PM - David Berlind claifies his position here and I responded here.]
The way I see it, if a product does everything I need it to and it happens to have features that I don't use, it's no skin off my nose unless those unused features affect me negatively. When it comes to software, something like Adobe Photoshop or Corel Draw will tend to have tons of features I'm never going to use. But if that software has all the features I do want and it packages those features in a usable user interface, I'm perfectly happy with it. Furthermore, just because some users don't use some features doesn't mean that they're not used by other users. Does it matter to you that you never touch more than 99% of the items in a grocery store? We simply can't assume that everyone will have the same habits and training level and that they will all use the same features.
Measuring the percentage of features individual users use isn't really meaningful; measuring whether features are used at all by any meaningful percentage of users is what's important. Do you buy even 1% of the total items in a grocery store? Does it matter to you that you never touch more than 99% of the items in a grocery store? Should we start saying that grocery stores are bloated? Of course not! What's more likely to happen is that if you can't find 5% of the products you're looking for, you'll just end up going to another store that does have everything you're looking for. [Update 7:00PM - Some readers have responded in the talkback that they don't think this is a valid analogy because they're not forced to buy out the entire grocery store. But all the food in a grocery stores doesn't cost $200 and it can't fit in 1% of my refrigerator and it won't stay fresh indefinitely. If that were the case, I'd be crazy not to buy out the entire grocery store.]
The kind of bloat that bothers me isn't feature bloat; what really bothers me is resource bloat. I'm talking about the kind of bloat that led me to conclude that OpenOffice.org is a pig when it comes to hogging all the memory and CPU resources. Microsoft Office in comparison consumes very little CPU and memory resources even though it contains so many features. But despite the fact that OpenOffice.org is such a resource pig, I'll use it over a Web-based Office any day of the week.
The only thing worse than waiting for OpenOffice.org to start up is having to wait for a Web interface to load. The only thing worse than waiting for a Web interface to load is when the application returns an error or when you accidentally hit the back button before you saved all your work. Anyone who's ever tried to type a long and spirited post on an online forum only to have all that work vanish into the World Wide Web knows the pain of tearing their hair out. There isn't a single Web-based Office application that I can take seriously.
ThinkFree's online Office Suite has the only user interface that can be taken halfway seriously, but the use of Java makes it so massively bloated that it makes OpenOffice.org look good. Each single applet and session uses over 100 megabytes right off the bat before the data is even loaded. While it's great that it lets you save to an online drive or a local drive, the fact that it forces you to first open the online drive makes it painfully slow. I have absolutely no interest in twiddling my thumbs while a Java Applet loads, so there are no usable Web-based Office applications, as far as I'm concerned. There are Java-free AJAX-based alternatives, but their user interfaces are even more primitive, and it isn't easy coding in AJAX.
If someone really wants to compete with Microsoft, create an application suite that's less than 3 MBs with SpreadSheet.exe, WordProcessor.exe, or Mail.exe that I can carry with me on a USB stick. For cross-platform compatibility, create a native executable for Linux and Mac OS. Then, I can instantly download it anytime off of the Internet or get it off the USB stick. The mere act of using the secure login should automatically load my personal configuration and give me access to all my data on the USB stick or on the Internet. If anyone doubts the effectiveness of slimmed down feature-rich applications, Ludvig Strigeus created uTorrent (merely 170 KB), and he took over the BitTorrent client market in a little over a year. uTorrent defied the conventional wisdom that people don't care about tight code and proved that people like their resources and time treated with respect.