Further WWDC 2004
I got a chance to watch the keynote last night and had some additional views.
Searchlight: I'm really impressed by this. This really is where desktop computing is headed. Using Quicksilver, I have had a taste of seach based computing, and hardly ever use the finder any more except for file management. This implementation is much like what BeOS had to offer, without having to create a new filesystem to deal with the metadata. It seems that Apple is really abstracting the file system away from applications and processes. It will be interesting to see how this system performs on anything other than a G5.
Dashboard: There has been a lot of muttering about this in the mac world, and for good reason. Much like Watson, Apple has again taken an idea from an independent developer and appropriated it for themselves without adding any additional value, or doing anything better. And yes, while it could be argued that Desk Accessories did much the same thing in the earlier systems, and that this is a logical extension of DAs, the implementation is almost a verbatim copy of Konfabulator. The thing that really kills me about this is that there is no reason to do it.
To be fair, I downloaded Konfabulator yesterday and decided to give it another spin. I still don't like it. Why? Well, for a number of reasons. Chief amongst them being that there were perhaps one or two widgets out of 660 currently developed that I would actually use. Their functionality is just better implemented in other solutions or applications. Do I really need to have 20 search widgets for different sites? No I don't. If I want to pull down a customized list of web cams, I will throw them in an HTML file and bookmark it in my browser. Also, and some would view this as a good thing I suppose, is that the widgets have no consistent interface design. Like the bad old days of skinning Soundjam, you would get horrible looking designs that take up half the screen and look like a big fish, just because someone could. Or you get a nice tight beautiful design with the controlling buttons unlabeled and hidden along an edge. UI guidelines are there for a reason. This is not to say that people couldn't make widgets that conform to UI guidelines, they just don't seem to want to.
In general I still have an issue with splitting applications into "Applications" and "Widgets". If Dashboard were an application like Konfabulator instead of being integrated into the OS, it wouldn't be such a big deal. People shouldn't have too much trouble understanding that widgets are modules that run inside an application and offer quick information when you need it. Making it part of the OS leaves a huge grey area. What's the difference between an iTunes controller widget and the minimized iTunes controller? Why do I have a calculator application and a calculator widget? Why isn't my network system preference a widget rather than a modularized piece of System Preferences?
This idea of a dashboard while being very cool looking, is ill conceived, and without a solid goal. It does not move the OS forward, like expose, it obfuscates it and makes it more confusing to use. If apple insists on including it, I hope they make it an App rather than an integrated part of the OS. Again, I don't really mind it as an app. If people find it useful, it's fine to offer this functionality in an application where the boundaries are clear to users between the OS and modularized functions within an application. I'm sure some people find Konfabulator an essential part of their computing experience, but I seriously doubt that most users will take advantage of this. It's glitz for the demo. Although the rippling that happens when new widgets are selected is really cool.
Xgrid: It was under reported, but steve mentioned that Xgrid would be included in the client edition of the OS by default (as well as the server version as reported on the server web page for tiger). This is going to be one of Apple's secret weapons in the coming years. I have said it before, and will say it again, pervasive grid computing will be a killer app in the future, especially as more and more home users begin editing their home movies, making their own music and digitizing their video and music collections at home. Large companies that use intensive applications in one department, and have legions of word processors in another will understandably be excited at the ability to better leverage their resources. The real trick is making every app take advantage of Xgrid that can, and making it painless to configure and administer. Perhaps a checkbox in Xcode that when coded will automatically search for parallel operations and optimize them for Xgrid. Apple can make this happen.
Server: Apple needs to go on the road with Server. It is incredibly impressive, and they need to do a carnival style, barker and all roadshow. Take a semi out, loaded up with xserves and demonstrate NT migration, show off the failover capabilities. Demonstrate the secure chat server and ichat 4 way video chatting. Tell them of the benefits of internal blogging for development and communication within a company. Apple is leveraging some really cool stuff into a great turnkey solution for companies and it's not doing a good job of hawking it.
Posted by Joe Mullins at June 29, 2004 10:48 AM | TrackBack

