The Sun Also Rises

Commentary

Sun users are a hardy bunch. In the last year alone, we have had one of our windowing servers (NeWS) declared defunct, our graphical interface (Open Look) abandoned, and our operating system updated to a totally incompatible version (from SunOS to Solaris).

So when I saw that Sun was moving to a totally new GUI, my system administrator and I started swearing in tandem. When all was said and done, I went off to start learning about NEXTSTEP and figure out why Sun hated us so much. I was delighted to find that Sun was actually going to be making our lives easier eventually.

Our work consists of imaging and modeling problems with radar for the Army Corps of Engineers, and we had already been looking into object technology as a solution to some of our programming dilemmas. We've been using Geographical Information Systems, image-processing packages, and homegrown models.

A problem with government data is that it comes in a variety of formats and its users access it through different of models, hardware, and software. Every time a new file format is created, translation programs have to be written to coordinate the new data with the old.

As researchers started learning more about the power that workstations offered, they would ask questions about how they could make their models work with other models. Looking for solutions has lead us into the realm of "frameworks" and object-oriented design to find answers.

The need to communicate with large data sets and processes over the network was also being addressed. The solution we have been looking forward to is Distributed Objects Everywhere (DOE). But it is only available at beta sites, and I hear its functionality is incomplete. The adoption of NEXTSTEP moves our development capabilities into the distributed-object realm on a hardened platform.

There are methods behind Sun's madness (at least in this case). Sun has been keenly aware of the problems its users are attacking and has backed such organizations as the Open GIS Foundation, which is spearheading the attempt to unify spatial- and temporal-modeling and analysis solutions using object technology. Support and investments into technology by companies like NeXT and IONA (a designer of Object Request Brokers in Ireland) give Sun users a feeling that all the pain will be worth it when their technology will still be a part of the current trend.

By backing such organizations, Sun has let its users know that it is committed to keeping them at the forefront of technology. Hopefully, that technology will not come at the expense of its users' sanity. A little backward compatibility would be nice.

Sun is reinventing itself for its users, and I think the users will be patient, as many of us realize that the move is for our own good. Like going from the typewriter to the word processor or from the mainframe to the client-server model, the move to objects is a painful one, with a lot of promised future benefits. If Sun can reinvent its desktop software, it will be viewed as Digital is now a company that has the fastest processor running three popular operating systems and excelling in interoperability. People have forgotten the monolithic system problems that Digital had, and people will forget Sun's desktop problems if it's successful with NEXTSTEP.

Sun seems to have learned from its past mistakes by adopting already-established technologies. This time, hopefully, there will be no word swallowing, as there was when Sun announced its move from OpenWindows to Motif. Projects designed on Solaris now will still work after OpenStep is introduced. World peace will be right around the corner.

Moving to a new GUI is nothing new. Sun users are well prepared for object technology, since they have had to take into account the possibility that even their OS may not be around for long. This situation fosters excellent encapsulation and quick prototyping skills. It's interesting that NEXTSTEP, which has been suffering from hardware-platform consistency, would end up teamed with a company that has equivalent software problems. Maybe we can all finally have a stable platform on which to get our work done.

Sam Gustman is a computer engineer and head of application development for the Geophysical Sciences Branch of the Cold Regions Research and Engineering Lab for the Army Corps of Engineers in Hanover, New Hampshire. His group uses Sun SPARCstations.