Recommendations

Some parts of the route forwards are still unclear. Will Apple remain in business, and will internal politics intrude into the way of getting Rhapsody out of the door? There isn't a zero risk option at this point, but some assumptions are reasonable.

Firstly, the announced timetable is a practical one. Despite rumours that the developer release may slip to September 1997, the full map is achievable. NeXTSTEP has been ported many times before, even to PowerPC. We should assume that a mostly NeXTSTEP developer release, running on some Macintosh configurations, is done.

The initial user release, with some Mac compatibility features, is almost too long a gap. It is possible that user interface politics may fill the timescale, which is undesirable. But given that internal demos have already been given of the "Blue Box" (the System 7 emulation environment) running Photoshop, Excel and Word, there is no reason to doubt it. However the demo was apparently running on the Coplan NuKernel, and hasn't yet been ported to run on MachOS (NeXTSTEP/OpenStep). Only significant problems could prevent the full release from also meeting target.

Although the Blue box means that System 7 applications will run, they are likely to run less efficiently, and with less integration to the environment, than native OpenStep applications. This will mean that there will be some pressure on developers to port to Rhapsody, although it is likely that most are playing a waiting game. Apart from Adobe and Metrowerks, there has been no great show of hands supporting full conversion; and John Warnock of Adobe has made mixed statements about full support for Rhapsody; these were later retracted, but their impact remains.

Developers will either accept the performance losses of running in the box, or will play a waiting game to see what indicators are revealed. Only existing NeXTSTEP developers are likely to move full steam ahead into Rhapsody, aiming at leveraging their existing advantages.

To port to Rhapsody a Mac developer will have to either use a porting kit, or rewrite from scratch. A NeXTSTEP developer will have to port to OpenStep. This is significant work. The official quote from NeXT is that it takes one man month to port 50,000 lines of code; in reality, a full project will take several times that long. Porting to Java using Netscape's IFC is only slightly longer than this.

The key application for many users is Microsoft Office. Without it, or a fully equivalent product, Rhapsody will not be a success. The Blue Box emulator will be available in the middle of 1998, although it is likely that the 1997 user release will support MS Office.

This means that there is a gap while either apps are ported to Rhapsody, or we wait for the Blue Box to be released. This gap will be from nine to eighteen months in length, during which the only available applications will be the OpenStep applications. This is a significant opportunity for NeXTSTEP developers.

As far as NeXTSTEP users are concerned, what action should they take? The options are to stay with NeXTSTEP indefinitely, to implement plans to port to Windows NT/95, or plan to port to Rhapsody.

Considering the reasons to stop using NeXTSTEP that applied in the recent past, these will mostly go away. The operating system will be back under support again, will continue to run on Intel hardware, and the Microsoft applications will be available. Exact timescale will need to be monitored, and performance of MS Office under emulation should be considered. The advantage of custom applications and cost of ownership still remain.

Given the current circumstances, it seems wise to continue to defer plans to migrate from NeXTSTEP, while continuing to evaluate both NT and now Rhapsody. The reasons to migrate immediately are going away, but there is no definite reason to cancel these plans entirely.

Users should keep a watching brief, while the viability of their existing situation has been extended.