Plans for the Future

Users

Given the situation prior to December 20, 1996, most users of NeXTSTEP were of the opinion that NeXT would not be enhancing the product, and had already made plans as to how they would migrate to other platforms.

Windows NT

In general, the options being considered by users were migrations to Windows NT, Windows 95; sometime in combination with either WebObjects or OpenStep/NT. Other options were not widely considered.

Because of the strategic importance of custom applications developed for NeXTSTEP, most users had decided to remain with NeXTSTEP for so long as it proved viable. This would continue until hardware support for current Intel add on boards, and the supply of NeXTSTEP drivers, had diminished to the point that hardware failures were interfering with the conduct of business. At that point the previously selected exit strategy would be implemented.

The cost of porting to OpenStep was considered to be too high by most users; it was often comparable to the cost of redeveloping the applications using design lessons learnt from the NeXTSTEP implementations. The additional cost of having to license OpenStep (for $500 per machine) for the deployment was considered to be a deal breaker in most cases, given the existing expenditure and cost of porting.

WebObjects was not an option for most users: the crude HTML interface is not appropriate for the scale and significance of application usually developed with NeXTSTEP.

ISVs

ISVs were in a similar situation. Most were doing general, although NeXTSTEP oriented, consultancy, and so found it easy to become independent of NeXT. The loss of revenue from NeXTSTEP specific products was so small as to not be worth considering.

OpenStep

Staying with the NeXT community was not really much of an option. Porting existing applications to OpenStep was far too expensive to be worth considering; and the cost of deployment licenses (between $100 and $500 per seat, depending upon volume) far too high for an independent application. So the potential for the uptake of OpenStep in either the NT or Solaris forms was inhibited.

The claimed porting cost (from NeXTSTEP to OpenStep) of one man month per 50,000 lines of code is close, but an underestimate. Allowing for proper testing and the more thorough programming required for a commercial application, this estimate can easily treble. It is also comparable to the initial cost of developing the application, or that of porting to a new, but comparable, environment, like Java with IFC.

WebObjects

WebObjects had a similar problem. The ideal version to use was the Enterprise product, priced at $25,000 per server. This was a larger price in most cases than the tag for developing the application to run on the server. In addition, NeXT were selling their consultancy services in competition with their former partners, making it even less attractive to work with them.

Java

These considerations, along with the general market flow in favour of Java, meant that most developers were actively considering moving to Java as a general strategy. Java is very similar in many ways to NeXTSTEP for developers, especially when used with Netscape's IFC framework. The cost of porting applications from NeXTSTEP to Java was estimated by several developers as being slightly more, but generally comparable, to the cost of porting to OpenStep. The only factors not in Java's favour was the relative weakness of the surrounding tools, and the immaturity of the market.

Pages designed by:

Paul Lynch
paul@plsys.co.uk

Last updated April 25, 1997.