Hello NeXT Community: I just had a very nice contact from an David exNeXTy that worked briefly at NeXT on the NRW so towards the very end of production:
He wrote nice bit about it , what perks my interest is Kevin Enderby , does any one in the forums know his whereabouts as there is a gentleman that trains Ferrari drivers lol might be him , sounds like fun. With the F1 movie .
A couple things below I did not know read an enjoy as I've never heard anything about the actual NRW development at NeXT , so all of this is amazing :)
:)
Thanks, Rob. That's very generous of you to give me some editorial space in your next newsletter. I'd love to contribute but unfortunately, I don't think I'd have much to offer. I was hired just when the initial NRW design was complete and right around the time that the first prototype was powered on in the lab, so I wasn't party to any of the initial design discussions nor did I have access to any of the design documents. Thanks, Rob. That's very generous of you to give me some editorial space in your next newsletter. I'd love to contribute but unfortunately, I don't think I'd have much to offer.
I was hired just when the initial NRW design was complete and right around the time that the first prototype was powered on in the lab, so I wasn't party to any of the initial design discussions nor did I have access to any of the design documents.
Although my background was in hardware design/architecture (I spent 10 years at Amdahl/Fujitsu participating in the design of high-performance pipelined CPUs), at NeXT I didn't do any ASIC design as my responsibility was in the area of performance.
So, I worked with the hardware guys in benchmarking performance and identifying bottlenecks. I also wrote a NeXTstep wrapper for a verilog simulator of NRW. Although my background was in hardware design/architecture (I spent 10 years at Amdahl/Fujitsu participating in the design of high-performance pipelined CPUs), at NeXT I didn't do any ASIC design as my responsibility was in the area of performance. So, I worked with the hardware guys in benchmarking performance and identifying bottlenecks. I also wrote a NeXTstep wrapper for a verilog simulator of NRW.
About the closest I got to any actual design decisions was when the processor was chosen. As I recall, the original plan was that it would be the 88110, which was the leading processor in Motorola's 88K RISC line at the time, but the Intel Pentium and IBM's PowerPC were also contenders.
I wasn't really in the decision-making loop, but as I recall no final decision had been made when the hardware group was shut down.
There was an effort to get funding from *Canon to continue the development but I believe that fell through (I had already left the company by then).
* "I think Canon used the NeXT assembly line to manufacture Canon Object Stations"
Outside of the lab,
the only other time I saw an
NRW running NeXTstep was I believe at the first NeXT reunion
where Kevin Enderby .......... * "Kevin are you out there, this is awesome?"
had performed a super-human effort to get all the machines used in development at NeXT since inception up and running,
starting with the first Sun SPARCStation that was used as the initial development platform for NeXTstep.
* "I'm not sure but think it may have been a Sun workstation , I think it was a Sun 3/60 , if that is the case up until which version of NeXTstep??"
Disclaimer - this account may have omissions and serious mistakes but after 35 years that's all I remember :). Members of the original design team would certainly have more intimate technical knowledge of the development timeline than I.
David
*Are comments from me Rob Blessin, I didn't realize how important Kevin was in the whole process how cool.
I found a couple of patents that are neat as well from Kevin and a few familiar names, explains the multi architecture , the dates granted are also interesting.
Patents by Inventor Kevin Enderby
Kevin Enderby has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
System and method for dynamically loading object modules
Patent number: 7743371
Abstract: A system and method for dynamically loading object modules is described herein. In one embodiment, the method comprising loading a first set of instructions into a memory unit, wherein the first set of instructions includes an unresolved reference to a second set of instructions, and wherein the loading includes replacing the unresolved reference with an address of a third set of instructions. The method also includes executing instructions of the first set and executing instructions of the third set to determine the address of the second set. The method also includes loading instructions of the second set into the memory unit, after determining the address of the second set. Additionally, the method includes executing instructions of the second set.
Type: Grant
Filed: March 10, 2004
Date of Patent: June 22, 2010
Assignee: Apple Inc.
Inventors: Steve Naroff, Kevin Enderby, Nick Kledzik
Method and apparatus for architecture independent executable files
Patent number: 5604905
Abstract: The present invention is a method that enables single release of applications for multiple architectures and operating systems and to provide ease of use of applications in multiple architecture environments. The present invention provides a single file that contains separate object code each of multiple architectures. A special header on the file identifies each section of object code and includes pointers to its starting location. When the file is to be executed on a particular architecture, the resident operating system identifies that block of object code most suited for that particular architecture and environment. That section of code is then loaded into memory for execution. Each architecture in the file is specified by CPU-type and CPU sub-type. For each CPU type or CPU sub-type, file offset, file size and alignment is also provided. Padded bytes are provided to place each member on its specific alignment.
Type: Grant
Filed: February 28, 1995
Date of Patent: February 18, 1997
Assignee: NeXT Software, Inc.
Inventors: Avadis Tevanian, Michael Demoney, Kevin Enderby, Douglas Wiebe, Garth Snyder
Method and apparatus for architecture independent executable files
Patent number: 5432937
Abstract: The present invention is a method that enables single release of applications for multiple architectures and operating systems and to provide ease of use of applications in multiple architecture environments. The present invention provides a single file that contains separate object code each of multiple architectures. A special header on the file identifies each section of object code and includes pointers to its starting location. When the file is to be executed on a particular architecture, the resident operating system identifies that block of object code most suited for that particular architecture and environment. That section of code is then loaded into memory for execution. Each architecture in the file is specified by CPU-type and CPU sub-type. For each CPU type or CPU sub-type, file offset, file size and alignment is also provided. Padded bytes are provided to place each member on its specific alignment.
Type: Grant
Filed: August 20, 1993
Date of Patent: July 11, 1995
Assignee: NeXT Computer, Inc.
Inventors: Avadis Tevanian, Michael DeMoney, Kevin Enderby, Douglas Wiebe, Garth Snyder
I have some of the design documents for NRW... we should probably archive those (assuming Apple won't be chasing us).
Quote from: gtnicol on June 18, 2025, 10:31:00 AMI have some of the design documents for NRW... we should probably archive those (assuming Apple won't be chasing us).
I think the documents should definitely be saved for the future. They have a value in preserving technological history. I don't think they have any other value these days.