Slashdot Mirror

Simon Janes wrote in with a very interesting editorial for us to read. This marks the first time that someone besides me has had an Editorial posted on Slashdot, but this piece is so worth reading that I just had to post it. Simon writes about the hardware and software industry, and specifically Sun. His conclusions are "out there" but each time I read this thing, I think he's more and more right. Pay attention guys because this is a big one. Everything after this point is written by Simon Janes.

After reading The Washington Post today (" Sun's Lonely Battle", 8 Feb 1998 H1, H15), it struck home to me that Sun needs to be given hope against Microsoft and we all need some kind of hope for a good machine architecture in return. First look at what has happened to other processor architectures in the last 12 months:

Alpha Compaq has stated nothing about what it plans to do with Digital's Alpha making everyone wildly speculate about what is is going to do with it. Some do not like how Compaq has made its hardware very Windows specific or hard to administer at the bare-metal level (no BIOS? you mean I need a floppy disk with Compaq tools? how INCONVENIENT!)

PowerPC Motorola is retargetting PowerPC towards embedded applications because of Apple's monopolistic intents. Personally, I think Steve Jobs has a deeper plan of revenge against Apple, and is running it into the ground under the guise of being a saviour. Motorola was *just* about to come out with a line of G3 based PC's and Apple ruined it. I can imagine a lot of people at PowerComputing are so agry with Steve Jobs, that they would surely create an instant medical condition if they met Steve on sight. Motorola ate a $100M loss because of it-- which was really immature (the "we won't play your game anymore" game), Motorola & IBM should have sued the collective pants off of Apple for turning over the PowerPC apple-cart (the "we won't play your game anymore" game played the CORRECT way).

Intel Intel's offerings are increasingly getting proprietary, and proprietary means expensive. When and if the Merced processor based on IA-64 comes out, it will require a massive compiler re-engineering on the hands of the Free Software Foundation/Cygnus/LDC. On top of that Intel is targetting the Merced for "mainframes". Once you append the word mainframe to a computer, the price of that computer multiplys by several hundred percent. AMD and Cyrix are working hard to present alternatives to Intel, but in a market dominated by Intel and Microsoft, I fear that they will have limited sucess in this arena.

StrongARM A majority of the StrongARM design team have left their posts rather than work for Intel, who apparently now owns the rights to StrongARM. Granted, the StrongARM isn't your super-duper high performance server-class type of microprocessor, but I felt that it would have made for an excellent base for laptop computers.

MIPS MIPS machines sit there in the background. Acer at one time had a MIPS machine you could buy, but that disappeared into the chasm of Wintel. Cobalt Microservers is a start-up who could potentially turn MIPS around. Silicon Graphics is constantly in trouble with shareholder lawsuits so therefore has a cloud over its future.

Where is the hope then? I feel that it is with Sun. The SPARC and UltraSPARC have not had this kind of mayhem perpetuated on them. They are open specifications of which many other companies have implemented and sell-- however Sun is the largest, and most recognizable of them. Much of the hardware Sun has produced is still usable today so except for volume, Sun hardware has a similar history to that of Intel-based hardware (starting from the 386).

Sun needs help for this lonely fight, and it has one excellent place to turn to for it: The Linux Developer Community. Sun needs to follow the two-step method of world domination:

1. Officially support the Linux developer community (LDC) to support all grades of Sun hardware, from the low end (which are already pretty well supported) to the higher end (which are hard for the LDC to support because of lack of access and documentation. Sun needs to recognize that even though its Solaris is good, people love to have choices when it comes to what they run on their hardware. Sun, when it takes this first step, which Digital has already taken, must follow it with the step Digital has not taken--
2. Recognize and promote Linux as an alternative operating system for workstations and servers and ship it as a preinstalled option. I followed a banner advertisement yesterday from Wired news to Digital's little press story on the Titanic being rendered on Alpha hardware. Linux was not even mentioned ONCE in this press statement from Digital. This is not support of Linux on Digital's part, this is a cover-up. (If you go to Digital's Sucess Stories site to find customers who use Digital hardware, you will not even find Linux listed as an operating system option. Another coincidence?) When and if Sun takes the first step of supporting the LDC, it must also follow with the second step-- Promoting of the alternatives as well.

Sun is fighting a lonely battle indeed, but the Linux Developer Cavalry is there, waiting in the woods for Sun's call. Sun, are you listening? Victory is two years away, shine in the next millenium.

Ok, now I'm (Rob) back with a few closing remarks. I think that RedHat has proven that you can take the Linux OS, treat the Linux Community good (RPM, RHAD) and produce a quality product that you can sell. Sun is in a similiar position- they control the hardware, but if they were to openly support and port to Linux, they could take advantage of the amazing power of the Linux Community. Then Sun could concentrate on the 'value added' stuff that makes their platform special. Sure, it means ditching Solaris, but it allows Sun to focus programmers on stuff that isn't already being done better, elsewhere. I'm sure a handful of specialists devoted to Sparc/Linux would produce an amazingly optimized Linux Kernel. This would produce the fastest possible Linux platform, and it woudl make the rapidly growing Linux Population buy Sun hardware instead of Digital.

It wouldn't be easy, but I suspect it could be done. What do you guys think?