Haiku needs more work
His Haiku is bad
It has too many onji
Unlike mine that is
Jonathan Schwartz, now the ex-CEO of Oracle-owned Sun, went in the same way as he ran the company, oddly: he tweeted a haiku to his followers. With an impish sense of humour he wrote: Today's my last day at Sun. I'll miss it. Seems only fitting to end on a #haiku. Financial crisis/ Stalled too many customers/ CEO no more He' …
Meat for fake Steve jokes
Still with him
Sun is a disaster, much of that might be McNealy's fault, but still, the legacy is awful. Lousy product line, YAF open source license. On an enterprise level, unless they already have significant legacy Solaris/Sparc code, why would anyone ever choose that these days?
Everyone with innovation skills left / customers saw the disaster / IBM and HP pillaged
Stockholders lost billions / IBM quickly became disinterested at any price / Oracle went to the dollar store and got a bargain
Spent $4B on STK and $1B on MySQL / Could not integrate them / MySQL costs billions more as acquisition took 10 months
Solaris 10 success / Open Source blinders / Solaris 11 disaster in the making as it is based on OpenSolaris not Solaris 10
Great career / ruined Sun / career Hara-kiri
Sun in its early life had some huge successes by making software and protocols freely available. They also saw some brilliant technology fail (NeWS, for example) after keeping it proprietary. They saw their lower end market attacked by GNU/Linux, and responded by making much of their software "pay for service" rather than "pay royalties upfront". I think this approach is right for system software, and can often be right for application and database software.
So, while Microsoft and Oracle can make money out of royalties on widely used software, Schwartz didn't have the base of widespread use. In the current age, increasing adoption while charging upfront royalties is almost impossible - all the widespread software items for which people now pay achieved their ubiquity at least a decade ago. So he made the decision that widespread usage without royalties would be more valuable to Sun than limited usage with royalties. As Richard Stallman once put it: "by saying 'you can't use my software unless you give me money first' you are massively disincentivising people to use it".
Oh, and by the way, your haiku doesn't scan correctly.
Jonathan was a horrible CEO. CEOs need to have the long view. They need to operate in future time, in order to make the decisions today which will pay off in five years.
Jonathan was preoccupied with the now. He could not see beyond the end of his bespectacled nose. He put forth crazy idea after crazy idea with no thought of long term implications (anyone remember "Mad Hatter"?). The result was it looked like Sun's strategy was to throw spaghetti at the wall to see what would stick.
Jonathan also seemed to want to turn Sun into a copy of RedHat with a hardware option.
Another mistake for Sun was to chase money losing HPC deals, the same kind of business that repeatedly bankrupted SGI and Cray.
Sun's x86 group, which started off okay, mainly due to the popularity of Opteron, collapsed when Sun was blind to Intel's comeback with Woodcrest and Clovertown. Sun's x86 group never recovered. Meanwhile the neglected SPARC products continued to produce most of Sun's revenue, and the new UltraSPARC-T and SPARC64 products grew while the x86 business shrank.
When all was said and done, the part of the Sun hardware division Oracle is keeping is the SPARC part of the business, along with the core Java and Solaris products, and jettisoning Sun's middleware and most of Sun's x86 hardware, which were Jonathan's babies.
One thing I will say Jonathan did very well on was lighting a fire under Solaris x86. The number of ISVs supporting Solaris x86, and number of hardware vendors supporting Solaris x86 was simply the result of a stellar effort by Sun. I hope the Oracle acquisition does not turn Solaris into an "Oracle only" product. It truly is a gem, and the only operating system that can scale on upcoming Nehalem-EX systems. With the pending collapse of the Itanium market (Tukwila will add nothing of merit to the Itanium platform), a Nehalem-EX/Solaris combination could become the UNIX alternative to IBM's POWER/AIX systems.
SPARC has been severly neglected and even if Larry spends billions he cannot in a few years
stop the squeeze of Nehalem EX and Power7 that will happen in 1Q of 2010.
SPARC64 is end of life except for a minor .06GHz upgrade a little more cache and updating the I/O.
Yellowstone Falls might be a competitive chip last year but it is 2012/2013 and only 4 cores per chip. I don't see why anyone would buy the Fujitsu M-class systems to be replaced by CMT when it is likely it will go the way of ROCK. Larry does not have patience for low ROI and as you said already killing most of the x86 line.
Not sure people need an "Unix" alternative to IBM's POWER/AIX systems. :-)
Biting the hand that feeds IT © 1998–2021