Thursday, 2 April 2009

Sounds like Sun is pretty much fucked, then ...

Doesn't sound very promising:

What we actually know amounts to very little. The Wall Street Journal printed a story based on information received that IBM was thinking of buying Sun. IBM said "no comment" as did Sun. Neither company denied that acquisition talks were taking place.

Then Intel CEO Paul Otellini said Sun had been shopping itself around to several potential purchasers. Again Sun did not issue a denial. Crucially Otellini has not denied he said this or altered what it is he said - he believes Sun has been hawking itself around potential purchasers.

What do the WSJ article and the Otellini comment tell us? They indicate that Sun's senior management doesn't believe Sun can prosper on its own and needs to be taken over. The widespread shopping around of Sun by the Schwartz-led management team tells us all that the Schwartz team - and the Sun board - have no faith in Sun's stand-alone survival prospects. They don't see prosperity for Sun and its shareholders in that direction.

Don't think I'll be advising anyone to invest in Sun tin, then. And what about Java and MySQL. I reckon anyone running their business on any of the above must be shitting themselves.

Hahahahahahahahahaha ....

Update: Rob asks me why this is a bad thing. Well, the problem is that a LOT (and I mean a LOT) of businesses use Sun hardware (mostly because Solaris is still one of the best UNIX implementations out there), even more websites and businesses depend on MySQL and almost everybody is affected in some way, shape or form by Java.

If the top brass at Sun are not confident that their business can stand on its own two feet, then it means that they are looking for an acquisition. The immediate concern for anyone who uses any of the above technology is that it makes the future uncertain. If you have no reason to believe the company in question has any problems, you carry on with conviction.

But when it's clear that Sun is on shaky ground, it calls into question large swathes of people's IT strategy. It's going to damage faith in the LAMP stack (Linux, Apache, MySQL, PHP) for many websites, it's going reverberate around the Java world and it's going to cause an uproar in the tin business.

Even if Sun is acquired, it does not necessarily mean that their technology will survive, either. Just ask the guys from Sequent, who saw their NUMA-Q technology quietly killed off, or Informix, who very nearly saw their database technology snuffed.

Personally, I can't see Java dying off because it's pretty easy to take the final steps to open source it (although that is no guarantee); MySQL is already open source and so that should survive (in some form or another -- it may splinter like Linux); but if I was running my business on Solaris kit now, I'd be shitting bricks.

In an ideal would, whoever bought Sun would do no harm to the software, and port Solaris to their propietary kit (because, IMHO, Sun's hardware is fucking rubbish and all the propietary Unices are crap! :o)

But in reality, this sounds very much like the death knell for Sun and Solaris; Java doesn't make Sun (or the new buyer) any money and MySQL is open source and open source is a very difficult business model.

So I think there could be a world of IT pain heading in, whether Sun gets bought or not.


Rich said...

Hey Clownie, could you explain briefly to the non-techies why this is bad news?

DaveA said...

If we look at Java there are 2 ways of looking at it. Firstly if they dump Java what will they replace it with,.Net? I think not, or is this a cunning plan by IBM to dominate the language market.

WTF lots of wanking material for Obo, kleenex anyone?

Obnoxio The Clown said...

@DaveA: The last time IBM dominated the language market, we got COBOL, so I'm seriously fucking hoping they don't.

I don't know what will happen to Java, but personally I can't see it going away. It would be a very brave vendor that gave Java a double tap. :o)

Kleenex is so last year, this year it's Timney Tissues.

Roger Thornhill said...

I thought Solaris had a good interrupt/prio/sched mechanism and their compartments idea was good, too, making use of large numbers of cores SPARC supports.

I suppose it might mean good news for Itanium and PPC, seeing as Xeon does not scale as well (unless tech is folded in from Itanium).

Bristol Dave said...

It would be a very brave vendor that gave Java a double tap. :o)

But a good one, nonetheless. I can't fucking stand Java.

Edwin Greenwood said...

I guess that validates the sinking feeling I got when Monty and Co flogged MySQL off to Sun.

In principle, since the source is out there, it could be sourceforged back into the community, but that's not going to sit well with managers who either don't have techies on tap or need to feel the contractual faux security of having a support company to blame.

Then again, I'm no longer involved in deploying and supporting LAMP stacks, so why should I care?

Obnoxio The Clown said...

@Brizzle: Java sucks donkey cock, but it's not as bad as the cunts who program it!

Edwin Greenwood said...
This comment has been removed by the author.
Edwin Greenwood said...

"The last time IBM dominated the language market, we got COBOL,..."

Ah yes, but they did give us PL/I too, the bastard child of COBOL and FORTRAN. Strange language, but it kind of grew on you. I quite missed it when I started working on PCs.

"I can't fucking stand Java."

Java is like applying for planning permission for a garden shed. Lots of fucking bureaucracy irrespective of the significance of the result.

Language war, anyone?

Obnoxio The Clown said...

Ah yes, PL/I was kind of cool, I remember it fondly. But it wasn't exacly dominant, was it? :o)

DaveA said...

I think Pl/1 is also influenced by Algol too. On Sun's O/S their future is assured as 50%+ of investment banks run Sun and most apps are Java. Like Sybase it is not going away either and £1 trillion of toxic debt is not going to rewritten overnight, where is the budget?

Pogo said...

Ah... Good old PL/1 - "combining the worst features of COBOL and Fortran". :-)

Actually, IBM surpassed themselves in the vileness stakes with RPG-II.

Edwin Greenwood said...

Yes, ALGOL definitely had a hand in the parentage of PL/I, particularly in terms of block structure and scope. That's what was nice about PL/I. Like PHP these days, it was never an a priori academic bondage and discipline project, more a practical coder's eclectic exercise, as in: "That feature looks good and seems to makes life easier, we'll have a bit of that."

Chaos, but entertaining chaos. You could happily poke around the manual for years, discovering exotic features like the AREA datatype (if memory serves), which in poorly implemented systems allowed you to poke around in the data left by the previous process to use that memory. ('Deadbeef' to you, too, mate!)

Java, Solaris, MySQL, etc, aren't going to disappear. There's too much capital invested in them. Though specialized support might get more pricey.

This is a concept difficult to get across to marketing and management types. I spent much of the last two decades looking after legacy systems, as in:

Management: Current standard is Oracle under NT4.

Moi:. OK, but this rather important app is written in pure client-side FoxPro, with the data on a NetWare server. To port it to this week's preferred standard will cost you about £500K.

Management: Ah. Carry on for now, then.

Moi: Can I at least have some money to keep it ticking over?

Management: Er, no. It does meet current company standards.

Rinse and repeat until a catastrophic failure occurs.

DaveA said...

@Obo: "The last time IBM dominated the language market, we got COBOL, so I'm seriously fucking hoping they don't"

There are four areas to a Java program, n'est pas?

Now calm down boys, I appreciate you cannot afford Mr Jacqui Smiff's porn bill, but on the next few buzz words, hands on the keyboard please for some Java jazz.

Polymorphism, software re-use, multiple inheritance, dynamic binding....

It comes off with a damp sponge.

Jack Maturin said...

Why is Java so damned complicated? The solution to the horrible EJB world was Spring and Hibernate, but just getting a simple web page up with Spring and Hibernate is like getting Gordon Brown to say "Yes Guv, I got it wrong". Yes, it's possible. Eventually. But it's painful, long-winded, and requires the use of lots of expensive experts to make it look good.

The whole IT contractor mularkey, at two and three grand a week, was all part of the bubble economy, and now that the bubble economy has gone south, very few are going to pay out shedloads to hire people smart enough to make Java work at the Enterprise level (without having to understand and alter about 242 disparate XML configuration files every time you want to blow your metaphorical nose).

They're going to want simple modular systems that bolt together easily and simply, that an entrepreneur can knock together themself on a PC, with professional looking results, to keep costs down and flexibility up.

I don't know what that kind of thing that may be (Microsoft develope tools anyone?), because it's been a while since I used much technology beyond an Excel spreadsheet and Visual Basic add-ons. But from what I've seen of Java, it ain't going to be the winner in that particular survival-of-the-fittest competition.

It's just too complicated, too expensive to deploy and maintain, and when you decide to use it, you find yourself trapped in the hands of the techies.

Question: My son wants to learn a programming language. Which one d'you reckon I should advise him to learn?

I'm leaning towards Visual Basic, because it comes free with Excel, and you can make pretty things real quick with it, on spreadsheets. Or should it be Perl, Java, C, C++, Lisp, Prolog, C#, whatever?

What is the most economically useful language for someone to learn? Not the most technically satisfying. Not the most perfect in terms of harmonious structure. But what is the most economically useful, which will deliver the most bangs per buck, and which will stand the test of time?

Obnoxio The Clown said...

C is still a good option or C++. If he can program in one language, it's not difficult to teach him another. Scripting languages like PERL or PHP are good for bang for the buck, which is why I do most of my programming in KSH. :o)

Jack Maturin said...

Cheers Obo,

Dear Lord, 'C'.

I'll have to go into therapy then, to get rid of my linked list nightmares before I beginneth the lesson.

Still, it's a nice progression. Learn 'C' for anatomy, 'C++' for physiology, and 'Perl' for biochemistry.

Korn Shell? You are a VERY BAD man. Still, it's a damn sight better than Bourne Shell or the simply appalling C-shell.

Ksh to bandage it all together.

It could work! ;-)