Update #1: To clear up any confusion, this article is focused on the business value of Java – it’s ability to control and generate revenue – not it’s value as a language or a technology. So the word “irrelevant” is applied to it’s business value, not if it’s a good technology or not.
Sun Microsystems announced today that it would be laying off 6,000 employees (18% of it’s workforce) in an attempt to better realign itself with “global economic climate”. The company will be focusing on:
- Application Development (?)
- Systems Platform (Solaris, Glassfish, Java, etc.)
- Infrastructure Development (Server Hardware, etc.)
With Sun’s stock (JAVA) falling from 20 at the same time last year to around 4 today it isn’t much of a surprise that the company has had to reel back costs and rethink some strategies going forward.
You have to wonder if the Java/Open Source-everything strategy Schwartz pushed is paying off or just making the open source folks happy while loosing Sun footing in the servers and platforms market.
With a strong push by the community and interested parties such as RedHat, IBM and Oracle, betting their server and platform offerings on Java/Linux and trying to drive that effort through the OpenJDK project it’s plausible to think that Sun’s own “official” Java release could become less and less important as we move forward.
Eclipse began this trend by completely invalidating the Swing UI stack in the JDK with the introduction of both SWT/JFace and the Eclipse platform itself (implicitly bringing along the importance of OSGi as an industry-leading design force)
With RedHat/JBoss (RHT), Oracle/WebLogic (ORCL) and IBM/WebSphere (IBM) trending mostly flat (RedHat’s value has slipped a bit) for the year and not falling in value like Sun, you can imagine there are some pretty powerful companies, completely vested in Java, that will take this omen of Sun’s slide as a moment to jump forward and become the new stewards of the direction for what used to be a language, but is now a platform.
You can even see some disruptive forces in the Java space like Spring getting their foot in the door and eventually kicking it wide open with “easier solutions” to enterprise Java – that so far the community has mostly agreed with as seen by it’s penetration.
Spring in and of itself is a platform; another means of approaching the Java development problems. Another approach to how to develop and deploy your Java applications (enterprise, client or otherwise).
The purpose of pointing all these pieces out is that there are a lot of big elephants in the room, with their businesses and revenue streams tied directly to the same platform and the one-time stoic steward of this platform is slowly showing signs of weakness. I’m proporting that we are going to see some interesting transitions starting in Java land; not beginning with direct attacks on Java or attempts to wrestle the language out of Sun’s control, but instead peripheral attacks on the language by devaluing it as a language itself and pushing forward as a platform or an engine as a means of execution for other languages controlled by other sources.
One quick example is Spring recently aquiring the Groovy/Grails camp. Not directly Java related, but gives them more leverage in the platform space because it is JVM friendly.
Interestingly enough in an attempt to give longer legs to Java, the Sun team is working hard on opening up the JVM to host alternative dynamic languages; for example, JRuby is becoming hugely popular.
With all these new doors opening, the key to controlling Java may not be through a kung-fu grip directly against the base platform, but by controlling the most popular (and best integrated) 1-off language that becomes the Next Big Thing(tm).
If there is a pending invalidation of the platform I’ve been curious why Sun has been whole-heartedly helping this effort forward with the work on the VM boosting success in hosting additional languages. Certainly Java-proper has enough leverage and penetration now (being through the hype-dissolution-acceptance cycle) that providing the ability to host additional dynamic languages isn’t going to help that penetration by leaps and bounds (I’m talking enterprise commitments end-to-end).
If Sun had a proprietary grip on Java and it’s deployment, then hosting more things on the platform is a great idea, because you still need the platform… but Sun doesn’t have that type of grip over Java. They have a commercial lock-in on the platform side by giving you a do-or-die option (GPL or Commercial License) but beyond that, if you are simply deploying solutions on the platform, Sun has no leverage over you any longer and you are free to innovate right past Sun’s control.
An example might be Linux in this regard. IBM, RedHat, Oracle and a few other big powerhouses all rely on Linux for their company strategies and while it serves as the platform for their technology stacks, Linux (the community, Linus or otherwise) have zero leverage or garuntee of benefit from that relationship. Linux in this case is irrelevant, no one is getting rich off of it, no one controls it and no one can. The gelatenous mass that is “Linux” has no leverage over itself or against these companies if it wanted, it’s a brain-dead participant in the server-stack-war that is being dragged along for the ride because it allows itself to be.
Won’t Sun share this same fate as Java advances and continues to become a platform for platforms. Spring with Groovy, RedHat with Python, IBM with what… PHP? Let’s give Oracle Ruby then (all hypotheticals)… at this point all the big boys now have their own “secret sauce”, it’s all rehosted on Java, Sun has no control over these new Next Big Thing(tm) and it’s platform will just get dragged along for the ride in the form of OpenJDK, ontop of Linux, hosting some proprietary solution that is made relevant by the huge backing force behind it.
We are likely going to see a resurgance of programming-language wars starting up soon as each company picks it mascot, builds tools around it, integration platforms for it and starts pushing their solution as God’s Honest Truth. The question won’t be “Java, Ruby, PHP?” it will be “IBM, RedHat, Oracle… Spring?”
Everything goes in cycles, we did this a decade ago, we’re bound to do it again.
Loading ...
15. November 2008 at 1:12 pm
The thing with cycles hits the point quite well. Take Ajax. A simple combination of Java or other language, JavaScript and a bit of XML. Everything has been around for more than 10 years now.
Java itself or even the Java Community Process won’t go away for at least another 10 years. Sun as a vendor may not be important in let’s say Application Servers based on Java EE. Despite a great debut, GlassFish 3 has made so far. I cannot say, if this is likely to change. Spring and its Application Platform is in some areas covering a lot of what a Java EE Reference Implementation like GlassFish also aims to be.
The commercial products are currently dominated by IBM, Oracle, RedHat and maybe very few others though most of them are less and less relevant. And of course Spring or Apache not to forget.
15. November 2008 at 3:50 pm
Werner,
Ajax is a good example of a missed opportunity; no one grabbed it by the horns and made it theirs and then dictated a standard across the community, so we have no less than 113 AJAX frameworks now.
I’m not saying they *should* have, organic growth can be awesome, just stating that if someone were able to wrestle it to the ground in the early days that probably would have gotten them a lot of control over the direction to take it.
I think that is what we are seeing with the surge of interest in dynamic languages, we have camps building up around a few select languages and then companies stepping in and grabbing up the camps to control them and shape how they move forward, hoping they grabbed the right one.
This doesn’t work unless you control enough mindshare to drive an idea forward, but everyone mentioned in this article owns enough mindshare, consulting contracts and resources to make that happen with whatever they choose… so I’ll be really interested in seeing what happens.
This might also lead to a fragmentation of sorts in the sweeping “Java Development” group, but not in the pure sense… in the sense of “This is what I do on Java”.
Still too early to tell, I just know that what used to be a very black and white “This is Java development” situation is now a very Grey “My Java IDE has dev tools for like 9 different languages now…”
Commercial developers (not open source/hardcore geniuses) are already sick of all the noise, so as soon as one of these folks can consolidate the best approach of all this mess down into a single proprietary stack, tied with tools and a server solution… that’ll be the beginning of it.
I full expect Spring Source to make a server purchase or move in the server space in the next year or so; maybe longer if the Groovy purchase fails. They want that server space as much as anyone else.
servers == hardware == lock-in == huge contracts
Tools were ruined by Eclipse, hard to make money on them any more… can’t make money on operating systems anymore… the only thing left is the support contracts and the hardware solutions.
15. November 2008 at 4:35 pm
Riyad,
Thanks a lot for your prompt remarks. It sounds funny to hear, Eclipse “ruined” the tools market from somebody working for a company selling Eclipse-based tools ?;-)
What you said to have failed for Ajax is something which more or less happened in the area of those tools. Hardly anybody with a sane mind will work a fully blown IDE nowadays. NetBeans still exists of course, but I cannot say whether Sun is able to afford its support after dropping other tools much longer? “Giving in” on the OSGi front looks like coming to reason probably because of economic issues making a fight for different standards and numbering schemes a bit pointless and hard to sit out…
Oracle is the only major vendor still locked-in with JDeveloper. A dinosaur with roots from JBuilder 2 where nothing significant has been done except plugins or extensions for Oracle built around it. Strange that all the refurbishing and straining keep-alive of this White Elephant keeps on, but Oracle can afford it even in times where they might loose to cheaper vendors on various fronts.
Should the shares fall further, even Sun could become a candidate for SpringSource to purchase some day, but I’m not sure, if they would dare all the hardware related stuff Sun’s got? Obviously a true Open Source company might benefit them if Sun ever really was on sale. BEA was clearly overpriced compared to now, but again Oracle will survive it. Maybe Larry Ellison might have to call of another Yacht being built?;-)
15. November 2008 at 6:13 pm
Hah, Ellison’s platinum yacht was reordered as a gold yacht, I wrote him a “Feel Better Soon” card.
I guess “ruined” is a strong word… maybe “changed the shit out of” or “permanently restructured what you could make money at in the tools market”. It went from being single-purpose tools to the more ethereal “value” that you wanted to pay for… and that could be just about anything.
You are absolutely right about NetBeans, it’s a cost center and Sun has been slowly moving resources off of it as time goes on. It’s only purpose is to serve as honey along the rim of the “funnel” that Sun plans on drawing people down, eventually ending in server hardware, and unfortunately for Sun it’s clear that Eclipse is where it’s at. Not starting a war, just that NetBeans/Sun can’t compete with the energy behind Eclipse. At some point NetBeans will be shelved or run in maintenance mode.
I’d expect Sun to buy an Eclipse tools company and restructure what their tool > server > os > hardware stack looks like. Maybe at that point they’ll pick their own mascot-language to tout… maybe JavaFX? Who knows.
Thanks for the comments Werner, always good to talk about this stuff with someone.
15. November 2008 at 10:36 pm
For future reference, since it’s spell-check-proof.
Loose: not tight or securely attached
Lose: to misplace or suffer loss
16. November 2008 at 9:12 am
Thanks a lot Benry, which of the 2 instances of “loose” did you actually mean, the one in Riyad’s original post, or that in one of my replies ?
I cannot say, if Eclipse Babel also requires English Language Champions, but should you consider it, I’ll ask our team lead whether he needs some help there ?;-)
@Riyad: you’re not talking about Sun possibly buying MyEclipse/Genuitec maybe ?!;-)
Cheers,
Werner
16. November 2008 at 9:48 am
haha, I’m embarrassed to say I’ve actually bought a domain name meaning to have the word “lose” in it, but bought the domain with “loose”… you’d figure I’d know how to spell that dang word by now, but it seems not.
I have on my New Years resolution sheet for this year: “Be smarter next year”, so we’ll see if 2009 is any different.
Werner, I really have no idea. Sun is *really* persistent when it comes to their NetBeans offering, and there are still a slew of independent IDE companies out there.
For a while I thought for political reasons that Sun would never acknowledge the Eclipse camp, but I think they are testing the waters of “acceptance” with the OSGi+Glassfish move… indirectly.
I just think in the long term you cannot ignore the power of the Eclipse ecosystem, at some point Sun will move their tools ontop of it… I’d be willing to put $1k on that. I don’t mean this year or maybe even next, but at *some point* they will. We saw Oracle do it and we saw BEA do it… heck we even saw JBuilder/CodeGear do it.
I think two years ago you could make the argument that if you were a big enough company, you could make a platform/toolset comparable to Eclipse, but it’s glaring now the amount of stuff you get for free building on Eclipse. The entry point is so much higher, the value so much higher that is *really hard* to argue for reinvention of the wheel anymore.
For example, the Sun team was restructuring code in NetBeans 5.x and I believe all the way up to 6.x to handle how IDE tasks could be fired off in the background and not block the UI for periods of time. They also rewrote the entire source code parsing infrastructure in 6.x.
How much did all *that* cost? Neither of those tasks are trivial, so figure a senior level salary, and I know the parser was atleast a year of work and god knows how much more work for maintenence and enhancements… so say 2-300k for all that work and all that QA and all those bug fixes… they would have gotten all that out of the box with Eclipse.
Now I’m not saying NetBeans is bad, I’m just saying the value of re-solving the same problems over and over, that a platform used by like 10k companies have already solved the hell out of, has no value to it. It makes less and less sense for Sun to burn so much money on platform concerns in NetBeans instead of tooling targetted towards their server and hardware platforms.
I did notice something *damn* interesting recently, on the NetBeans download page (here: http://download.netbeans.org/netbeans/6.5/rc/) the structuring for the pre-configured download packages are already targeting entire platforms like PHP, C, Ruby and Java.
If Sun’s future lies in Ruby and it seems they are pretty serious about it with JRuby surging forward and from what I’m seeing big support all around from Sun for it (VM and Tools), then they’ll probably make a move and buy a company like Aptana.
If there isn’t big-iron money in Ruby yet, and Sun’s revenue is still coming strong from their servers being sold with the Glassfish stack and Java EE spec-compliant apps (or big Spring apps or whatever), then I think a Java EE centric IDE would make more sense.
Hell… buy both, integrate them, and dominate the market… they could do that I guess too.
16. November 2008 at 10:06 am
Aptana could be a hot candidate for a further SpringSource purchase to add strength behind their new “toy” Grails btw
As is Tasktop whom they already work with quite heavily.
Of course being one of the most “LAMP” centered shops, they could appeal to Sun from their MySQL perspective.
Something I cannot say if they’ll keep it on the other hand. Their focus you mentioned above is a bit vague on that, but clearly as a “Platform” I’d say it is a hidden gem (maybe not the Holy Grail, but even Grails isn’t the whole meal for real IT projects, it is merely a rapid POC tool to impress some execs in a hurry
for Sun if they are smart enough to keep it?
16. November 2008 at 10:12 am
Very good pts Werner… it looks like there are a few different directions these consolidations could go.
I do fully expect Spring to get considerably more aggressive over the next 2 years trying to grab their piece of the server market. Their currently income is from consulting, so their IDE offering will be used to hook people into the “Way of the Spring/Jedi”, but beyond that… I suppose I wouldn’t be surprised by a server acquisition (MortBay/Jetty?).
Very interesting stuff to discuss.
19. November 2008 at 12:44 pm
Very interesting thoughts here, I’ve blogged about them on JavaWorld:
http://www.javaworld.com/community/node/1887
Regards -
Athen O’Shea
Editor, JavaWorld
19. November 2008 at 1:55 pm
Athen,
You did an excellent job summarizing the article; much better than I attempted to do before the JL community turned rabid
Thanks for the insight and pushing the conversation forward. I think if people ignore this and sit around with smiles assuming Java wins everything, they’ll be unpleasantly surprised in 5 years when the market (driven by the enterprise vendors) takes a hard right into a realm we didn’t anticipate.
19. November 2008 at 2:48 pm
Thanks, Riyad – I guess that’s what editors are good for
Seriously, glad to help: your thoughts are worth a deeper look. Let me know if you’d like to write them up in a developed article for JavaWorld.
Athen -
athen at javaworld .com