Even though I never had a formal course in computer science I can solve problems none of the computer science people I work with can touch because of the math I can do.
I've often said that software development comes down to two polar opposites: at one, it's all about people. The other, it's all about computer science.
Math is the universal key.
It's an important key, but it's certainly not the only one. One can be just as effective in life and career if they become a master of understanding people. That's especially true for business software development.
As someone who has been working professionally with C# for years, and who has found it to be a much faster way to deliver quality production code, I contend that C# is the true benchmark: Java has already been superceded.
C# is a good language, and for many types of rich Windows applications I agree it is more productive approach than Java + Swing; the event and delegate model is very intuitive.
Having said that, Java is also very productive; a lot of the problems have been the era in which older applications were written. People didn't know a lot, and abused the language. leading to monstrosities.
ASP.NET is also a very productive approach to web sites, but it isn't really significantly productive than Java + Struts, or JSF.
I've seen a few.NET monstrosities already; the platform is so new (only 4 years since the 1.0 GA) that they haven't really trickled out to the public to the extent that Java failure have after over 10 years of use.
Java's not that much of a mess, it's just a big thing that targets lots of people, and so there's a lot to it.
Also note, XDoclet isn't a part of the Java platform, it's a community project. There's plenty of examples of that in any language/platform... CPAN for Perl, Ruby gems for Ruby, etc...
Deployment descriptors aren't broken, the issue is that they don't recognize that developers sometimes make policy decisions at the early parts of a project, so it make sense to keep that declarative stuff close to the code. Yet you still need the external XML file to override the declarative policy, as needed. EJB 3 adopts this approach....
But it's typically a complicated relationship. Something that performs very well with one user could very well be scalable; the issue tends to be the amount of indirection one goes through to ensure scalability.
Put another way, performance can be had by maximizing the usage of available resources. Scalability under concurrent load is achieved by minimizing resource usage, especially resources that can only be used by one user at a time.
Good software that exhibits "scalable performance" is about knowing what approach to pick, and when.
$16B+ in revenues annually. That's 4x eBay's current revenue and 3.5x Google's current revenue.
eBay's market cap is slightly higher at the moment, and Google's is in the stratosphere (does it really deserve to be valued more than IBM??)
Having said that, eBay and Google are more profitable, and Google is growing revenue faster. Apple's current 60%+ quarterly growth needs a new hit product to sustain.
Armed conflict is power of might over right. And minority over majority. It's , by nature, anti-democratic. And , as far as I can tell, a majority of Americans picked this administration - twice. So what you're basically saying is, if I don't agree with everyone else, I'll just have to kill them to get my way.
My point: There have been lots of bloody revolutions, most of which left things in worse shape than before. Oh, sure, they were one step to an eventual "better world", but I don't think you'd like to live in France during Robespierre, nor a businessman in Russia during the Bolshevik revolution. Speaking of that one, Russia arguably is still looking for that "better world"; I don't think it's a stretch to argue that particular revolution was for the worse.
My second point - lots of regimes have fallen in recent years due to non-violent resistance, which also seems to be very effective.
Your analogy does not hold. Corporate capitalism is not at all like democracy in practice. Shareholders have long ago abdicated responsibility for corporate actions.
Firstly, control of the organization is seperate from the elected officials (board members), whereas in a democratically elected government, budgetary control is in the hands of the elected members.
Secondly, voting is, in the vast majority of cases, done by "proxy" to the management team, which selects the board members it wants. Shareholder activism is only present when major institutional shareholders decide they want a change in policy, or a raider comes in to raise the share price.
This is very far from responsible government.
Having said all of this, it's unlikely that democracy is an appropriate organizating structure for a business. Certainly, there is a role for employee ownership and for collective action, etc. But businesses exist to create markets and increase society's wealth, neither of which has anything to do with the level of freedom in how its policies are decided. Democracy doesn't automatically generate good decisions.
Democracy isn't necessarily a good thing at all times, as it can actually inhibit freedom if it is not counter-balanced. "Tyranny of the majority", for example.
and any government defying its principles is deficient, if not questionably moral
This is way too stringent. Firstly, what princpals are you refering to? There are many, not completely compatible, views on what a democracy really is.
Secondly, democracy doesn't necessarily product good or moral decisions. A democratic & free organization of humans can decide to do some horrendous things, such as kill other people they don't like.
then why does the same not hold true for corporations?
A managed organization (it's largely irrelevant if it's a corporation) is a pluralist entity within society; each has a certain role in society. Businesses, for example, exist to create markets and to increase the productive wealth of society through those markets (i.e. generate a profit).
Since democracy doesn't necessarily make "good" decisions (i.e. create products or services people need, or do so profitably), it doesn't make sense for the organization of a business to be democratic.
The problem with business, unfortunately, is that lots of bad theories and flawed laws about shareholder value and profit motive have led many businesspeople to forget that, basically, they're leaders in social institutions. And while they're not responsible for the "common good" per se, they at least must be responsible for their impacts, particularly on communities and the environment.
CEOs are just little Maoist dictators at heart. They share more with the reality of the Chinese rulers than they do with you, me or Thomas Paine.
The above statement has no basis in the reality of a modern large business, nor do modern large businesses have any reasonable resemblance to Mao Zedong's principles or tactics.
Family isn't important to everybody though; a lot of it depends on your background. "Home is where the hate is" isn't too far off the mark for some. In such cases one creates their own support network through friends & collegues.
It also depends on whether you have kids. Increasingly I find affluent couples are both working and are choosing not to have them; thus "family" really isn't a motivation.
This issue isn't about elitism or hypocrisy, it's that you rarely play WoW for 4 hours -- that's a "minimal" session. No other form of computer game that I've seen has the power to suck your time away. The sheer amount of life that game sucks away from you is sickening when you really start to think about it (90+ hours to reach level 60).
I would argue - and so would the Mono team - that.NET is more open as a platform than J2EE.
I suppose that depends on what you mean by "open". There are many more choices of application server with J2EE than there are with.NET, which arguably means it is more "open". Some specifications have been extremely successful at creating many options (JDBC, JMS, JCA, etc.). Whereas many parts of.NET aren't even standardized (ASP.NET, ADO.NET), and there are only three major implementations that I've seen, though perhaps I'm behind the times (I'm referring to MIcrosoft's closed branch, the Rotor branch, and Mono).
My argument is that without the economies of scale made possible by the Internet, those individual accomplishments would never have reached the point that they have now
Absolutely true. One could even argue that without the captial influx from both public and private capital (VCs & public companies), most of the popular open source wouldn't be where it's at. Most full-time contributors are on corporate payroll to write OSS, which is being funded either through complementary products (hardware, consulting, support) or is just a capital sink until they figure out how to make money with it.
In general, your article is thought provoking, though I disagree on some of the points of how people perceive the role of ideology in OSS. I'm going to draw a strawman scenario here to illustrate my view of how OSS mindshare grows, hopefully it's not too far from reality:
The OSS Hype Cycle
There is no core group of ideologues that really matters anymore. Perens and ESR did good things to hype OSS in the late 1990's, but I don't think they're doing much now to increase its hype. Today, the hype cycle is fed by a large group of in-the-trenches developers that are ideologues because their don't get much personal value out of their jobs and are trying to attach themselves to a larger cause. They're frustrated with the proprietary software they're forced to use that just doesn't work the way they want it to (regardless whether their way is actually better). This leads mostly to pro-OSS postings on blogs and websites, like Slashdot, TheServerSide.com, O'Reilly Network, or whatnot.
These posts, along with their voice on projects, eventually leads to influence thought leaders inside and outside their company, looking for the next trend to exploit. Joe Developer will promote the OSS-solution-du-jour for their project, and explain its wonders to his team leads and the public, mostly based on cool-factor and some anecdotal statements about its productivity. Examples abound, such Ruby on Rails, or MySQL + PHP, or the plethora of Java frameworks.
Comment: I'm not challenging that these tools actually make life better at times, but I am concerned with two things: the influence is usually based purely from a narrow "professional lens" -- I'm a developer, I only care about developer values, and I choose tools that make me feel more productive or cool, regardless of consequences outside my area of expertise. Business factors (which often are also architectural factors) are rarely considered. Secondly, that there is such chaos and splintering in the market going on due to OSS development that qualitiy is suffering. People are going "meta" and developing more and more tools for themselves instead of using old, proven tools that have lost the cool-factor, or might be proprietary.
To continue the story, these in-the-trenches IT or ISV developers influence their team leads, who, in smaller companies with less bureaucratic oversight on licensing / legal concerns, influence their directors, and open soruce gets used on a project. Successes are bound to occur, especially if the requirements are modest, and performance demands are light, and availability requirements loose. Pundits and bloggers pick up on these modest successes and run with it, claiming that all infrastructure software
It's very common for a Chairman / majority shareholder, or even maybe CEO to use personal pronouns when talking about business. I don't think it's glaring at all, but then again, this is Slashdot....
An interesting analysis, but I'm not sure I agree with your lack of faith in currency markets. Oh, the US dollar is going to have problems, I have no doubt, and gold is a decent place to park your wealth, but a lot of the problems with the dollar are already priced in.
Greenspan prints new money and sends it out at 1-4% interest. The lemmings get this money as increased wages and easy loans. The lemming sees this money at 4% can make him 9% in the market, so he invests there. Millions of other lemmings do the same, causing a supply shortage based on the increased demand that the Federal Reserve created -- causing the stock prices to go up. Artificial price increase.
In the late 1990's bubble period, I would agree, however now I believe most of the increase is generally due to institutional investment (pensions, some mutual, etc.)...
Houses are going up 10% annually! Said lemmings then invest in the housing markets, hoping to be the "flip it!" millionaire by 40. Except houses are depreciating assets, just like cars and clothes and food. Houses need maintenance, new houses replace old ones -- houses go down in price in a free market. Why did our houses go up? Easy credit combined with ridiculously low interest rates.
People were burned in the 1980's by the house market collapsing... If they didn't learn then....
Gold and silver are true stores of wealth above and beyond bank accounts -- I can easily liquidate gold in an emergency, but when it comes to impulse buying, I don't do it. In order for me to make an impulse purchase I have to take the step to sell my gold (every town has a pawn shop that offers up to 98% spot for gold), and by the time I care to cash in my wealth store, the impulse is over.
But this is akin to the old arguments that a floating currency is bad, and we should return to the gold standard --it's a debate on what bedrock of faith the market resides upon -- faith on the system itself vs. desire for shiny objects. The main argument for currency is its liquid nature vs. gold, but as you say, if you don't really need liquidity, more power to ya....
Are you actually suggesting that the VM is writing memory pages to swap disk? Egad. I've not heard this as a problem with the parallel collector; typically I lock the VM to shared memory (on UNIX platforms) to ensure it won't actually hit the swap...
As for the BEA JRockit license, good point -- you sound like an ISV, thus there probably would need to be some sort of arrangement with BEA on it.
Generally I've found that you can tune the Sun Hotspot VM's generations to minimize pause times, though this requires a GC analyzer to determine what's taking so long. Ensuring objects stay in the eden generation, for example, ensures they get the faster parallel garbage collector -- the "stop the world" collector only runs on older generations. I've run heaps with a couple of gigs of RAM and it's pretty rare to see it pause for several minutes -- a few seconds, sure, but that depends on the application.
Another thought, have you tried BEA JRockit? It has both a concurrent generational and parallel garbage collector which you can swap at runtime, and also has a dynamic GC setting to optimize for "minimal pause times" (-Xgcprio:pausetime).
Unfortunately, it's only on x86, EM64T, AMD64, and Itanium for now (Windows + Linux), but a Solaris AMD64 and SPARC version is due in the coming weeks. It's free for general use, but requries a license if you want to use the management features / GUIs.
Really, do you know what you're talking about? Every language in.NET has the same "features" if you're referring to the fact that they all have access to the.NET standard assemblies + Windows API, etc and that they are all (obstensibly) Turing complete.
Err, the point is that common language interoperability is based on a certain subset of language features that tend to be limited by C#'s features, which seems to be where Microsoft's mroe innovative language features are showing up.
You can absolutely use COBOL or Eiffel or whatnot, but the language constructs for talking across assemblies are limited in expressiveness based on the CLI spec. I generally think this tends to be "enough" for most practical purposes.
The other thing to note is the number of customers actually using COBOL or Eiffel on.NET for commercial applications are absolutely miniscule.
It's not a big lie, but it's not perfect. And it's arguably more pervasive that other so-called "cross platform UI toolkits", which if you'll recall in the late 1990's, were pretty rare...
It's interesting, the amount of hype being generated really depends on who you talk to. In some people's worlds,.NET is a complete non-player. In other people's worlds, it's all that matters; Java is a fading memory (and what is this weird Ruby thing)?
Based on what I see in financial institutions, I agree that many IT managers have latched onto.NET as the next "silver bullet", sometimes helped along by Microsoft doing CIO or CEO-level pitches. But I find architecture groups (good ones, anyway) are a bit more level headed, they see Java remaining important, particularly with the advent of these lightweight containers. We've spent 10 years as an industry building a marketplace for standard tooling, utilties, infrastructure and support for Java, it would be a shame to start re-inventing the wheel again for another platform. I would rather see new languages adapted to run on the existing Java infrastructure (the VM, the API's that work, etc.)
Sure, Microsoft is big enough to build a.NET infrastructure ecosystem single-handedly, but I'm not sure the software industry can afford to do a "rebuild" at this point, which was largely justified by Y2K and the rise of the Internet. Java's timing was absolutely imeccable to get the most impact at a time where things were changing the most. We may, of course, have a similar era of change in the coming decades, but I think it may be a while before Java's success / adoption cycle is replicated.
Java evolved out of C++, which also has a gargantuan spec, and C, which also is spec-based (at least the K&R book is something you'll need to read). And you need to understand the gargantuan operating system libraries to get anything done, which are different on UNIX vs. Windows vs. Mac OS X. And there are no standard UI libraries across platform, though you have a number of choices...
Java is more productive than C or C++ were, which is what its creators really wanted it to be. It's not a scripting language, which has a different value system.
Lots of people like to point out flaws in the corporate form of organization, and there are many areas of law that should be reformed / cleaned up in North America. In particular, we need to fix the relationship between management & shareholders -- it ignores the reality that shareholders have largely abandoned responsibility as "owners", but also tends to enshrine in law the idea that management can make unethical decisions in support of shareholder interests.
Having said that, the broader problem is the not the veil of the corporation, it's the veil of the organization. Many organizations that commit unethical or evil acts are private partnerships or syndicates or even loose collectives. Going back in history, we see atrocities committed in the name of many organizations: the nation, the state, the church, the race, the tribe, the interest group. Humans commit evil acts and use organizations as their tool. This is not to say that organizations are bad and we should all become anarchists (interesting idea in theory, as it may be).
My point is two-fold: all human organizations sow the seeds of their own destruction. They can be reformed and renewed, but without adapting to its external environment it will die, through irrelevance or overthrow. Secondly, revolution rarely works as intended; a revolt doesn't necessarily mean the new order will be better than the old order. More likely, it's "Here comes the New Boss, Same As The Old Boss".
The author may be very bright and have great insights in procedural programming, but he simply does not understand data modeling, or it was not taught to him in school, or he chose to ignore it when it was.
I've been waiting all thread for someone to finally point this out. RoR is one of the most productive things to come out of the web programming community in a while, but its creator and community are doing a lot of damage to a very important discipline.
The Rails community, containing many data management novices eager to avoid having to *think* (this is different than doing less work!), follows right along. Some of the stuff in the Rails mailing lists and blogs is quite silly and flies in the face of data management best practices.
In fact, many people on the blogs are starting to say that there ARE no data management best practices, as data integrity is bunk, and the database shouldn't manage the meaning of the data. They point to the the whole integration database vs. application database distinction, but that ignores the risks you're taking when you create poor data structures for the sake of programmatic simplicity. And then there's the bigger observation that databases have been the most proven way to integrate systems over the past few decades; messaging works too but is much newer and arguably harder to get working properly.
Then again, I'm an SOA advocate in my day job, so arguably this is where things are heading anyway. I'm around a lot of large organizations in recent years, and there are not very many "web + database" projects anymore; they've built them all. Now it's about integrating these silos as services, to improve the quality of data in the enterprise, consistency in experience, flexibility in business process, etc.. Whereas RoR just seems content with creating yet another layer....
The primitive / wrapper hassle is a minor annoyance, at best. And like.NET, there's the "autoboxing" feature in Java 5 that makes this more or less go away, though you may have to understand its own quirks (i.e. you get consistency at the expense of easy predictability).
No it doesn't. It means the percentage of the installed base.
Sez you. Marketshare is the share of market, which implies money. Installed base is a fundamentally different measure. Arguably installed base means something is more influential / has more mindshare. But it doesn't say anything about money exchanging hands.
I don't know if it's appropriate or not but that's the way it's always been measured.
No, actually. Plus it is hard to point to an accurate way of measuring the install base. Publication surveys tend to have a statistically skewed population and thus are very inaccurate. Basing things in terms of revenue is also inaccurate (shelfware is pervasive) but tends to at least be quantifiable and verifiable.
1) The majority of java applications are running on free software. When you add tomcat + spring + jboss + JONAS etc this number probably dwarfs the installed base of websphere, BEA or Oracle
I would qualify this. The majority of Java applications in production, supporting commerical entities with overall revenues of $10 million or more are running BEA or IBM. The rest probably are running free software or something really cheap. My evidence for this assertion? Well statistically there was the Middleware Company's / TheServerSide.com's 2004 J2EE survey where BEA and IBM's actual production installed share according to the survey was much larger than JBoss.
And anecdotally, I've been a consultant internationally for over 6 years, I've worked at several major wall street firms, two international banks situated in Japan, several major telcos, a couple of publishers, some e-commerce, the U.S. Navy, different branches of the Government of Canada and 3 provinces, and for 3 large systems integrators, and in my eyes and my collegues eyes, you see IBM a bit more than BEA and Oracle catching on quick. JBoss is in pockets, but mostly for development work and only if they're using EJB's. Tomcat on the other hand is pervasive.
And I have never, in my entire career of consulting, nor have I ever heard of anyone seeing a production application that involves financial transactions use an open source JMS product. It probably exists, but I would never recommend it.
2) Corporations frequently get the application server thrown in for free when they sign support contract bundles from the likes of IBM and ORACLE (and to much lesser degree BEA).
Given that IBM is still #1 in terms of marketshare according to IDC and Gartner, and that's measured by revenue, they're certainly not giving it away for free. Heavily discounted, and bundled, sure. But they make money on this stuff. THough if you scratch at the figures (again, marketshare according to IDC or Gartner - which measure share of market, not installed base), you see that IBM basically does give away WebSphere on AIX and Windows platforms but charges out the leg for Mainframe platforms. Which is likely why BEA has the marketshare lead on UNIX/Linux.
Oracle doesn't give away the farm except in an enterprise license situation where it's a burn-down or an all-you-can-eat, which isn't really a give away, it's a discount.
3) Corporation are free to choose the open source containers because the open source containers have proven themselves to be "enterprise ready". They are deprived of free and enterprise ready application server if they choose the.NET development stack.
Disagree. JBoss is not, in my opinion, "enterprise ready". Neither is MySQL. Linux and Apache are. JBoss is free if you want an unsupported application server.. That's it. Most of the install base that you cite don't buy support contracts from JBoss, Inc. But if you do, it's pretty affordable. But then again, IBM's low-end offerings and BEA WebLogic Express are also price competitive with JBoss, and contrary to JBoss' rhetoric have very good support quality.
And I see.NET deployed a tremendous amount more than JBoss..NET as a whole is becoming quite pervasive. Developers that use it tend to love it. If there's a dark horse that's "free" that's going to kill BEA and IBM, it would be this....
So it's innacurate to say the J2EE market is dominated for expensive app servers, those app servers are actually free or close to free.
Considering it's a 50/50 split between BEA and IBM across the board in most production systems, with Oracle catching onto third, and BEA doesn't sell a complementary product, I guarantee you BEA's software is not remotely close to free.
That's why I refered you to javalobby or magazine surveys. They tend to more representitive of the java development ecosystem as a whole.
I challenge that. WTF is an "ecosystem" and why does it matter? Marketshare implies a market. Marketshare traditionally means "the proportion of revenue generated in a market segment by one organization". By that definition, JBoss is a laggard -- they have 100 employees, obviously they're doing something right, but IBM's WebSphere application server division + global services have a cast of thousands, and BEA has around 4000 employees. BEA is still growing (albiet slowly) and takes in over $1.1B dollars annually, mostly on products based on WebLogic Server.
Is it really appropriate to measure marketshare in terms of number of installs? That's certainly harder to measure (most installs don't necessarily imply the software is "doing" anything). Apache is an exception because of surveys like netcraft, that can actually count external sites, that show the actual "web server marketplace" is tiny, since the top 2 (Apache and IIS) are essentially free. JBoss, Inc. exists to make money, so naturally they love a mass install base as a potential swath of support contract opportunities, but that's the only real benefit.
Furthermore, I would say the most installed Java server by far is Tomcat. JBoss isn't a web server / container. Most J2EE applications are not made with EJB these days. There's less reason to choose JBoss when you can go Spring + Tomcat, or Spring + Resin or even BEA WebLogic Express. And I wouldn't choose JBoss for its stellar JMS implementation, I'd rather go with IBM or BEA, or even a small/fast player like Fiorano.
if anyone actually listened to your recommendation, the Internet would still be an academic curiosity and we'd all be vegging out to shitty interactive TV.
There's a saying.. any programming paradigm that passes a certain threshold of mathematical reasoning will remain a niche. HTML succeeded because it was dumb.
Slashdot Mirror
Even though I never had a formal course in computer science I can solve problems none of the computer science people I work with can touch because of the math I can do.
I've often said that software development comes down to two polar opposites: at one, it's all about people. The other, it's all about computer science.
Math is the universal key.
It's an important key, but it's certainly not the only one. One can be just as effective in life and career if they become a master of understanding people. That's especially true for business software development.
As someone who has been working professionally with C# for years, and who has found it to be a much faster way to deliver quality production code, I contend that C# is the true benchmark: Java has already been superceded.
.NET monstrosities already; the platform is so new (only 4 years since the 1.0 GA) that they haven't really trickled out to the public to the extent that Java failure have after over 10 years of use.
C# is a good language, and for many types of rich Windows applications I agree it is more productive approach than Java + Swing; the event and delegate model is very intuitive.
Having said that, Java is also very productive; a lot of the problems have been the era in which older applications were written. People didn't know a lot, and abused the language. leading to monstrosities.
ASP.NET is also a very productive approach to web sites, but it isn't really significantly productive than Java + Struts, or JSF.
I've seen a few
Java's not that much of a mess, it's just a big thing that targets lots of people, and so there's a lot to it.
Also note, XDoclet isn't a part of the Java platform, it's a community project. There's plenty of examples of that in any language/platform... CPAN for Perl, Ruby gems for Ruby, etc...
Deployment descriptors aren't broken, the issue is that they don't recognize that developers sometimes make policy decisions at the early parts of a project, so it make sense to keep that declarative stuff close to the code. Yet you still need the external XML file to override the declarative policy, as needed. EJB 3 adopts this approach....
But it's typically a complicated relationship. Something that performs very well with one user could very well be scalable; the issue tends to be the amount of indirection one goes through to ensure scalability.
Put another way, performance can be had by maximizing the usage of available resources. Scalability under concurrent load is achieved by minimizing resource usage, especially resources that can only be used by one user at a time.
Good software that exhibits "scalable performance" is about knowing what approach to pick, and when.
$16B+ in revenues annually. That's 4x eBay's current revenue and 3.5x Google's current revenue.
eBay's market cap is slightly higher at the moment, and Google's is in the stratosphere (does it really deserve to be valued more than IBM??)
Having said that, eBay and Google are more profitable, and Google is growing revenue faster. Apple's current 60%+ quarterly growth needs a new hit product to sustain.
Armed conflict is power of might over right. And minority over majority. It's , by nature, anti-democratic. And , as far as I can tell, a majority of Americans picked this administration - twice. So what you're basically saying is, if I don't agree with everyone else, I'll just have to kill them to get my way.
My point: There have been lots of bloody revolutions, most of which left things in worse shape than before. Oh, sure, they were one step to an eventual "better world", but I don't think you'd like to live in France during Robespierre, nor a businessman in Russia during the Bolshevik revolution. Speaking of that one, Russia arguably is still looking for that "better world"; I don't think it's a stretch to argue that particular revolution was for the worse.
My second point - lots of regimes have fallen in recent years due to non-violent resistance, which also seems to be very effective.
Your analogy does not hold. Corporate capitalism is not at all like democracy in practice. Shareholders have long ago abdicated responsibility for corporate actions.
Firstly, control of the organization is seperate from the elected officials (board members), whereas in a democratically elected government, budgetary control is in the hands of the elected members.
Secondly, voting is, in the vast majority of cases, done by "proxy" to the management team, which selects the board members it wants. Shareholder activism is only present when major institutional shareholders decide they want a change in policy, or a raider comes in to raise the share price.
This is very far from responsible government.
Having said all of this, it's unlikely that democracy is an appropriate organizating structure for a business. Certainly, there is a role for employee ownership and for collective action, etc. But businesses exist to create markets and increase society's wealth, neither of which has anything to do with the level of freedom in how its policies are decided. Democracy doesn't automatically generate good decisions.
If Democracy is supposed to be such a good thing
Democracy isn't necessarily a good thing at all times, as it can actually inhibit freedom if it is not counter-balanced. "Tyranny of the majority", for example.
and any government defying its principles is deficient, if not questionably moral
This is way too stringent. Firstly, what princpals are you refering to? There are many, not completely compatible, views on what a democracy really is.
Secondly, democracy doesn't necessarily product good or moral decisions. A democratic & free organization of humans can decide to do some horrendous things, such as kill other people they don't like.
then why does the same not hold true for corporations?
A managed organization (it's largely irrelevant if it's a corporation) is a pluralist entity within society; each has a certain role in society. Businesses, for example, exist to create markets and to increase the productive wealth of society through those markets (i.e. generate a profit).
Since democracy doesn't necessarily make "good" decisions (i.e. create products or services people need, or do so profitably), it doesn't make sense for the organization of a business to be democratic.
The problem with business, unfortunately, is that lots of bad theories and flawed laws about shareholder value and profit motive have led many businesspeople to forget that, basically, they're leaders in social institutions. And while they're not responsible for the "common good" per se, they at least must be responsible for their impacts, particularly on communities and the environment.
CEOs are just little Maoist dictators at heart. They share more with the reality of the Chinese rulers than they do with you, me or Thomas Paine.
The above statement has no basis in the reality of a modern large business, nor do modern large businesses have any reasonable resemblance to Mao Zedong's principles or tactics.
Family isn't important to everybody though; a lot of it depends on your background. "Home is where the hate is" isn't too far off the mark for some. In such cases one creates their own support network through friends & collegues.
It also depends on whether you have kids. Increasingly I find affluent couples are both working and are choosing not to have them; thus "family" really isn't a motivation.
This issue isn't about elitism or hypocrisy, it's that you rarely play WoW for 4 hours -- that's a "minimal" session.
No other form of computer game that I've seen has the power to suck your time away. The sheer amount of life that game sucks away from you is sickening when you really start to think about it (90+ hours to reach level 60).
I would argue - and so would the Mono team - that .NET is more open as a platform than J2EE.
.NET, which arguably means it is more "open". Some specifications have been extremely successful at creating many options (JDBC, JMS, JCA, etc.). Whereas many parts of .NET aren't even standardized (ASP.NET, ADO.NET), and there are only three major implementations that I've seen, though perhaps I'm behind the times (I'm referring to MIcrosoft's closed branch, the Rotor branch, and Mono).
I suppose that depends on what you mean by "open". There are many more choices of application server with J2EE than there are with
My argument is that without the economies of scale made possible by the Internet, those individual accomplishments would never have reached the point that they have now
Absolutely true. One could even argue that without the captial influx from both public and private capital (VCs & public companies), most of the popular open source wouldn't be where it's at. Most full-time contributors are on corporate payroll to write OSS, which is being funded either through complementary products (hardware, consulting, support) or is just a capital sink until they figure out how to make money with it.
In general, your article is thought provoking, though I disagree on some of the points of how people perceive the role of ideology in OSS. I'm going to draw a strawman scenario here to illustrate my view of how OSS mindshare grows, hopefully it's not too far from reality:
The OSS Hype Cycle
There is no core group of ideologues that really matters anymore. Perens and ESR did good things to hype OSS in the late 1990's, but I don't think they're doing much now to increase its hype. Today, the hype cycle is fed by a large group of in-the-trenches developers that are ideologues because their don't get much personal value out of their jobs and are trying to attach themselves to a larger cause. They're frustrated with the proprietary software they're forced to use that just doesn't work the way they want it to (regardless whether their way is actually better). This leads mostly to pro-OSS postings on blogs and websites, like Slashdot, TheServerSide.com, O'Reilly Network, or whatnot.
These posts, along with their voice on projects, eventually leads to influence thought leaders inside and outside their company, looking for the next trend to exploit. Joe Developer will promote the OSS-solution-du-jour for their project, and explain its wonders to his team leads and the public, mostly based on cool-factor and some anecdotal statements about its productivity. Examples abound, such Ruby on Rails, or MySQL + PHP, or the plethora of Java frameworks.
Comment: I'm not challenging that these tools actually make life better at times, but I am concerned with two things: the influence is usually based purely from a narrow "professional lens" -- I'm a developer, I only care about developer values, and I choose tools that make me feel more productive or cool, regardless of consequences outside my area of expertise. Business factors (which often are also architectural factors) are rarely considered. Secondly, that there is such chaos and splintering in the market going on due to OSS development that qualitiy is suffering. People are going "meta" and developing more and more tools for themselves instead of using old, proven tools that have lost the cool-factor, or might be proprietary.
To continue the story, these in-the-trenches IT or ISV developers influence their team leads, who, in smaller companies with less bureaucratic oversight on licensing / legal concerns, influence their directors, and open soruce gets used on a project. Successes are bound to occur, especially if the requirements are modest, and performance demands are light, and availability requirements loose. Pundits and bloggers pick up on these modest successes and run with it, claiming that all infrastructure software
It's very common for a Chairman / majority shareholder, or even maybe CEO to use personal pronouns when talking about business. I don't think it's glaring at all, but then again, this is Slashdot....
An interesting analysis, but I'm not sure I agree with your lack of faith in currency markets. Oh, the US dollar is going to have problems, I have no doubt, and gold is a decent place to park your wealth, but a lot of the problems with the dollar are already priced in.
Greenspan prints new money and sends it out at 1-4% interest. The lemmings get this money as increased wages and easy loans. The lemming sees this money at 4% can make him 9% in the market, so he invests there. Millions of other lemmings do the same, causing a supply shortage based on the increased demand that the Federal Reserve created -- causing the stock prices to go up. Artificial price increase.
In the late 1990's bubble period, I would agree, however now I believe most of the increase is generally due to institutional investment (pensions, some mutual, etc.)...
Houses are going up 10% annually! Said lemmings then invest in the housing markets, hoping to be the "flip it!" millionaire by 40. Except houses are depreciating assets, just like cars and clothes and food. Houses need maintenance, new houses replace old ones -- houses go down in price in a free market. Why did our houses go up? Easy credit combined with ridiculously low interest rates.
People were burned in the 1980's by the house market collapsing... If they didn't learn then....
Gold and silver are true stores of wealth above and beyond bank accounts -- I can easily liquidate gold in an emergency, but when it comes to impulse buying, I don't do it. In order for me to make an impulse purchase I have to take the step to sell my gold (every town has a pawn shop that offers up to 98% spot for gold), and by the time I care to cash in my wealth store, the impulse is over.
But this is akin to the old arguments that a floating currency is bad, and we should return to the gold standard --it's a debate on what bedrock of faith the market resides upon -- faith on the system itself vs. desire for shiny objects. The main argument for currency is its liquid nature vs. gold, but as you say, if you don't really need liquidity, more power to ya....
Are you actually suggesting that the VM is writing memory pages to swap disk? Egad. I've not heard this as a problem with the parallel collector; typically I lock the VM to shared memory (on UNIX platforms) to ensure it won't actually hit the swap...
As for the BEA JRockit license, good point -- you sound like an ISV, thus there probably would need to be some sort of arrangement with BEA on it.
Generally I've found that you can tune the Sun Hotspot VM's generations to minimize pause times, though this requires a GC analyzer to determine what's taking so long. Ensuring objects stay in the eden generation, for example, ensures they get the faster parallel garbage collector -- the "stop the world" collector only runs on older generations. I've run heaps with a couple of gigs of RAM and it's pretty rare to see it pause for several minutes -- a few seconds, sure, but that depends on the application.
Another thought, have you tried BEA JRockit? It has both a concurrent generational and parallel garbage collector which you can swap at runtime, and also has a dynamic GC setting to optimize for "minimal pause times" (-Xgcprio:pausetime).
Unfortunately, it's only on x86, EM64T, AMD64, and Itanium for now (Windows + Linux), but a Solaris AMD64 and SPARC version is due in the coming weeks. It's free for general use, but requries a license if you want to use the management features / GUIs.
Really, do you know what you're talking about? Every language in .NET has the same "features" if you're referring to the fact that they all have access to the .NET standard assemblies + Windows API, etc and that they are all (obstensibly) Turing complete.
.NET for commercial applications are absolutely miniscule.
Err, the point is that common language interoperability is based on a certain subset of language features that tend to be limited by C#'s features, which seems to be where Microsoft's mroe innovative language features are showing up.
You can absolutely use COBOL or Eiffel or whatnot, but the language constructs for talking across assemblies are limited in expressiveness based on the CLI spec. I generally think this tends to be "enough" for most practical purposes.
The other thing to note is the number of customers actually using COBOL or Eiffel on
The write once, run anywhere "Promise" is a big lie. Try to write a swing app and run it on windows, linux and OSX to see what i mean.
Azeurus
IntelliJ IDEA
And while not Swing-based... Eclipse??
It's not a big lie, but it's not perfect. And it's arguably more pervasive that other so-called "cross platform UI toolkits", which if you'll recall in the late 1990's, were pretty rare...
It's interesting, the amount of hype being generated really depends on who you talk to. In some people's worlds, .NET is a complete non-player. In other people's worlds, it's all that matters; Java is a fading memory (and what is this weird Ruby thing)?
.NET as the next "silver bullet", sometimes helped along by Microsoft doing CIO or CEO-level pitches. But I find architecture groups (good ones, anyway) are a bit more level headed, they see Java remaining important, particularly with the advent of these lightweight containers. We've spent 10 years as an industry building a marketplace for standard tooling, utilties, infrastructure and support for Java, it would be a shame to start re-inventing the wheel again for another platform. I would rather see new languages adapted to run on the existing Java infrastructure (the VM, the API's that work, etc.)
.NET infrastructure ecosystem single-handedly, but I'm not sure the software industry can afford to do a "rebuild" at this point, which was largely justified by Y2K and the rise of the Internet. Java's timing was absolutely imeccable to get the most impact at a time where things were changing the most. We may, of course, have a similar era of change in the coming decades, but I think it may be a while before Java's success / adoption cycle is replicated.
Based on what I see in financial institutions, I agree that many IT managers have latched onto
Sure, Microsoft is big enough to build a
Java evolved out of C++, which also has a gargantuan spec, and C, which also is spec-based (at least the K&R book is something you'll need to read). And you need to understand the gargantuan operating system libraries to get anything done, which are different on UNIX vs. Windows vs. Mac OS X. And there are no standard UI libraries across platform, though you have a number of choices...
Java is more productive than C or C++ were, which is what its creators really wanted it to be. It's not a scripting language, which has a different value system.
Lots of people like to point out flaws in the corporate form of organization, and there are many areas of law that should be reformed / cleaned up in North America. In particular, we need to fix the relationship between management & shareholders -- it ignores the reality that shareholders have largely abandoned responsibility as "owners", but also tends to enshrine in law the idea that management can make unethical decisions in support of shareholder interests.
Having said that, the broader problem is the not the veil of the corporation, it's the veil of the organization. Many organizations that commit unethical or evil acts are private partnerships or syndicates or even loose collectives. Going back in history, we see atrocities committed in the name of many organizations: the nation, the state, the church, the race, the tribe, the interest group. Humans commit evil acts and use organizations as their tool. This is not to say that organizations are bad and we should all become anarchists (interesting idea in theory, as it may be).
My point is two-fold: all human organizations sow the seeds of their own destruction. They can be reformed and renewed, but without adapting to its external environment it will die, through irrelevance or overthrow. Secondly, revolution rarely works as intended; a revolt doesn't necessarily mean the new order will be better than the old order. More likely, it's "Here comes the New Boss, Same As The Old Boss".
The author may be very bright and have great insights in procedural programming, but he simply does not understand data modeling, or it was not taught to him in school, or he chose to ignore it when it was.
I've been waiting all thread for someone to finally point this out. RoR is one of the most productive things to come out of the web programming community in a while, but its creator and community are doing a lot of damage to a very important discipline.
The Rails community, containing many data management novices eager to avoid having to *think* (this is different than doing less work!), follows right along. Some of the stuff in the Rails mailing lists and blogs is quite silly and flies in the face of data management best practices.
In fact, many people on the blogs are starting to say that there ARE no data management best practices, as data integrity is bunk, and the database shouldn't manage the meaning of the data. They point to the the whole integration database vs. application database distinction, but that ignores the risks you're taking when you create poor data structures for the sake of programmatic simplicity. And then there's the bigger observation that databases have been the most proven way to integrate systems over the past few decades; messaging works too but is much newer and arguably harder to get working properly.
Then again, I'm an SOA advocate in my day job, so arguably this is where things are heading anyway. I'm around a lot of large organizations in recent years, and there are not very many "web + database" projects anymore; they've built them all. Now it's about integrating these silos as services, to improve the quality of data in the enterprise, consistency in experience, flexibility in business process, etc.. Whereas RoR just seems content with creating yet another layer....
The primitive / wrapper hassle is a minor annoyance, at best. And like .NET, there's the "autoboxing" feature in Java 5 that makes this more or less go away, though you may have to understand its own quirks (i.e. you get consistency at the expense of easy predictability).
No it doesn't. It means the percentage of the installed base.
.NET development stack.
.NET deployed a tremendous amount more than JBoss. .NET as a whole is becoming quite pervasive. Developers that use it tend to love it. If there's a dark horse that's "free" that's going to kill BEA and IBM, it would be this....
Sez you. Marketshare is the share of market, which implies money. Installed base is a fundamentally different measure. Arguably installed base means something is more influential / has more mindshare. But it doesn't say anything about money exchanging hands.
I don't know if it's appropriate or not but that's the way it's always been measured.
No, actually. Plus it is hard to point to an accurate way of measuring the install base. Publication surveys tend to have a statistically skewed population and thus are very inaccurate. Basing things in terms of revenue is also inaccurate (shelfware is pervasive) but tends to at least be quantifiable and verifiable.
1) The majority of java applications are running on free software. When you add tomcat + spring + jboss + JONAS etc this number probably dwarfs the installed base of websphere, BEA or Oracle
I would qualify this. The majority of Java applications in production, supporting commerical entities with overall revenues of $10 million or more are running BEA or IBM. The rest probably are running free software or something really cheap. My evidence for this assertion? Well statistically there was the Middleware Company's / TheServerSide.com's 2004 J2EE survey where BEA and IBM's actual production installed share according to the survey was much larger than JBoss.
And anecdotally, I've been a consultant internationally for over 6 years, I've worked at several major wall street firms, two international banks situated in Japan, several major telcos, a couple of publishers, some e-commerce, the U.S. Navy, different branches of the Government of Canada and 3 provinces, and for 3 large systems integrators, and in my eyes and my collegues eyes, you see IBM a bit more than BEA and Oracle catching on quick. JBoss is in pockets, but mostly for development work and only if they're using EJB's. Tomcat on the other hand is pervasive.
And I have never, in my entire career of consulting, nor have I ever heard of anyone seeing a production application that involves financial transactions use an open source JMS product. It probably exists, but I would never recommend it.
2) Corporations frequently get the application server thrown in for free when they sign support contract bundles from the likes of IBM and ORACLE (and to much lesser degree BEA).
Given that IBM is still #1 in terms of marketshare according to IDC and Gartner, and that's measured by revenue, they're certainly not giving it away for free. Heavily discounted, and bundled, sure. But they make money on this stuff. THough if you scratch at the figures (again, marketshare according to IDC or Gartner - which measure share of market, not installed base), you see that IBM basically does give away WebSphere on AIX and Windows platforms but charges out the leg for Mainframe platforms. Which is likely why BEA has the marketshare lead on UNIX/Linux.
Oracle doesn't give away the farm except in an enterprise license situation where it's a burn-down or an all-you-can-eat, which isn't really a give away, it's a discount.
3) Corporation are free to choose the open source containers because the open source containers have proven themselves to be "enterprise ready". They are deprived of free and enterprise ready application server if they choose the
Disagree. JBoss is not, in my opinion, "enterprise ready". Neither is MySQL. Linux and Apache are. JBoss is free if you want an unsupported application server.. That's it. Most of the install base that you cite don't buy support contracts from JBoss, Inc. But if you do, it's pretty affordable. But then again, IBM's low-end offerings and BEA WebLogic Express are also price competitive with JBoss, and contrary to JBoss' rhetoric have very good support quality.
And I see
So it's innacurate to say the J2EE market is dominated for expensive app servers, those app servers are actually free or close to free.
Considering it's a 50/50 split between BEA and IBM across the board in most production systems, with Oracle catching onto third, and BEA doesn't sell a complementary product, I guarantee you BEA's software is not remotely close to free.
That's why I refered you to javalobby or magazine surveys. They tend to more representitive of the java development ecosystem as a whole.
I challenge that. WTF is an "ecosystem" and why does it matter? Marketshare implies a market. Marketshare traditionally means "the proportion of revenue generated in a market segment by one organization". By that definition, JBoss is a laggard -- they have 100 employees, obviously they're doing something right, but IBM's WebSphere application server division + global services have a cast of thousands, and BEA has around 4000 employees. BEA is still growing (albiet slowly) and takes in over $1.1B dollars annually, mostly on products based on WebLogic Server.
Is it really appropriate to measure marketshare in terms of number of installs? That's certainly harder to measure (most installs don't necessarily imply the software is "doing" anything). Apache is an exception because of surveys like netcraft, that can actually count external sites, that show the actual "web server marketplace" is tiny, since the top 2 (Apache and IIS) are essentially free. JBoss, Inc. exists to make money, so naturally they love a mass install base as a potential swath of support contract opportunities, but that's the only real benefit.
Furthermore, I would say the most installed Java server by far is Tomcat. JBoss isn't a web server / container. Most J2EE applications are not made with EJB these days. There's less reason to choose JBoss when you can go Spring + Tomcat, or Spring + Resin or even BEA WebLogic Express. And I wouldn't choose JBoss for its stellar JMS implementation, I'd rather go with IBM or BEA, or even a small/fast player like Fiorano.
if anyone actually listened to your recommendation, the Internet would still be an academic curiosity and we'd all be vegging out to shitty interactive TV.
.. any programming paradigm that passes a certain threshold of mathematical reasoning will remain a niche. HTML succeeded because it was dumb.
There's a saying