Now, to be honest, I really don’t care about that, it’s just digital detritus after all. But there are really creepy, nasty, vile corporate interests that do care a lot about it. And as much as I think they’re vile, immoral, creepy entities whose members will surely rot in hell, I have to let them, at least momentarily, invade my normally quite pure thoughts… and I gasp in horror at the thought of invertantly violating their corrupt business practices- er I mean copyright laws… and…

Anyway I killed the process midstream.

This is one of those things you don’t think about until something crappy like that happens. Normally a VCS process is simple- commit, comment, done! I really don’t think about it. It’s habit. So what happens when you think you’re committing 100 or so files and maybe that amounts to 1MB and suddenly whoops! And you kill the process?

Well, SVN really has trouble dealing with killed processes, obviously. But thankfully it cleans up after itself really well!

svn cleanup

That’s it. It’s easy. But I didn’t see an svn cleanup option in subclipse, so I did it from the command line.

Amazon is the best place to buy books, cds, electronics and other memorabilia. But what happens if you want your hands in all the pies?

After a sloppy launch of Kindle 1, a year ago, Amazon released Kindle 2, hoping that it would be the next best thing, something of a revolutionary device like IPhone. Dreams. Stephen King introduced the product at the launch; I was waiting for Kate Perry next, who would sing “I read a book, I sucked at it.”

Amazon — please respect people’s intellect. We are not buying acne creams.

Ok. Some features:

• No Touch screen.
• Larger screen? It’s smaller than a book. the screen is double that of your IPhone. 35-40% real estate wasted on the device. Just look at the picture once. Why? Because there’s a keyboard on it. Yes — A KEYBOARD.
• Too many buttons. There are seven buttons on the device. Remember Microsoft Vista? There were seven options too — logoff/sleep/hibernate/shutdown/whatever1/whatever2/hangMyself.
• Displays in B&W.
• Operating System? Inspite of being based on Linux, this still is one black box where you can read a book, one at a time. More cameras are going to have built in wifi, smart phones are becoming mini computers, cars can park themselves. If someone thinks they want to make a “revolutionary” product that will change the landscape of the world and there is no software to manage the device. What is the point of having an OS when you can’t have software running on it?
• No API for third party applications. Imagine the applications built around reading books, social networking, custom viewers, virtual bookshelves or even mini applications to store your favorite quotes from books. I expect a dedicated ebook reader (the one that claims “version 2″) to be smarter than just displaying books in black and white pages. If you had inbuilt wifi, and a browser, you could probably get “recommendations” or lookup a dictionary or even sort your library according to genre. It’s quite naive that Amazon limits the device to such a small scope. Imagine Playstation releasing PS4 and it only plays Sony games downloaded from Sony store and all the games are arcade games, single player. And they are 16-bit color.
• Stylus? Don’t you want to mark your favorite sentences? Or do you want to click menu->highlight->cursor down (12 times)->press spacebar (on the keyboard!) 17 times->press another button to “highlight.” Now, did I say, there was another button to save it?
• foldable lcd screen or flexible displays? No. It looks good for a 1999 ipod equivalent of an ebook reader.
• Doesn’t read pdfs or any other format.
• No Wifi or browser.
• Can you do anything apart from reading books (bought from amazon?) – No. Same reasons as above.
• Does it read books in your own voice? No. There’s a robotic voice which is worse than the “your call is important to us. Please hold, while ….”

Fine, this is a book reading device. I get it. But what’s special about it? What’s so “groundbreaking” about it that I should stop buying books and buy a Kindle? I haven’t seen a single Kindle in subways, libraries or Starbucks, in the last one year. The original Kindle sales are not disclosed to public, clearly seems like a catastrophic failure for a product that had hopes of “changing the way people read books, forever.”

It seems like a step back to the lalaland with outdated technology and mediocre features. For 350$ and seven buttons and a keyboard and a black n white display and lack of add-on software, its a bad investment one could ever make. Amazon raves about saving 5-10$ on a book, but never counts on the 10$, one would get when the same book is sold. We read books. We share books. We give it to our friends. We distribute them. Reading & sharing books is a culture. We need products that can enhance this experience and add more value, not redefine it for someone to be a monopoly. I can’t believe Amazon thinks that “Kindle” is still the future of reading … I just can’t. The worst strategy than Microsoft. They thought, that the whole world will “browse the web” in IE.

As for me, my admiration and honor goes for a little unknown company that developed Stanza for IPhone/IPod, for FREE. If only Amazon had 1% of their passion …

Some pics here.

However, some of you may benefit from what I have learned NOT to do over the past 2 years. I have learned plenty of “DO NOT’s” over time, but the last 2 years, especially the last 7 months has been the grand festival of “DO NOT’s”.

1. Do not let an “expert” with unsubstantiated credentials tell you how to do your job. Better yet, avoid working with an “expert” whenever possible. These “experts” like to talk down to you and blame you for everything that goes wrong. They never admit to contributing to the problem. They also insist that his/her approach is the best, even if God (or other supreme being) were to tell him/her that it’s wrong. An “expert” will try to control you and the project at any cost. An “expert” will no doubt run the project into the ground or into a persistent vegetative state. If you find yourself working with an “expert”, take precautionary measures and have an escape plan.

2. Do not let users perform their daily job duties with an application that is not fully developed or production ready, especially if it’s not even beta ready. Don’t let users rely on an application that is undergoing constant development and hasn’t even been tested. Users will no doubt work their dirty little hands into your project and get you to make changes at the drop of the hat, because their jobs depend on it! Don’t expect a manager to help you out in this situation. If things have reached this point, no manager can simply stop this evil cycle. Now you have a situation where there is a direct impact on the success of the company. Because the application is not even fully developed or tested, there will already be bugs galore. The users relying on your application will find even more bugs that stop them in their tracks. Everything becomes a critical defect. Trust me, if you get into this situation, you will want to die, because you are already dying a slow and painful death.

3. Do not develop in an environment that has absolutely no process. You don’t need to be process gung-ho and implementing CMMi level 5, but you should have some basic sort of SDLC process in place. A development group without processes usually fails to have standards as well. The “as long as it works” mantra comes into play.

In this situation, you will end up with horrible spaghetti code where developers are doing whatever the hell they want. You will have no ability to monitor code quality. You might end up with code that looks like it was written by someone who is half-way through their first programming course (CSCI 101). Hey, they just learned how to write arrays, and want to show off that skill, as much as they can. Your “code review” will take place when you have to take over their project, at which point you’ll ask yourself, “What kind of dumbass could possibly write this &*@%!@#!!!????” Of course by then the bad code has permeated its way into the deep inner bowels of the application, like a virus, making it nearly impossible to correct without throwing it all away and rewriting from scratch. It’s sort of like when Windows gets jacked-up every year or so and your best solution is to “format C:” and start over.

In addition, the extent of your QA will probably be limited to the success of your compiler and a few keystrokes and mouse clicks.

no process = bad applications (hopefully most of you know that already)

4. Do not get involved in a project where a phone call, simple face-to-face conversation, or a note on your desk from a user is all that is required for you to make a change. This is closely related to #3, but it’s worth emphasizing a bit. You need to have a way to document change requests and prioritize those requests, even if it a simple excel spreadsheet on a shared network drive. This is just one more way that you can be trampled to death by users. Not to mention the fact that you are developer and you have no business answering the phone or communicating with a user directly. That’s what BA’s are for. You lack the people skills required to interact directly with users and you should be damn proud of that. Additionally, there is great satisfaction to be had when you can make users punch themselves in the face. By that I mean something like this:

User: “What is this thing doing? That’s not the way it’s suppose to work!”
You: “Yes it is.”
User: “No, it’s not.”
You: “Yes it is, you requested it, see?”
User: “Ah DAMMIT!” (punches self in face)

5. Do not succumb to the “at least we still have jobs” outlook on your career. Yes, the economy sucks right now and layoffs suck. I hate seeing people bringing in boxes to pack up their stuff, especially when they have done a great job at the company for 10, 15, 20 years or more. That’s like saying, “Thanks for all your great years of service, now be a team player and piss off, yeah?” There are other things employers will do to cut costs though. They might squash benefits, slash salaries, and eliminate incentives. This is especially likely to happen if the executives made stupid decisions (likely in any case) that ran the company into the ground. You don’t have to pay for their mistakes. This industry is still viable. You can find good opportunities elsewhere. Don’t be afraid to explore and don’t feel bad for leaving. If you really believe in the company and feel that you want to help “get through this tough time together”, then stay and try to maintain a positive attitude. But remember, you don’t owe the company anything.

That’s it for now. Maybe you learned something, maybe not. You may have only learned that I am just an ass, if you didn’t know that already. Either way, I hope it wasn’t a boring read. I’d better get back to “work.”

Me                               : Can I get a room for two, tonight, near SF Airport?
Customer Service (CS:) : I’m sorry sir, but it’s past midnight, do you want me to book for the next day?
Me                               : I don’t know which watch you’re wearing but mine says 10:12pm.
CS:                               : But we book hotels only in Central time
Me                               : So … ?
CS                                : The “system” doesn’t allow us
Me (eureka)                  : No wonder, I didn’t get any results online
CS                                : (silent, probably checking her nails)
Me                               : What do you suggest?
CS                                : Call the hotel directly

So, I called the two other travel agencies, same story. Looks like they all use the same *** system. It’s shocking to see that this BUG probably exists and known by the agencies but they haven’t done anything to fix it. Dealing with timezones maybe a little tricky but bringing down the business in a 2 hour window, because a system that “ignores” user’s timezones, I think is lame. (shame)

PS: I did get some hotel, for 25% more price but that’s another story.

public interface Formatter <t extends Foo> {
    public List<list<string>> format(List<t> foos);
}

Class FooFormatter implements Formatter and is strongly typed

public class FooFormatter<foo> implements Formatter {
    public List<list<string>> format(List<foo> beans) {
        return null;
    }
}

When you compile the above (or see the red marks in your IDE), you get the following compiler error:-

FooFormatter is not abstract and does not override abstract method format(java.util.List) in Formatter

What is so wrong with that? Here’s the correct version:-

public class FooFormatter implements Formatter<foo> {
    public List<list<string>> format(List<foo> beans) {
        return null;
    }
}

Fair enough, I get it, but seriously — what does the previous compiler error mean?

“Strongly typed classes,” can be good at times, but the syntax of Java Generics is so complicated, it makes you wonder about the implementation of the same. Just think of instantiating a HashMap containing a set of Generic classes, which itself contains another Generic classes. The declaration and initialization is so long, it might well cross the 150 meter boundary in your IDE.

Here are some rants on why Java Generics sucks. One more here and a very useful Generic FAQS here

1. Reduced memory foot print? No.
2. Reduced distribution size? No.
   Java distro is 123MB and JEE with over 500 plugins is > 400 MB
3. Loads faster in memory? No.
   With a 3GB Ram, Intel core 2 duo 3.4 Ghz on a Macbook Pro, there isn’t a noticeable difference in performance. Strangely, it feels like the same product 4 years back, inspite of the computer configurations being quadrapled in this time-period.
4. Plugin manager seems improved with a Software Update option. (mimics IntelliJ Idea) but is not really that user friendly. For instance, download the jee distribution and there were a host of plugins that are not necessarily useful for all. The problem is, you cannot uninstall these, the button itself is disabled.
5. I don’t understand Plugin management. Let’s say, you need a rich facelets editor with live view. This is what you need to do for a standalone Java distro now. Download the entire JBoss Tools plgins (70MB). Download Web Developer plugins (~100MB) and then a EMF plugin (def > 50MB). Congratulations, the plugin works now with a “little” dependency on other plugins (~200MB).Try downloading java distro and add a Web Development Tools bundle, it comes with 15 plugins bundles and you have to download all of them. The best part is webtools doesn’t come with a single adapter by default. After downloading 100MB, it’s desirable to have atleast a tomcat plugin. But for that you have to download another 70 MB of Jee plugins. In web tools mode, downloading server adapters fails midway.It’s good to reusue java component libraries but to have a dependency that cannot work without loading an entire plugin suite — I don’t understand this architecture.
6. Why can’t we update eclipse from within eclipse? Do we need to completely erase existing one and reinstall a new one and then move all our workspaces, projects and test for the rest of the day? Talk about modular development.
7. Sometimes if one plugin breaks, it breaks the rest. For instance, code completion doesn’t work. Automatic imports fail because of a faulty plugin. Try this — with jboss tools plugin in the jee distribution, setup a hibernate project and open hibernate perspective. It crashes eclipse. Configur a JBoss Seam project, the server view crashes. Could not create the view: org/eclipse/wst/server/ui/internal/view/servers/ServerAction
8. The Data Tools Platform plugins is a joke. Does anyone use it in real world, compared to a quality standalone DB Query Tool (many available for free)? And now eclipse jee distribution forces us to use it in a jee distro. One cant even configure a url using service name (forces SID name).
9. There are several other redundant features like “Create a new Servlet” button.
10. Why aren’t SVN & Maven out of the box features and a first class citizens of eclipse? (after 4 yrs of releases.) There are third party distributions but they don’t work as desired. For instance I get this error when trying to connect to the new SVN plugin (not subclipse) SVN: ’0×00400006: Validate Repository Location’ operation finished with error: Selected SVN connector library is not available or cannot be loaded.
    If you selected native JavaHL connector, please check if binaries are available or install and select pure Java Subversion connector from the plug-in connectors update site.
    If connectors already installed then you can change the selected one at: Window->Preferences->Team->SVN->SVN Client.
    Selected SVN connector library is not available or cannot be loaded.
    If you selected native JavaHL connector, please check if binaries are available or install and select pure Java Subversion connector from the plug-in connectors update site.
    If connectors already installed then you can change the selected one at: Window->Preferences->Team->SVN->SVN Client.

11. Can’t connect to the Internet even after proxy settings. Is it really difficult to import proxy settings from system browser by default? I can’t get Help to work within the IDE. It keeps throwing — “There is no internet connection”, which can’t be true as the plugins were download via the same proxy setting.
12. It would be much more helpful to read Tips and Tricks & understand New features as compared to reading page no 134 in section 4.3 in a help document online.
13. Mylyn is a revolutionary way to manage tasks form within an IDE but lets face the reality — I haven’t seen or met one team or any of my past projects who actually use this. Every company (small or big) usually has it’s own S/W development model. This again is an example of why eclipse tries to force you to use some features which it “thinks” is standard. I don’t want Mylyn for now, but I cannot get rid of it either in 3.4. (3.3, I could atleast uninstall after a horrifying manual cascade delete of all the other dependencies).
14. Why is Plugin Dev Environment included by default?
15. With web development plugins, add JSF capabilities and try to remove. It doesnt remove.
16. JPA plugin doesn’t work. It completely messes up and throws compilation errors instead. Doesn’t identify theid generators (compiles fine on cmd line) Duplicate generator named “id-generator” defined in this persistence unit Comment.java
    xxx-ejbs/src/main/java/xxx/entity line 16 JPA Problem Marker.
    Even after getting it work, it seems pretty useless. Do we really need a plugin to add a “@NamedQuery” graphically?My worry is that some of the plugins promote the developer to becoming dumb rather than being productive or make them a better programmer
    
17. Tried to update WST tools plugin on eclipse java version. It hangs at 56%. Same with JEE tools plugin. It works after closing and reopening eclipse.
18. The Workspace setup is still painful sometimes, cannot nest folders. Why can’t it be intelligent enough to ignore it?
19. Got a bunch of random errors on using throughout the day: - java.util.concurrentModification on closing/opening a proj Ok, The good things:

1. breadcrumb navigator in the editor. This is the best feature introduced in 3.4. Simple yet powerful. Combining this with Alt+Shft+R or Alt+Shft+T, you would never need a package explorer view.
2. Fine grained formatting for annotations. Overall minor improvements for formatting are great!
3. Some new refactoring tools like “Extract class refactoring”, “quick assists” are cool.
4. I haven’t tried but Real Time shared editing is amazing.
5. Dynamic Language Toolkit is a nice extension
6. P2 Provising, the newly designed Runtime etc are cutting edge features but as a web developer — I doubt if it adds to the developer productivity

What I really want from eclipse …

Ideally this is what I would want from eclipse (which is not too much to ask considering it’s there already)

• Download a bare bones eclipse IDE with a distro size of < 50 MB. It should have *all* the java editor capabilities – refactoring + formatting + intelligent editing, code completion
• all the current java/server debugging facilities
• Just 2 Standard server adapters shipped by default — Tomcat/Jetty. And an easy extension plugin to other servers (JBoss, glassfish etc .. )
• a good XML/XSD plugin (able to edit and code complete)
• a good jsp,jsf,facelets,Xhtml plugin (with a possibility of live view)
• Ant, Maven — out of the box
• junit, testng out of the box
• cvs, svn and the existing Team Synchronizing out of the box
• Simplified workspace, project management. (try to run a helloworld in eclipse, it takes more steps than your vi editor)
• Of course, it should be a *Fast IDE”

For everything else, keep it as simple non-dependable plugins. You want Grails? Fine, add grails plugin.

No — don’t ask me to go buy myeclipse. It comes with *all the frameworks support in the world* and I don’t want them. Do you want to goto a buffet everyday? Or do you prefer a-la-carte?

I do respect Open Source and the Eclipse IDE foundaton, I’ve been using it for over 5 yrs now but I’m beginning to worry about it’s future inspite of it being free, as it seems to be going down the path of what Microsoft did with Vista — giving us features which we don’t need and assume a lot of things about developers, trying to satisfy *every kind of developer* on the planet; like using Microsoft Office 2008 when all you need for is a simple rich document with tables and formatting.

Simplicity is dying.

Lets go back to the 90s and early 2000s’, we all loved Winamp. It was simple, light weight and did the job. Open a file from your hard disk and play. Add dsp plugins or skins. Great! Now, that is simplicity. And then came these monsters — ITunes, Windows Media Player, Real Payer and took over the industry. Sadly Winamp died. The same, I’m afraid is happening with eclipse. With bloated Enterprise features (which don’t really add much to the productivity in my opinion), trying to compete with every vendor out there, adding features like reporting tools (BIRT?), it’s becoming insane everyday.

The idea of just working on a simple web application environment with a feature rich editor is dead now. I have to work with a 400MB monster to write a web app, which takes up 150MB of heap space average and lets accept it — it’s *slow*, makes me wonder what is my “Core 2 Duo processor with 256MB Graphics card” really adding value.

I want my old eclipse. Where is my a-la-carte?

[Update: 1 hours 30 minutes] Ok, they added a nicer error message. But click on the link, it gives back the same Http/1.1 Service Unavailable message back! What’s worst is — It does actually work sometimes, but when you refresh, you’re back to this screen. Hilarious stuff Amazon, keep us entertained.  Â

[Update: 1 hour] Â Amazon.com down for over one hour and counting. Tech Crunch.com features this on it’s home page at this point. Tech Dirt reports the same

Wow! You don’t see these messages often.

Reminds you of your old school webapp, when you first tried. If Amazon.com sells thousands of books books per hour (hundreds of thousands, in peak season), then how much are they loosing per millisecond, for this error message? Not to mention, the thousands of other stuff that they sell.

 It would be an interesting math.

Even consultants and recruiters alike, market profiles as “J2EE consultants” as if they have only worked with one framework and version (“2″) throughout their careers. Most developers today pretty much have experience with Struts, Spring, Hibernate, Ajax and a host of other smaller to larger framework APIs, a lot of them being open source. It’s surprising that the Java world hasn’t come up with an innovative name yet to call such a developer.

Here are a few suggestions: –
-JOS (Java Open Source) Developer
-JEM (Java Enterprise Middleware) Developer
-JUI (Java UI) Developer
…

The cold reality that software companies try to ignore is that ANY major piece of software needs to be rewritten from the ground up eventually. They age, get bit-rotted, are obsoleted by new technology and generally become less and less supportable. A ground-up rewrite preserves most of the perceived value of the code, because the existing system can serve as a template for how the hard things were solved the last time around.

I rest my case.