I'm a Mac user by choice that is currently typing this from Kubuntu 9.04 with Windows 7 executing in a background. I was impressed with Windows 7 until I tried to play with networking. It wasn't working initially (turned out to be a conflict with VirtualBox's default 10.0.2.x NAT range) but I needed to see how Windows was configured (DHCP, etc.). So I go to Control Panel -> Network and Internet -> View Network Status and Tasks -> Change Adapter Settings. Then I left-click on "Local Connection". Now, to get the IP and gateway info, I have to click "View Status of Connection". To view whether it's set up for DHCP check the DNS servers, I have to click "Change Settings of this Connection", right-click on "Internet Protocol Version 4 (TCP/IP)" and click "Properties".
By contrast, here is what I do in Mac OS X: Click the apple menu, click System Preferences. Click Network. There I'm presented with a list of all of my adapters. I click on one and I see everything I need.
In Kubuntu, I click the knetwork-manager applet and click "Manage Connections"
Now, which of the three is the most usable? Keep in mind that as I was troubleshooting the networking issue I had with my VM, I constantly had to repeat those steps. What is that, like 10 clicks? Look at the menu names... "View Network Status and Tasks", "Change Adapter Settings"... is this supposed to be intuitive?
Another thing, is that Windows pops up every time I jump on a network and asks me if it's a home network, work network, or public network, and initializes stuff for me (including home groups, which I don't want). Now, this is fantastic for end users, and a great feature. But as a power user running 3 levels of NAT at home (local net, work VPN, and NAT VMs) it is infuriating to have the details hidden from me and not know how to get to them.
I would definitely recommend Windows 7 to anyone who wants to plug something in and have it just work and be done with it (which is I think the point). But the configuration is hellish, so if you like to tinker, think twice.
I'm not well-versed in the distinction between netbooks and nettops, but I have seen $200-$300 devices that are bigger than phones and smaller than laptops with and without built in 3G support.
Laptops are approaching the price of desktops, and portability, and that's with comparable system specs. A lot of employees do little more than tinker with documents, spreadsheets, and E-mail/calendaring software, so there is no great need for horsepower. So in the office I see three types of devices: workstations (for developers/IT professionals); dumb terminals connected to Citrix servers (or similar) for stationary employees or non-personal stations (think call center); and laptops for managers, salespeople, and employees prone to frequent travel or telework.
It's called communism, and it fails because humans are lazy and greedy. If you can somehow persuade humans to stop being lazy and greedy, then pretty much any system of government will suffice.
It will happen. 5 years ago, a family had one central PC that they all had to fight over. Now they all sit in the living room with their individual smartphones and nettops that they can afford because they are subsidized by the carrier. They are cheaper and everyone has their own device with 24/7 access, and every two years you can upgrade cheap. More than enough for text, IM, E-mail, web browsing, money management, research and report writing, casual gaming, and FaceSpace. Heck they're selling a wireless printer/scanner/fax/copier at Target for less than $100. They have NAS devices wireless interfaces, and you can plug a USB 2.0 backup drive into those. Who needs a server or central PC?
You'll still have gamers and hobbyists purchasing PCs, but they will be more likely to be specialized high-end machines.
On the server side you'll still have big metal, whether it's some distributed cluster or mainframes of a sort. Got to serve the cloud.
For anyone just entering a CS/CIS program in college, smart money's on embedded programming or multi-core parallel programming on distributed/clustered systems. (And it won't be too long before someone starts sticking multiple cores in cell phones, if they haven't already...)
I have an Apple MacBook I use as my primary PC at home, and I love Apple products, but Java on Mac OS X sucks. It's slow, always at least 2 years behind the current version, and the Cocoa LAF make Swing apps look terrible.
I bought my MacBook with the intention of making it my primary development laptop (for school--we used a lot of esoteric languages and libraries), but I found Java next to unusable. I use Ubuntu at work on a Dell D830 and the the difference is night and day in terms of performance (same system specs). My next laptop will probably be some commodity Intel machine with Linux with lots of cores, RAM, etc.
(I will say this, though, the only laptop I've used with a decent trackpad and sleep/resume is a MacBook, so if those things are important to you, you might want to consider one.)
There are a lot of people who spend a great deal of time documenting OSS projects in an ad-hoc way. They have an itch, they play around with it for several days, and then write short how-tos on how to go about scratching that particular itch. The web is full of articles, blog posts, and forum posts of this nature. The problem isn't that there aren't enough people willing to document and make this stuff accessible, it's the massive amount of experimentation that has to go into figuring out how it can/should work in the first place. A small problem with limited scope and defined "success" is motivating and can be documented through trial and error. Looking at an application holistically, someone who isn't intimately familiar with the code does not know where to even start. It's not like there are specification documents to go by.
If the developer defines the requirements and functionality, then the developer must document the implementation at least in a rudimentary way to make it accessible to a technical writer. A technical writer (say, someone from O'Reilly for instance) can then make it accessible to a broader audience.
I think the best thing an OSS project can do is actively maintain and participate on a Wiki and a forum and from them derive a "How do I..." knowledge base where users can learn by example.
A large part of my contribution comes from forum participation. If I have an issue, and I don't have a support contract, then I take it to the forums and iron it out there. That way the next person that comes along with the same problem can learn from that, which helps the community. Bug reports help in the same way, provided you do enough investigation on your side and provide useful scenarios and test cases. And since you are troubleshooting with others from the community, the solution is not "owned" by any one person or company. If your company doesn't like that, then they can purchase support and work their problems out privately with the vendor.
I also code and submit patches at home for projects unrelated to my employer's business. In a way, my use of F/OSS software is of benefit to me. I'm sure my employer couldn't care less if I used Windows XP and Notepad to code, but thankfully I have the latitude to choose my own tools and I choose to use Linux with a F/OSS IDE because it makes me more efficient and preserves my sanity. So I try to give back on my own time in return for the benefits I receive personally.
You may be bound by NDAs and noncompetes that preclude the above, but even promoting F/OSS software is also a way of contributing. Driving demand for F/OSS solutions will increase adoption and support on the vendor side. Most businesses like to maintain support contracts, so your endorsements can help F/OSS expand in the enterprise and increase support revenue.
Glassfish may continue to live on as the reference implementation for the JEE stack. It's open source so it's not just going to vanish. It already uses TopLink (from Oracle) as its EJB3 persistence implementation and already plays very nicely with Eclipse and especially NetBeans (which they now also own). If Oracle was smart, they would build an enterprise Java development stack around NetBeans, Glassfish, ADF Faces, and Oracle DB and reap the benefits of a massive installed user base and the contributions of the open source community. It would be healthy competition for RedHat's JBoss+Eclipse+Postgres stack (although I'd take the latter any day.) Heck, all its missing really is an enterprise DB and a big name to provide support.
Not sure where WebLogic is with EJB3, ESB, and other emerging APIs and technologies, but Glassfish is already there, so you would think it would be a simply cost/benefit analysis in terms of the cost of bringing one up to the level of the other.
I doubt the JCP will be unaffected by this acquisition. I think Oracle had to make a lot of concessions and put a lot of money into refactoring their app server to support the final EJB3 spec. I'm sure they would have preferred to continue in whatever direction they were headed, both because it would have been cheaper and because it would have meant less of an impact their existing customers and applications.
Of course, now that they've put the effort into bringing their app server up to date and making it standards compliant, perhaps they will keep playing nice, but then again, it's their ball now, and they can take it home whenever they want to.
No, because currently the JCP moves along at an acceptable pace and tries hard to accommodate all the major players. IBM has a solid open source track record, so I don't doubt that they will work with the open source community to ensure compatibility. That being said, it's still two separate VMs, and when you factor in nonstandard VM options like heap size, garbage collection, permgen space, etc, you end up running into a lot of unintentional issues for apps developed against one running on the other.
Oracle has its own VM that they've tweaked so that you can run Java stored procedures, and I know from experience that it has compatibility issues. So even if IBM and OpenJDK play nicely, if one third (or more) of the Java base is running on even a slightly incompatible VM you will have issues with adoption.
And of course all of this is cynical speculation, but I wouldn't put it past Oracle to optimize their Java stuff to work best with their operating system and their database. My primary concern is not so much with binary compatibility as it is with the JCP and extensions to the language.
Re:Java 8 Preview
on
Oracle Buys Sun
·
· Score: 3, Interesting
I think completely open source. But that doesn't matter. Supposing it forks, you would in essence end up with 3 different dialects of Java: Oracle, IBM, and RedHat (FOSS) which may or may not remain binary compatible. Having Sun as an external arbiter with no direct commercial interest in the success of one stack versus another meant that you had agreement among the major interests and a unified direction. (Sun does/did have an app server stack, but it's always been more of a reference implementation.)
EJB3 was almost a carbon-copy of Hibernate, which was the most advanced, feature-rich ORM implementation at the time. Had Oracle been in charge at the time, the scales might have tipped in favor of TopLink, which would have left the OSS community playing catch up trying to implement a less elegant solution geared toward one vendor's RDBS. Granted, TopLink has essentially been open-sourced (it is the EJB3 implementation used by Glassfish), but that may not have happened either if Oracle had control over the reference app server stack.
Oracle doesn't have the commitment to open standards and open source that Sun does. I don't trust them to continue to open up new technologies and allow much community participation. I expect closed, buggy extensions that will ultimately be imitated by 1001 open source knock-offs , leading to fragmentation like we've never seen before.
The market is converging on stack-oriented development, so perhaps this is inevitable. It seems now that instead of simply knowing a language you have to know a particular IDE, DB, and app server as well. This is just another step in that direction.
And one last point. I've always found the Sun Java forums to be helpful, but I've never had much luck with Oracle forums. I think this is bad news for the community all around.
Re:Java 8 Preview
on
Oracle Buys Sun
·
· Score: 4, Interesting
Seriously, anyone who has taken a close look at what Oracle has done to Java with JDeveloper and Oracle AS knows that this will not be good for Java. Oracle is famous for not implementing standard API calls and instead providing proprietary methods and super classes to implement basic functionality (JDBC BLOBs, web services, etc.) Vendor lock-in is one thing, but their ideas and designs are just ugly and unwieldy.
They had started to play nice with EJB3 and TopLink, but now they have absolutely no reason to keep doing so. They now have much more weight in the JCP process (if the JCP even continues to exist) and they can now push out better ideas from competitors. I'm very apprehensive about the future of Java.
I realize this is a troll, but I'll bite. I've used.NET and I was very impressed with the visual design tools. The ASPX + codebehind setup was very intuitive and powerful, and the DnD/WYSIWYG interface worked well and as expected. I also loved C# (which is vastly superior to Java, IMO), especially the notion of properties, events, and expression execution (the former of which are an improvement over the pre-existing JavaBeans spec, but the latter of which was innovative and copied poorly in Java Unified EL). I changed jobs shortly after.NET 1.1 came out, so I don't know what subsequent versions had, but my biggest gripes with.NET were:
* Poor support for web service interoperability (especially document/literal)
* No ORM solution (NHibernate is out now, but I don't know if.NET proper has an answer for this)
* Reliance on all sorts of other MS products (IIS, SQL server, DTS, etc.)
* Unintuitive/cubersome deployment methodology (xcopy, DLL versioning, difficult to share libraries across sites, etc.)
* This may be my own ignorance, but I also don't remember any kind of object pooling, RIM-IIOP analogue, or distributed computing frameworks (ala remote EJBs and clustering).
* The IIS-based security system also left a lot to be desired.
And the number one gripe, of course, is the archaic NPE deep in some MS code that I can't look at because it's not open source, and waiting 3-4 days to get an answer or resolution from MS tech support. I much prefer to have source code and message boards.
The shops that ran Tomcat generally used JDBC or some in-house persistence framework. In either case, there were no facilities for dirty checking or optimistic locking. So if user A loads an object into his session and modifies it; and user B loads the same object into her session and modifies it, then the changes made by whomever commits first are lost. Small shops generally ignore this problem as it happens infrequently enough to warrant serious development time, and usually end users catch the error and fix it manually. Or they will identify a particularly problematic section of the app and implement some kind of DB-driven pessimistic locking (select for update) which has its own issues (stuck locks, timeouts, etc.).
JPA fixes this by making dirty-checks and optimistic locking functionality part of the spec. You can still choose not to use it, but it's trivial to use and generally most tools will enable it by default.
The other issue is transaction scope. I've seen implementations where every user gets a connection and holds it for his entire session. Then I've seen implementations where a new connection is retrieved from a connection pool before every JDBC statement. When you use container-managed persistence, then you are forced to deal with things like transaction scope, and the defaults are usually sane enough to use without additional configuration.
Finally, if you are executing multiple transactions against multiple resources, then JTA+XA saves you the hassle of implementing your own home-grown 2-stage commit process to handle distributed transactions. A lot of places will just let the commits happen when they happen and then manually fix problems that pop up do to partially completed transactions, but that's sloppy and easily avoided with JEE.
I've never seen a serious commercial product shipped independently of an application server. I've always though of an application server to be more of a framework. As a product author/vendor, I make choices about which application server best serves my product. Then I build around it. When I sell to end users, I either enter into some kind of licensing/support agreement with the application vendor, or I require the end user to purchase licenses and support themselves.
The spec exists so that developers can write components that can interoperate with various application servers. For example, I can write a data connector for a proprietary mainframe application and use JCA to include that in distributed XA transactions. That resource adaptor can be used on virtually any application server, but the way it is deployed and configured will differ.
Or I might write a custom JAAS LoginModule to support some kind of proprietary authentication requirements. Again, the ability to integrate into an application server is undisputed, thanks to the specifications, but the mechanisms by which it is integrated is left to the implementor.
To support full drop-in interoperability, Sun would have to control all of the details of deployment and configuration, and then there goes competition. Everyone would just grab a copy Glassfish and rebrand it, because how then would you innovate?
As a developer I'm glad that there are differences in application servers. The various classloading strategies, configuration interfaces, and monitoring tools ensure healthy competition and innovation. And if something becomes ubiquitous (or problematic enough) then it gets added to the standard. (EJB3 is largely Spring+Hibernate, standardized, and Seam+Facelets is on its way to standardization through JSF 2 and WebObjects.)
One last thing. I've worked in shops that ran Tomcat plus a few third party libs (i.e. Struts and Hibernate), and I've worked in shops that have used the full JEE stack. In the former, we had infrequent but persistent problems with transactions (lost updates, etc.) that required implementing a lot of proprietary transaction management code. Using the full JEE stack you have things like JTA and JPA that manage locking for you, which can be painful if you don't know what you're doing. But it forces you to deal with real issues rather than pretending they don't exist, which is a must for true enterprise applications. The same can be said for resource pooling, security, managing web sessions and conversations, and many other issues.
Seconded. I have a G3 iBook, a G5 iMac, and an Intel MacBook that suspend and restore within 1 second (with automatic hibernate if the battery dies while suspended). Not only that, but wireless generally connects and is ready to use within 1-3 seconds as well. I think I may have remotted each maybe 5 times throughout their respective lifespans, and then only for updates. I've used Dell, HP, Gateway, and Sony machines with mixes of Linux and Windows and none of them can do this reliably. It's really disappointing that other vendors can't seem to get this right.
They use a modified version of Eclipse called Ellipse designed for "learning". Given that Eclipse is more of an IDE SDK than an IDE itself, this makes sense. (Having developed Eclipse plugins in the past I find the documentation to be plentiful and robust, and the learning curve relatively mild. And since everything in Eclipse is a plugin, there are lots of examples to go by.) I doubt this is a deliberate dig at Sun as a competitor.
Oh the acronyms get uglier than that:) RedHat Enterprise Development Studio (or somesuch) is LJPJE, an integrated blend of Linux, JBoss, Postgres, and Java (JSF + Seam + JBPM), and Eclipse. Great stack but not quite as catchy as LAMP. (Even for "Apache" it would be Tomcat, which still lends to an over-abundance of consonants: LTPJE).
Open source Java is fantastic news for anyone competing with Oracle (which has their own bastardization of Java) and Microsoft/.NET.
The plurality voting system sucks. If we could rank each candidate or "grade" each candidate on a scale and then add up the total scores, then we would have viable third parties. Rather than fret about the party or candidate you oppose winning, you can vote your conscience knowing that your second- and third-place picks will count. Think of how many more people would have voted for Paul in the Republican primary had they not been overly concerned about preventing Huckabee from beating Romney...
Well it does serve to educate the student about copyright infringement and prove that the student is fully aware of the consequences of his actions in the event that he does engage in piracy. This could be a move to limit the school's liability in civil suits involving its students. If the school forces the students to continually and consistently demonstrate that they understand what copyright infringement is, then the school wouldn't necessarily be liable for the actions of its students. (I hesitate to say "can't" be held liable, although this does seem like more than a good faith effort on their part.)
Not that I advocate this. It just looks like the university caving to corporate interests to protect themselves.
Legally, would it matter whether the water mark expired or whether it did not specifically exempt fair use? A watermark is just an identifier. So if you were found with a watermarked file, and 1) the date it was found was after the copyright expiration date; or 2) the situation was deemed "fair use" by judge/jury/arbitrator, then wouldn't you be off the hook? Of course, there's always the cynical so-what-it-will-be-abused-anyway perspective, but all things being fair and equal, watermarks wouldn't necessarily be bad things unless abused.
You're missing the point. The administration wants Libby to be exonerated through the appeals process, so a pardon won't happen until the appeals process has run its course. A successful appeal looks much better than a pardon, as you can see from the majority of the posts here. However, in the meantime, the administration doesn't want him to serve a sentence for a conviction that may be overturned. It's only 30 months, and the appeals process definitely will not have concluded in that time. If the conviction is eventually overturned he will have already served time for a crime the appeals court says he did not commit. If the judge had allowed Libby to delay serving his sentence until after the appeal, then Bush would not have felt the need to commute Libby's sentence. (Evidently they are pretty confident that an appeal will be successful.)
So it's not cowardice (since when has Bush given any consideration to his opposition?), it's just political strategy.
Admins need to be able to see the message and move the messages around, but they don't need to see the content. If you're transmitting sensitive data, encrypt it!
Slashdot Mirror
I'm a Mac user by choice that is currently typing this from Kubuntu 9.04 with Windows 7 executing in a background. I was impressed with Windows 7 until I tried to play with networking. It wasn't working initially (turned out to be a conflict with VirtualBox's default 10.0.2.x NAT range) but I needed to see how Windows was configured (DHCP, etc.). So I go to Control Panel -> Network and Internet -> View Network Status and Tasks -> Change Adapter Settings. Then I left-click on "Local Connection". Now, to get the IP and gateway info, I have to click "View Status of Connection". To view whether it's set up for DHCP check the DNS servers, I have to click "Change Settings of this Connection", right-click on "Internet Protocol Version 4 (TCP/IP)" and click "Properties".
By contrast, here is what I do in Mac OS X: Click the apple menu, click System Preferences. Click Network. There I'm presented with a list of all of my adapters. I click on one and I see everything I need.
In Kubuntu, I click the knetwork-manager applet and click "Manage Connections"
Now, which of the three is the most usable? Keep in mind that as I was troubleshooting the networking issue I had with my VM, I constantly had to repeat those steps. What is that, like 10 clicks? Look at the menu names ... "View Network Status and Tasks", "Change Adapter Settings" ... is this supposed to be intuitive?
Another thing, is that Windows pops up every time I jump on a network and asks me if it's a home network, work network, or public network, and initializes stuff for me (including home groups, which I don't want). Now, this is fantastic for end users, and a great feature. But as a power user running 3 levels of NAT at home (local net, work VPN, and NAT VMs) it is infuriating to have the details hidden from me and not know how to get to them.
I would definitely recommend Windows 7 to anyone who wants to plug something in and have it just work and be done with it (which is I think the point). But the configuration is hellish, so if you like to tinker, think twice.
I'm not well-versed in the distinction between netbooks and nettops, but I have seen $200-$300 devices that are bigger than phones and smaller than laptops with and without built in 3G support.
Laptops are approaching the price of desktops, and portability, and that's with comparable system specs. A lot of employees do little more than tinker with documents, spreadsheets, and E-mail/calendaring software, so there is no great need for horsepower. So in the office I see three types of devices: workstations (for developers/IT professionals); dumb terminals connected to Citrix servers (or similar) for stationary employees or non-personal stations (think call center); and laptops for managers, salespeople, and employees prone to frequent travel or telework.
It's called communism, and it fails because humans are lazy and greedy. If you can somehow persuade humans to stop being lazy and greedy, then pretty much any system of government will suffice.
It will happen. 5 years ago, a family had one central PC that they all had to fight over. Now they all sit in the living room with their individual smartphones and nettops that they can afford because they are subsidized by the carrier. They are cheaper and everyone has their own device with 24/7 access, and every two years you can upgrade cheap. More than enough for text, IM, E-mail, web browsing, money management, research and report writing, casual gaming, and FaceSpace. Heck they're selling a wireless printer/scanner/fax/copier at Target for less than $100. They have NAS devices wireless interfaces, and you can plug a USB 2.0 backup drive into those. Who needs a server or central PC?
You'll still have gamers and hobbyists purchasing PCs, but they will be more likely to be specialized high-end machines.
On the server side you'll still have big metal, whether it's some distributed cluster or mainframes of a sort. Got to serve the cloud.
For anyone just entering a CS/CIS program in college, smart money's on embedded programming or multi-core parallel programming on distributed/clustered systems. (And it won't be too long before someone starts sticking multiple cores in cell phones, if they haven't already...)
I have an Apple MacBook I use as my primary PC at home, and I love Apple products, but Java on Mac OS X sucks. It's slow, always at least 2 years behind the current version, and the Cocoa LAF make Swing apps look terrible.
I bought my MacBook with the intention of making it my primary development laptop (for school--we used a lot of esoteric languages and libraries), but I found Java next to unusable. I use Ubuntu at work on a Dell D830 and the the difference is night and day in terms of performance (same system specs). My next laptop will probably be some commodity Intel machine with Linux with lots of cores, RAM, etc.
(I will say this, though, the only laptop I've used with a decent trackpad and sleep/resume is a MacBook, so if those things are important to you, you might want to consider one.)
There are a lot of people who spend a great deal of time documenting OSS projects in an ad-hoc way. They have an itch, they play around with it for several days, and then write short how-tos on how to go about scratching that particular itch. The web is full of articles, blog posts, and forum posts of this nature. The problem isn't that there aren't enough people willing to document and make this stuff accessible, it's the massive amount of experimentation that has to go into figuring out how it can/should work in the first place. A small problem with limited scope and defined "success" is motivating and can be documented through trial and error. Looking at an application holistically, someone who isn't intimately familiar with the code does not know where to even start. It's not like there are specification documents to go by.
If the developer defines the requirements and functionality, then the developer must document the implementation at least in a rudimentary way to make it accessible to a technical writer. A technical writer (say, someone from O'Reilly for instance) can then make it accessible to a broader audience.
I think the best thing an OSS project can do is actively maintain and participate on a Wiki and a forum and from them derive a "How do I..." knowledge base where users can learn by example.
A large part of my contribution comes from forum participation. If I have an issue, and I don't have a support contract, then I take it to the forums and iron it out there. That way the next person that comes along with the same problem can learn from that, which helps the community. Bug reports help in the same way, provided you do enough investigation on your side and provide useful scenarios and test cases. And since you are troubleshooting with others from the community, the solution is not "owned" by any one person or company. If your company doesn't like that, then they can purchase support and work their problems out privately with the vendor.
I also code and submit patches at home for projects unrelated to my employer's business. In a way, my use of F/OSS software is of benefit to me. I'm sure my employer couldn't care less if I used Windows XP and Notepad to code, but thankfully I have the latitude to choose my own tools and I choose to use Linux with a F/OSS IDE because it makes me more efficient and preserves my sanity. So I try to give back on my own time in return for the benefits I receive personally.
You may be bound by NDAs and noncompetes that preclude the above, but even promoting F/OSS software is also a way of contributing. Driving demand for F/OSS solutions will increase adoption and support on the vendor side. Most businesses like to maintain support contracts, so your endorsements can help F/OSS expand in the enterprise and increase support revenue.
Glassfish may continue to live on as the reference implementation for the JEE stack. It's open source so it's not just going to vanish. It already uses TopLink (from Oracle) as its EJB3 persistence implementation and already plays very nicely with Eclipse and especially NetBeans (which they now also own). If Oracle was smart, they would build an enterprise Java development stack around NetBeans, Glassfish, ADF Faces, and Oracle DB and reap the benefits of a massive installed user base and the contributions of the open source community. It would be healthy competition for RedHat's JBoss+Eclipse+Postgres stack (although I'd take the latter any day.) Heck, all its missing really is an enterprise DB and a big name to provide support.
Not sure where WebLogic is with EJB3, ESB, and other emerging APIs and technologies, but Glassfish is already there, so you would think it would be a simply cost/benefit analysis in terms of the cost of bringing one up to the level of the other.
I doubt the JCP will be unaffected by this acquisition. I think Oracle had to make a lot of concessions and put a lot of money into refactoring their app server to support the final EJB3 spec. I'm sure they would have preferred to continue in whatever direction they were headed, both because it would have been cheaper and because it would have meant less of an impact their existing customers and applications.
Of course, now that they've put the effort into bringing their app server up to date and making it standards compliant, perhaps they will keep playing nice, but then again, it's their ball now, and they can take it home whenever they want to.
No, because currently the JCP moves along at an acceptable pace and tries hard to accommodate all the major players. IBM has a solid open source track record, so I don't doubt that they will work with the open source community to ensure compatibility. That being said, it's still two separate VMs, and when you factor in nonstandard VM options like heap size, garbage collection, permgen space, etc, you end up running into a lot of unintentional issues for apps developed against one running on the other.
Oracle has its own VM that they've tweaked so that you can run Java stored procedures, and I know from experience that it has compatibility issues. So even if IBM and OpenJDK play nicely, if one third (or more) of the Java base is running on even a slightly incompatible VM you will have issues with adoption.
And of course all of this is cynical speculation, but I wouldn't put it past Oracle to optimize their Java stuff to work best with their operating system and their database. My primary concern is not so much with binary compatibility as it is with the JCP and extensions to the language.
I think completely open source. But that doesn't matter. Supposing it forks, you would in essence end up with 3 different dialects of Java: Oracle, IBM, and RedHat (FOSS) which may or may not remain binary compatible. Having Sun as an external arbiter with no direct commercial interest in the success of one stack versus another meant that you had agreement among the major interests and a unified direction. (Sun does/did have an app server stack, but it's always been more of a reference implementation.)
EJB3 was almost a carbon-copy of Hibernate, which was the most advanced, feature-rich ORM implementation at the time. Had Oracle been in charge at the time, the scales might have tipped in favor of TopLink, which would have left the OSS community playing catch up trying to implement a less elegant solution geared toward one vendor's RDBS. Granted, TopLink has essentially been open-sourced (it is the EJB3 implementation used by Glassfish), but that may not have happened either if Oracle had control over the reference app server stack.
Oracle doesn't have the commitment to open standards and open source that Sun does. I don't trust them to continue to open up new technologies and allow much community participation. I expect closed, buggy extensions that will ultimately be imitated by 1001 open source knock-offs , leading to fragmentation like we've never seen before.
The market is converging on stack-oriented development, so perhaps this is inevitable. It seems now that instead of simply knowing a language you have to know a particular IDE, DB, and app server as well. This is just another step in that direction.
And one last point. I've always found the Sun Java forums to be helpful, but I've never had much luck with Oracle forums. I think this is bad news for the community all around.
Seriously, anyone who has taken a close look at what Oracle has done to Java with JDeveloper and Oracle AS knows that this will not be good for Java. Oracle is famous for not implementing standard API calls and instead providing proprietary methods and super classes to implement basic functionality (JDBC BLOBs, web services, etc.) Vendor lock-in is one thing, but their ideas and designs are just ugly and unwieldy.
They had started to play nice with EJB3 and TopLink, but now they have absolutely no reason to keep doing so. They now have much more weight in the JCP process (if the JCP even continues to exist) and they can now push out better ideas from competitors. I'm very apprehensive about the future of Java.
I realize this is a troll, but I'll bite. I've used .NET and I was very impressed with the visual design tools. The ASPX + codebehind setup was very intuitive and powerful, and the DnD/WYSIWYG interface worked well and as expected. I also loved C# (which is vastly superior to Java, IMO), especially the notion of properties, events, and expression execution (the former of which are an improvement over the pre-existing JavaBeans spec, but the latter of which was innovative and copied poorly in Java Unified EL). I changed jobs shortly after .NET 1.1 came out, so I don't know what subsequent versions had, but my biggest gripes with .NET were:
* Poor support for web service interoperability (especially document/literal) .NET proper has an answer for this)
* No ORM solution (NHibernate is out now, but I don't know if
* Reliance on all sorts of other MS products (IIS, SQL server, DTS, etc.)
* Unintuitive/cubersome deployment methodology (xcopy, DLL versioning, difficult to share libraries across sites, etc.)
* This may be my own ignorance, but I also don't remember any kind of object pooling, RIM-IIOP analogue, or distributed computing frameworks (ala remote EJBs and clustering).
* The IIS-based security system also left a lot to be desired.
And the number one gripe, of course, is the archaic NPE deep in some MS code that I can't look at because it's not open source, and waiting 3-4 days to get an answer or resolution from MS tech support. I much prefer to have source code and message boards.
Yes, thank you for the correction.
The shops that ran Tomcat generally used JDBC or some in-house persistence framework. In either case, there were no facilities for dirty checking or optimistic locking. So if user A loads an object into his session and modifies it; and user B loads the same object into her session and modifies it, then the changes made by whomever commits first are lost. Small shops generally ignore this problem as it happens infrequently enough to warrant serious development time, and usually end users catch the error and fix it manually. Or they will identify a particularly problematic section of the app and implement some kind of DB-driven pessimistic locking (select for update) which has its own issues (stuck locks, timeouts, etc.).
JPA fixes this by making dirty-checks and optimistic locking functionality part of the spec. You can still choose not to use it, but it's trivial to use and generally most tools will enable it by default.
The other issue is transaction scope. I've seen implementations where every user gets a connection and holds it for his entire session. Then I've seen implementations where a new connection is retrieved from a connection pool before every JDBC statement. When you use container-managed persistence, then you are forced to deal with things like transaction scope, and the defaults are usually sane enough to use without additional configuration.
Finally, if you are executing multiple transactions against multiple resources, then JTA+XA saves you the hassle of implementing your own home-grown 2-stage commit process to handle distributed transactions. A lot of places will just let the commits happen when they happen and then manually fix problems that pop up do to partially completed transactions, but that's sloppy and easily avoided with JEE.
I've never seen a serious commercial product shipped independently of an application server. I've always though of an application server to be more of a framework. As a product author/vendor, I make choices about which application server best serves my product. Then I build around it. When I sell to end users, I either enter into some kind of licensing/support agreement with the application vendor, or I require the end user to purchase licenses and support themselves.
The spec exists so that developers can write components that can interoperate with various application servers. For example, I can write a data connector for a proprietary mainframe application and use JCA to include that in distributed XA transactions. That resource adaptor can be used on virtually any application server, but the way it is deployed and configured will differ.
Or I might write a custom JAAS LoginModule to support some kind of proprietary authentication requirements. Again, the ability to integrate into an application server is undisputed, thanks to the specifications, but the mechanisms by which it is integrated is left to the implementor.
To support full drop-in interoperability, Sun would have to control all of the details of deployment and configuration, and then there goes competition. Everyone would just grab a copy Glassfish and rebrand it, because how then would you innovate?
As a developer I'm glad that there are differences in application servers. The various classloading strategies, configuration interfaces, and monitoring tools ensure healthy competition and innovation. And if something becomes ubiquitous (or problematic enough) then it gets added to the standard. (EJB3 is largely Spring+Hibernate, standardized, and Seam+Facelets is on its way to standardization through JSF 2 and WebObjects.)
One last thing. I've worked in shops that ran Tomcat plus a few third party libs (i.e. Struts and Hibernate), and I've worked in shops that have used the full JEE stack. In the former, we had infrequent but persistent problems with transactions (lost updates, etc.) that required implementing a lot of proprietary transaction management code. Using the full JEE stack you have things like JTA and JPA that manage locking for you, which can be painful if you don't know what you're doing. But it forces you to deal with real issues rather than pretending they don't exist, which is a must for true enterprise applications. The same can be said for resource pooling, security, managing web sessions and conversations, and many other issues.
Seconded. I have a G3 iBook, a G5 iMac, and an Intel MacBook that suspend and restore within 1 second (with automatic hibernate if the battery dies while suspended). Not only that, but wireless generally connects and is ready to use within 1-3 seconds as well. I think I may have remotted each maybe 5 times throughout their respective lifespans, and then only for updates. I've used Dell, HP, Gateway, and Sony machines with mixes of Linux and Windows and none of them can do this reliably. It's really disappointing that other vendors can't seem to get this right.
They use a modified version of Eclipse called Ellipse designed for "learning". Given that Eclipse is more of an IDE SDK than an IDE itself, this makes sense. (Having developed Eclipse plugins in the past I find the documentation to be plentiful and robust, and the learning curve relatively mild. And since everything in Eclipse is a plugin, there are lots of examples to go by.) I doubt this is a deliberate dig at Sun as a competitor.
Oh the acronyms get uglier than that :) RedHat Enterprise Development Studio (or somesuch) is LJPJE, an integrated blend of Linux, JBoss, Postgres, and Java (JSF + Seam + JBPM), and Eclipse. Great stack but not quite as catchy as LAMP. (Even for "Apache" it would be Tomcat, which still lends to an over-abundance of consonants: LTPJE).
Open source Java is fantastic news for anyone competing with Oracle (which has their own bastardization of Java) and Microsoft/.NET.
The plurality voting system sucks. If we could rank each candidate or "grade" each candidate on a scale and then add up the total scores, then we would have viable third parties. Rather than fret about the party or candidate you oppose winning, you can vote your conscience knowing that your second- and third-place picks will count. Think of how many more people would have voted for Paul in the Republican primary had they not been overly concerned about preventing Huckabee from beating Romney...
Well it does serve to educate the student about copyright infringement and prove that the student is fully aware of the consequences of his actions in the event that he does engage in piracy. This could be a move to limit the school's liability in civil suits involving its students. If the school forces the students to continually and consistently demonstrate that they understand what copyright infringement is, then the school wouldn't necessarily be liable for the actions of its students. (I hesitate to say "can't" be held liable, although this does seem like more than a good faith effort on their part.)
Not that I advocate this. It just looks like the university caving to corporate interests to protect themselves.
Legally, would it matter whether the water mark expired or whether it did not specifically exempt fair use? A watermark is just an identifier. So if you were found with a watermarked file, and 1) the date it was found was after the copyright expiration date; or 2) the situation was deemed "fair use" by judge/jury/arbitrator, then wouldn't you be off the hook? Of course, there's always the cynical so-what-it-will-be-abused-anyway perspective, but all things being fair and equal, watermarks wouldn't necessarily be bad things unless abused.
Tim Burton? Who would Johnny Depp play, Bilbo? I can't wait to see the flashbacks of Bilbo's twisted upbringing and the genesis of his daddy issues.
Seriously, did you see what that guy did to Charlie and the Chocolate Factory?
You're missing the point. The administration wants Libby to be exonerated through the appeals process, so a pardon won't happen until the appeals process has run its course. A successful appeal looks much better than a pardon, as you can see from the majority of the posts here. However, in the meantime, the administration doesn't want him to serve a sentence for a conviction that may be overturned. It's only 30 months, and the appeals process definitely will not have concluded in that time. If the conviction is eventually overturned he will have already served time for a crime the appeals court says he did not commit. If the judge had allowed Libby to delay serving his sentence until after the appeal, then Bush would not have felt the need to commute Libby's sentence. (Evidently they are pretty confident that an appeal will be successful.)
So it's not cowardice (since when has Bush given any consideration to his opposition?), it's just political strategy.
Admins need to be able to see the message and move the messages around, but they don't need to see the content. If you're transmitting sensitive data, encrypt it!