The article doesn't mention one of the most interesting aspects: Hadoop is Java.
It seems like Microsoft has had a policy of not touching Java with a ten foot pole for some time now. At least, since .NET came out, they've been pretty proactively anti-Java.
"Sun Microsystems, the creator of Java, sued Microsoft in October 1997 for incompletely implementing the Java 1.1 standard.[2] It was also named in the United States v. Microsoft antitrust civil actions, as an implementation of Microsoft's Embrace, extend and extinguish strategy. In 2001, Microsoft settled the lawsuit with Sun and discontinued its Java implementation"
Glad to see they're acting more mature these days.
And I'm sure they will. . . but lately Microsoft seems to have grown up and learned to not create incompatible extensions. Perhaps they're finally starting to realize that nobody really loves them, so if they refuse to be nice to other children then they'll just get kicked out of the schoolyard. (Ballmer and upper management may not have, but they're facing the real possibility of a mutiny right now.) So it'll hopefully just end up being the kind of situation where if you switch platforms you'll lose some of the bells and whistles that were provided in a library on the old platform. Just like what would happen if you switch from one open-source platform to another.
What concerns me more is:
> the company is providing “connectors” for moving data between the two.
Microsoft loves to sink a lot of time into gee-whiz crap that lets their MVCs throw together gee-whiz crap using GUI tools without ever writing any code. They love it so much that they often waste all their effort on that stuff without bothering to get the fundamentals in order, and the end result is unusable garbage like much of ADO.NET and early versions of Entity Framework. That one sentence makes me fear more of the same old garbage: We'll get a port of Hadoop that comes with a plethora of toys in the System.DraggyDroppy namespace for impressing the suits, but is crippled by a broken set of bridge libraries for the programmers.
Which is fine. A Windows port of Hadoop isn't really necessary. We can just keep using REST or whatever to keep the Windows servers talking to the servers that do actual work, and Microsoft can keep chasing its own tail while it slides into irrelevance, and everyone can be happy.
> We'll get a port of Hadoop that comes with a plethora of toys in the System.DraggyDroppy namespace for impressing the suits
Sadly, for programmers who work where the impressionable suits reign, it's usually the suits who pick the tools their underlings will have to use to assemble whatever was conceived in the demos. They'll end up having to deploy Hadoop on Windows 8 Server and talk to it using the vile offspring of Clippy and Hadoop that lives under System.DraggyDroppy.
> Perhaps they're finally starting to realize that nobody really loves them, so if they refuse to be nice to other children then they'll just get kicked out of the schoolyard.
Have you ever seen an application in Java that actually works in more than one system without any env-specific code at all and does not behave in a slightly incompatible way? I haven't. (apart from toy apps which exist only to prove the opposite)
Fair enough. I did just read this - "Various parts of Hadoop are executed using shell scripts that will only work on a Linux shell.". Sounds like you need Cygwin (currently). Guess those are the bits that will be ported.
To me, this sounds like a pragmatic continuation of their 'big data' plans on Azure -- gotta go after what the current user base is used to using -- and probably compliments their own Dryad project (http://research.microsoft.com/en-us/projects/dryad/)
The company is now working to port the Hadoop platform to Windows — it was built for use atop Linux — and Doug Leland, general manager of product management for SQL Server, told Wired that the company plans to eventually release its work back to the open source community.
Haha, some things never change. A code dump of major changes developed in isolation. I'm sure the Hadoop commuity will love that.
It's really great to see Microsoft moving on to embracing change, first WP7 and now this. A tendency and a positive attitude towards evolving the company is noticable.
24 comments
[ 15.1 ms ] story [ 1058 ms ] threadIt seems like Microsoft has had a policy of not touching Java with a ten foot pole for some time now. At least, since .NET came out, they've been pretty proactively anti-Java.
"Sun Microsystems, the creator of Java, sued Microsoft in October 1997 for incompletely implementing the Java 1.1 standard.[2] It was also named in the United States v. Microsoft antitrust civil actions, as an implementation of Microsoft's Embrace, extend and extinguish strategy. In 2001, Microsoft settled the lawsuit with Sun and discontinued its Java implementation"
Glad to see they're acting more mature these days.
They still intend to port Hadoop to Windows. There is little to prevent extending it.
What concerns me more is:
> the company is providing “connectors” for moving data between the two.
Microsoft loves to sink a lot of time into gee-whiz crap that lets their MVCs throw together gee-whiz crap using GUI tools without ever writing any code. They love it so much that they often waste all their effort on that stuff without bothering to get the fundamentals in order, and the end result is unusable garbage like much of ADO.NET and early versions of Entity Framework. That one sentence makes me fear more of the same old garbage: We'll get a port of Hadoop that comes with a plethora of toys in the System.DraggyDroppy namespace for impressing the suits, but is crippled by a broken set of bridge libraries for the programmers.
Which is fine. A Windows port of Hadoop isn't really necessary. We can just keep using REST or whatever to keep the Windows servers talking to the servers that do actual work, and Microsoft can keep chasing its own tail while it slides into irrelevance, and everyone can be happy.
Sadly, for programmers who work where the impressionable suits reign, it's usually the suits who pick the tools their underlings will have to use to assemble whatever was conceived in the demos. They'll end up having to deploy Hadoop on Windows 8 Server and talk to it using the vile offspring of Clippy and Hadoop that lives under System.DraggyDroppy.
I for one love Microsoft. There, I said it.
(Disclaimer: I work for Microsoft.)
It's OK. It's not your fault.
They also prove it's possible to write portable code in Java. ;-)
* Open source SDK for Java (and Eclipse): http://www.windowsazure4j.org/ * Gigaspace running their products on Azure: http://www.gigaspaces.com * More info (admittedly from a UK-based Azure consultant): http://www.theregister.co.uk/2011/09/09/java_on_windows_azur...
To me, this sounds like a pragmatic continuation of their 'big data' plans on Azure -- gotta go after what the current user base is used to using -- and probably compliments their own Dryad project (http://research.microsoft.com/en-us/projects/dryad/)
Haha, some things never change. A code dump of major changes developed in isolation. I'm sure the Hadoop commuity will love that.