I installed the Cuda Toolkit, installed the 32bit and 64bit versions of JCuda’s DLLs/JARs and downloaded the JCublas sample application.

While the compile finished fine, the app immediately terminated – once it tried to load JCuda’s interface DLLs (called using JNI). The VM printed the more or less useless message: ‘JCudaRuntime-windows-x86.dll isn’t a win32-application’.

After various efforts to make the app load the DLLs, I posted a question on the JCuda forum. Short after, Marco13 – I suppose, he’s the project owner of JCuda – pointed me to the right direction:

I installed the 64 bit JDK version for Windows, passed -D=64 to the VM – and the JCublas sample application finished orderly. Thx, Marco!

Although a not that fast Nvidia GeForce 210 runs in my PC, the speedup was enormous: JCuda based matrix multiplication of size 2500 x 2500 finished 40-times faster than equivalent pure JAVA code [which utilized the 4 cores too].  Not bad

An important aspect to note: JCuda comes with JCublas [linear algebra], JCufft [fast fourier transforms) and JCudpp [data parallel primitives]. These libs provide access to ‘pre-configured’ high-level CUDA routine. E.g., a matrix multiplication is  basically a call to a JCublas provided routine.

This mode of operation is much simpler, than to write  a so called CUDA Kernel, a piece of syntactically enriched C code. On the other side, CUDA Kernels are more versatile.

exoPC :: 599.- USD, 8.9” multitouch, Windows 7 oder Linux

A promissing device comes from exoPC. Here’s a video from engadget.com. The device comes with a single-core Atom N270 and plays video on its 1024×600 touchscreen quite good:

Have a look inside the exoPC: Image galery. The device is planned to be shipped March 2010. All in all, it’s quite good equipped: 2 GBByte RAM, 32 GByte SSD [4.5 GByte reserved for OS], full-size SD Card slot, Atom 270 @ 1,6 GHz, WiFi 802.11 b/g, Bluetooth, Lan 10/100, 3 x USB, Audio in/out, VGA – and an 1.3 Megapixels webcam. Compare the exoPC with other here device on this page. Product photos.

joojoo :: 499.- USD. 12.1-inch tablet-appliance

joojoo is a nice looking device of enourmous 12.1 inc size. While other system come with Microsoft’s Windows oder Apple’s MacOS, joojoo comes with its own pure browser-based environment. As far as the web-site explains, each and everything runs in a browser. Thus, joojoo doesn’t run native applications – but Flash and Java. joojoo originated from the never released crunchPad.

NotionInk’s ADAM :: [ARM Cortex-A9 MPCore]

NotionInk’s ADAM features an Cortex-A9 MPCore processor running Android, a 10.1” screen with 1024×600 pixels and a 180 degree Swivel Camera.

Unlike many other tablets, Adam features dual display, which runs either in transmissive ePaper mode or in transflective LCD mode. It’s Android distribution will be optimized for the large 10” display size. Not to mention 3G, WiFi and multi-touch.

The Cortex-Ap MPCORE and NVIDIA Tegra T20 accelerate Flash 10.1 and HD video – at 140 hrs audio or 16 hrs video playback.

NotionInk in India-based.

HP’s Slate-PC :: Details unknown

HP’s Slate-PC will be much smaller than the planned device of Apple. Gizmodo.de did write a review. A HP sponsored promotional interview:

Microsoft Courrier :: UX study

This review of Microsoft’s Courier looks nice:

Please note, that Microsoft’s Courier is currently not more than an UX study, not a real product. Anyway, it looks very promising.

Yinlips :: details unknown

Another iPad clone, likely from china, is Yinlips. Found by pmtoday.com.

t.b.c.

Actually, I returned it today, since Amazon’s PDF conversion service is by no means able to convert more than trivial text.

Here is part of my input PDF, which I created using LaTeX:

And this is the formula part, that Amazon converted for Kindle as seen on the device:

Beside the ill-converted diacritical letters  - ä converts to “a – the formula is now a useless string the characters.

In case you’d like to display mathematics/scientific text, don’t buy a Kindle.

In 60 days, I’ll give iPad a try. Since even iPhone – as MacOS X – displays good, but small, iPad will do it brilliant.

In case you try to auto-upgrade too, here is the way to manage the task anyway:

• - First, download the WP 2.9 upgrade archive file [download link]
• - Unpack the archive file to a local volume [your Mac OS X system unpacks automatically]
• - Start you FTP-client and connect to your WP virtual host [for Mac OS X, you might wish to download the FTP client Cyberduck]
• - Finally select all files within the archive and drag them to you FTP-client’s window

Now, it’s time to run the real upgrade process. Enter this URL

http://www.yourSiteName.com/wp-admin/upgrade.php

or, if WP is installed in a folder named e.g. myBlog, enter this URL

http://www.yourSiteName.com/myBlog/wp-admin/upgrade.php

At this point, you might end up with a blank page or one, which indicates a HTTP error 500. In this case, have a look at my previous post, which provides a solution regarding PHP setting of certain providers (namely 1&1).

After that, enter one of the above URLs again. WordPress should now present a small upgrade wizard, which explains the next steps to do. A typical upgrade process should finish within a few minutes (at least mine finished soon).

For other types of application, this approach tends to make development complicated.

Suppose, you need to create a note taking application. The app should be able to take notes. A note should have a title, a text body and an arbitrary number of optional fields, a use may use to add small snippets of text. Hm, an arbitrary number of fields, which you don’t know up front? How would you implement them in your favorite DB system?

Create a new table, which keeps each single optional field? Probably too inefficient.
Serialize optional fields and drop them into a single attribute of a table? Hm, complicated. How would you manage search operations?

Without pointing towards a relational solution, I’d propose to use a non-relational DBMS system. Especially a schema-free DBMS.

Schema-free solutions

A schema-free DBMS is one, which allows to store any kind of data of any structure. You don’t need to define your DBs tables and attributes up-front. Simply create new attributes of as soon as you need a new attribute.

Think of a schema-free DBMS as one, which manages key-value coded tuples [XCode users: Key-Value coding comes to mind].

In case you’d like to have a look, your first Google hit might point you to CouchDB, which meanwhile is a project of the Apache foundation. CouchDB is an Erlang-based solution. Thus you not only need to download and install CouchDB, but an Erlang runtime too.

Recently I stumbled upon MongoDB, which comes as a no-dependencies solution. Download it and start mongod, the MongoDB server. That’s it. Finally, you might wish to download and install a MongoDB driver for your preferred language.

Yet, MongoDB doesn’t come with an AS3 binding. To get access to MongoDB from within your Flex/Air app, various ways exist. One might use MongoDB’s PHP-driver, add some AMFPHP magic and finally expose your MongoDB’s API to flex.

Resources

A recorded webinar by MongoDB’s parents
A MongoDB presentation held at the NOSQL conference 06/2009

PS: Actually, it is kind of funny, how IBM developed SQL: The original intention was to create a tool for managers to easily retrieve data from DB systems – not a tool for developers.

After the I finished the upgrade, the front-end rendered fine, while I constantly got an error from with ADMIN site: HTTP 500 error.

In case your site runs on a server hosted by 1&1 too, you have to do two things to get ADMIN running again:

1. Create a php.ini in WP’s root directoy and add this single line of text:

memory=20MB

2. Edit WP’s .htaccess file  and add

AddType x-mapp-php5 .php

as the very first line.

Today, I tried to develop a method the create shallow and deep copies of objects. Finally, I prepared this method, initially without the line marked with __MARK__:

static private function clone( source:Object, deep:Boolean = false ):Object
  {
    if ( null == source ) {
    return null;
    }
    var copy:Object;
    if ( true == deep ) {
       copy =  ObjectUtil.copy( source );
    } else {
      copy = new Object();
      for( var key:Object in source ) {
        if (  ( 'mx_internal_uid' != key ) ) {     // __MARK__
          copy[ key ] = source[ key ];
        }
     }
  }
  return copy;
} // clone

Without the line marked by __MARK__, the method copies (more or less raw) objects, but the dataProvider-driven controls gets in trouble: List-based control add each managed object a ‘mx_internal_uid‘ to identify each data object. Thus, if you copy the ‘mx_internal_uid’ property, e.g. a DataGrid will no longer be able to distinguish the original source object or the clone(source) object.

Thus, it’s appropriate to not copy ‘mx_internal_uid‘ from source to clone.

Twitter in plain english

Wikis in plain english

Google Docs in plain english [Likely to get closed down soon?]

Social networking in plain english

RSS in plain english

Social bookmarking in plain english

Blogs in plain english

Podcasting in plain english

Googl Reader in plain english

Finally: Zombies in plain english

And electing a US president in plain english

Most likely, you missed to assign the correct mime-type! Don’t forget, that you need to tell browsers, that files with *.air extensions actually are Adobe AIR installer files.

If you run Apache, adding the correct MIME type is simple:

AddType application/vnd.adobe.air-application-installer-package+zip .air

As always: Big trouble, small solution.Don’t forget, that this setting is absolutely vital! You need to set it – even if some browsers might be able to run the AIR install package with it. Most browser won’t…