Upcoming Apple laptop and iPod drought

Looks like the rumors about Apple redesigning its laptop computers is true. In a recently bulletin Apple has asked resellers to stock up on 3 weeks worth of laptop order and 4 weeks worth of iPod orders. I have heard of Apple’s decision on redesigning the laptops, but the reason for a iPod shortage is not very clear. The latest incarnation of the iPod is relatively new. So chances are less that there would be big changes in the iPods design. A plausible reason might be that Apple would be concentrating most of its manufacturing resources on the laptops thereby starving the demand for iPods.

There have been little reported about the new laptops, Some of the things that we can expect are a one-piece all aluminum body and a glass track pad with finger gestures. Finger gestures made their debut in laptops through Apple’s MacBook Air laptops. The feature no doubt enhances usability but apart from the cool factor adds very little to the performance of the laptops.

Also its rumored that Apple might be going back to its roots and fabricating its own chipsets thus moving away from Intel. This move, in spite of Apple enjoying a one of a kind deal with Intel, is quite interesting.

Indicators point out that the changes are more likely on MacBook line rather than on the MacBook Pro line of laptops. Also there have been no reports of the desktop systems(iMac and Mac Pro) being overhauled.

So why would Apple try to have systems with different architectures in terms of the their chipsets? Having different chipsets makes the job more difficult for the engineers since now they would have to put in more development effort on their software (specially in the Mac OS) to ensure that they work consistently on both platforms.

Maybe Apple plans to move away from Intel in a phased manner and has decided to start with the MacBooks. An experiment of sorts? Could be. Moving their entire portfolio away from Intel would surely point to a company that does not have a very clear idea of where it wants to go.

The new laptops are expected to hit the retail shelves in September this year. I surely suggest that people contemplating buying MacBooks hold on till that time.

Interesting Cuil

This is an addition to my blog entry on Cuil taking on Google. Here are some pretty interesting search results from Cuil and Google. Search for Cuil on Cuil. Here are the resutls:

Surprisingly, searching for Cuil does not list Cuil.com!!! Is this an act of modesty or did the creators in all their technological zeal forget to add their own site to their index?

Now search for Google on Cuil:

Well Cuil seems to believe in playing fair as far as search results go.

The new Apple

A few rumors are doing the rounds with regards to Apple and its line of notebook computers. How far they are true is difficult to say but they are quite interesting.

Firstly, Apple seems to be redesigning the look and feel of its laptops. The present designs are actually pretty dated. The latest design for the laptops incorporate a single piece aluminum body. Also the track pad is going to be glass. Well, trust Apple to push the boundary when it comes to choosing material for constructing its computers. The track pad would respond to finger gestures like the iPhone and more recently MacBook Air. Good but not great. It would rather be more beneficial if the laptop screen behaves like that of the iPhone. Imagine zooming in and pinching photographs right on the screen. A Microsoft Surface like behavior for the screen would be even better.

The second rumor is the more interesting one. Apple, it seems, is looking for alternatives to Intel chipsets for its next generation of laptops and has decided not to go in with the Montevina chipset. Very interesting news indeed. Apple’s switch to Intel is pretty recent and apparently the marriage seems to be already over. But what are the alternatives? AMD chips or the IBM Cell processor? Neither seem very likely. Apple’s acquisition of PA Semi, a mobile CPU developer, suddenly seems to make sense. Is Apple planning to go back to its old ways of using its own chipsets? Does not seem very unlikely given that in the past Apple used to do this. Very recently Psystar tried to sell its own computers with Mac OS X installed. Even though Apple is currently suing the company, Psystar has definitely touched a nerve here. So it may be possible that the Psystar episode has prompted Apple to go back to its original ways.

Lets wait and watch.

Oracle speaks on Kubuntu: Installing Oracle 10g on Kubuntu Linux

Arguably Oracle 10g Enterprise Edition is the most commonly used database for enterprise applications. Financial institutions, government organizations, manufacturing companies, etc. all use this database for their enterprise needs. This blog talks about installing Oracle 10g Enterprise Edition on Kubuntu 8.0.4. Ever since its launch, Kubuntu and its GNOME cousin Ubuntu, have become a very popular flavor of GNU Linux for the desktop. I have been using the Ubuntu versions for quite some time now, and needless to say I have not felt the need to switch back to any other OS ever since.

During my attempts at installing Oracle, I was hugely helped by Tomasz Węgrzanowski blog. He deserves a big thanks. Its available at:
http://t-a-w.blogspot.com/2007/11/installing-oracle-10g-enterprise.html

I followed most of the steps described in Tomasz’s blog, but some I found were not required. So here’s my attempt at a slightly simpler instruction set.

It is assumed that Oracle 10g database software is already downloaded and there is a working version of Kubuntu 8.0.4. These software can be downloaded from:

• Oracle: http://www.oracle.com/technology/software/products/database/oracle10g/htdocs/10201linuxsoft.html
• Kubuntu 8.0.4: http://www.kubuntu.org/getkubuntu

Also, the steps are for the 32-bit version of the database.

My installation target directory was /opt/oracle10g. Make sure that the target installation directory is writable by the user under which you want to install Oracle.

Oracle installation expects a few binaries in certain directories, namely basename and awk. The installer expects both these binaries to be available under /bin. So we need to create a couple of links to these directories:

prosenjit@vito:~$ sudo ln -s /usr/bin/basename /bin/basenameprosenjit@vito:~$ sudo ln -s /usr/bin/awk /bin/awk

You can test these by invoking /bin/awk and /bin/basename from a console window.

Once this is done, its time to run the installer. To do so change into the directory where you have unzipped the Oracle distribution and invoke the following command:

prosenjit@vito:~$ ./runInstaller -ignoresysprereqs

The -ignoresysprereqs flag instructs the installer to ignore the supported Linux distribution check. Since Kubuntu is still not supported by Oracle as an official release platform, this flag is necessary.

Running the installer should show up a GUI. Follow the instructions till the installer asks you to run one or more scripts. For me there was only one script:

/opt/oracle10g/root.sh

The location of the script depends on the target directory you have chosen and also on the operating system. Once this is done, come back to the GUI and perform the next steps. This should install Oracle properly in your system.

To start Oracle certain environment settings have to be done. Usually I prefer to put these entries into my .bashrc file. That way I do not have to do so everytime I want to start or stop Oracle. Here are my settings:

export ORACLE_HOME=/opt/oracle10gexport LD_LIBRARY_PATH=$ORACLE_HOME/libexport ORACLE_OWNR=prosenjitexport PATH=$PATH:$ORACLE_HOME/binexport ORACLE_SID=oradab10 (During installation I had created a databasewith the SID oradab10)

Thats about it. Other than this you might need to do some work on the listener.ora and tnsnames.ora file.

My tnsname.ora file has the following:

oradab10 =(DESCRIPTION =(ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))(CONNECT_DATA = (SERVER = DEDICATED) (SID = oradab10) ))

My listenr.ora has the following entry:

SID_LIST_LISTENER =(SID_LIST =(SID_DESC = (SID_NAME = PLSExtProc) (ORACLE_HOME = /opt/oracle10g) (PROGRAM = extproc))(SID_DESC = (SID_NAME = oradab10) (ORACLE_HOME = /opt/oracle10g)))

To start Oracle first you need to start the listener. Open a console window and type in the following command:

prosenjit@vito:~$ lsnrctl startLSNRCTL for Linux: Version 10.2.0.1.0 - Production on 30-JUL-2008 22:59:26

Copyright (c) 1991, 2005, Oracle.  All rights reserved.

Starting /opt/oracle10g/bin/tnslsnr: please wait...

TNSLSNR for Linux: Version 10.2.0.1.0 - ProductionSystem parameter file is /opt/oracle10g/network/admin/listener.oraLog messages written to /opt/oracle10g/network/log/listener.logListening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC1)))Listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=127.0.0.1)(PORT=1521)))

Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EXTPROC1)))STATUS of the LISTENER------------------------Alias                     LISTENERVersion                   TNSLSNR for Linux: Version 10.2.0.1.0 - ProductionStart Date                30-JUL-2008 22:59:26Uptime                    0 days 0 hr. 0 min. 0 secTrace Level               offSecurity                  ON: Local OS AuthenticationSNMP                      OFFListener Parameter File   /opt/oracle10g/network/admin/listener.oraListener Log File         /opt/oracle10g/network/log/listener.logListening Endpoints Summary...(DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC1)))(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=127.0.0.1)(PORT=1521)))Services Summary...Service "PLSExtProc" has 1 instance(s).Instance "PLSExtProc", status UNKNOWN, has 1 handler(s) for this service...Service "oradab10" has 1 instance(s).Instance "oradab10", status UNKNOWN, has 1 handler(s) for this service...The command completed successfully

Now start the Oracle database with the following command:

prosenjit@vito:~$ dbstartProcessing Database instance "oradab10": log file /opt/oracle10g/startup.log

To verify that the Oracle process was started invoke the following command at the console:

prosenjit@vito:~$ ps -ef|grep ora

You should get an output similar to:

root      5602  5497  0 21:15 ?        00:00:00 hald-addon-storage: polling /dev/scd0 (every 2 sec)1000      6277     1  0 22:59 ?        00:00:00 /opt/oracle10g/bin/tnslsnr LISTENER -inherit1000      6351     1  0 23:00 ?        00:00:00 ora_pmon_oradab101000      6353     1  0 23:00 ?        00:00:00 ora_psp0_oradab101000      6355     1  0 23:00 ?        00:00:00 ora_mman_oradab101000      6357     1  0 23:00 ?        00:00:00 ora_dbw0_oradab101000      6359     1  0 23:00 ?        00:00:00 ora_lgwr_oradab101000      6361     1  0 23:00 ?        00:00:00 ora_ckpt_oradab101000      6363     1  0 23:00 ?        00:00:00 ora_smon_oradab101000      6365     1  0 23:00 ?        00:00:00 ora_reco_oradab101000      6367     1  0 23:00 ?        00:00:00 ora_cjq0_oradab101000      6369     1  0 23:00 ?        00:00:00 ora_mmon_oradab101000      6371     1  0 23:00 ?        00:00:00 ora_mmnl_oradab101000      6373     1  0 23:00 ?        00:00:00 ora_d000_oradab101000      6375     1  0 23:00 ?        00:00:00 ora_s000_oradab101000      6379     1  0 23:00 ?        00:00:00 ora_qmnc_oradab101000      6383     1  1 23:00 ?        00:00:00 ora_j000_oradab101000      6385     1  0 23:00 ?        00:00:00 ora_j001_oradab101000      6387     1  0 23:00 ?        00:00:00 ora_j002_oradab101000      6389     1  0 23:00 ?        00:00:00 ora_j003_oradab101000      6391     1  0 23:00 ?        00:00:00 ora_j004_oradab101000      6395     1  0 23:00 ?        00:00:00 ora_q000_oradab101000      6397     1  0 23:01 ?        00:00:00 ora_q001_oradab101000      6400  6249  0 23:01 pts/1    00:00:00 grep ora

To stop a running instance issue the following command:

prosenjit@vito:~$ dbshut

This will shutdown your database, but the listener is still running. To stop the listener:

prosenjit@vito:~$ lsnrctl stop

LSNRCTL for Linux: Version 10.2.0.1.0 - Production on 30-JUL-2008 23:03:15

Copyright (c) 1991, 2005, Oracle.  All rights reserved.

Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EXTPROC1)))The command completed successfully

Thats pretty much it. Oh, I had faced a problem while starting the Oracle instance when I rebooted the machine. It turned out to be a problem with the /etc/oratab file. The entry in my /etc/oratab file is:

oradab10:/opt/oracle10g:Y

The ending Y seemed to make all the difference.

Hope this helps.

Knol takes on Wikipedia

Yet another war! Are we not already tired of all the wars that are taking place in Iraq, Afghanistan and other places? Not to mention the constant war against terror. Why then another war?

The only difference is this time its on the grounds of technology and new products. Very recently Google announced a new product named Knol. Its supposed to take on the dominance of Wikipedia. Was this necessary? Well a lot of things were not necessary but Google still decided to do those and needless to say they have been very very successful in it.

Google describes Knol as ‘an authoritative article about a specific topic’. I dug around for a while and what I understand is its basically Google’s Wikipedia with a few improvements thrown in. For instance, the credentials of an author are verified in Knol and thus the postings might have more credibility. Another nice thing about Knol is the user comments section. Pretty nice to have. The thing with user comments is that they can very quickly degenerate into meaningless battles. So its a nice to have feature with limited benefits.

A lot of people have criticized Google for reinventing the wheel with its ‘me too’ applications. Well I do not see why someone should not attempt at inventing a better wheel. And Google has been very good with reinventions. Look at Gmail or Google Talk. There were established webmail providers and internet chat services before Google came along with its offerings. Google’s offerings were compelling and now have a huge user base. They thought things differently. Thats what differentiated Google’s apps from the rest. Even Google’s search engine displaced established players like Alta Vista, Lycos and Yahoo!. So Google itself can be termed as a reinvention.

The same does not seem to apply to Knol though. I for one, find Wikipedia to be a very useful site with a lot of information thats really worthwhile. It also offers a very easy and largely effective search facility. In the past a lot of reports have exposed several weaknesses of Wikipedia including spurious postings. Ever since those reports, Wikipedia has tightened things up. But inspite of all this, is a new Wikipedia-like application justified? Is Knol really immune to the problems that Wikipedia faced? Only time will tell.

I would have really liked Google to come up with a product that would have enabled Wikipedia to be more authoritative, credible and secure. Why not add Google’s method of verifying credentials to Wikipedia? Why not add the user comments feature to Wikipedia? Why not get Google to collaborate more with the community and thus generate better results? Why not leverage the already vast article base on Wikipedia rather then creating the same?

Even though most of Google’s products are free to use, as a company they don’t seem to be interested in contributing to existing efforts in the community. Rather they prefer to go their own way. Maybe thats their strategy. Maybe thats their goal. Maybe they intend to become to the internet what Microsoft is to the desktop world.

Just google Cuil

So there’s yet one more new search engine on the horizon. Its Cuil (pronounced Cool, don’t know how that works). Anyway. Cuil makes certain tall claims. Its got a new way of ranking pages that takes the content of pages into consideration rather than other aspects that were or are used by other search engines. It also claims that it has the highest number of pages in its index, even more than Google and Yahoo!. I remember about 4 years back Yahoo! having released the number of pages in its index. Ever since there haven’t been any reports of such numbers from any of the leading search engines.

Nevertheless, it was worth a try since it comes from people who have very impressive credentials. The leading team consists of Ph.D.-s who have worked for such technology giants like Google and IBM. So their product is definitely going to be interesting.

The first impressions were not so bad. The search engine is fast and it has a very interesting way of showing the search results. Here’s a screen shot:

Pretty cool or cuil, eh!!! Gives you a feel of browsing through a catalog – quite refreshing after the boring result pages of Google. My interest grew and so I thought of taking it for a test ride.

After having used it for sometime my reactions are mixed. Like I mentioned before, the presentation of the results is really a refreshing change. But the site leaves a lot to be desired. For one, it does not seem to be stable. Maybe its going through the initial jitters. I wanted to search for JBoss Seam integration in the NetBeans IDE. Well here are the results:

Wow! No results! Disappointing. Wait there’s more. I did a few more searches and then thought of searching for this one once more. And here’s what I get:

What’s going on? Clearly it shows that the search engine is far from stable. Well if its not stable, then it should have at least the tag ‘Beta’ somewhere on it. The home page of Cuil does not have the word ‘Beta’ anywhere. So a person would expect this to be a fully baked, production version of the product. Unfortunately it does not seem to be so. Also, Google provides a nifty feature where it tries to guess the correct spelling for a potentially misspelled word – something missing in Cuil.

The relevance of the search results is also quite questionable. I searched for the phrase ‘jboss seam projects on netbeans ide’ on both Cuil and Google. Clearly, Google showed up the more relevant hits. Here are the screen shots:

On Cuil:

Now on Google:

Even though Cuil has still a long way to go if it intends to pose a serious challenge to Google, I still liked their approach. The software’s got potential and quite a few things going in favor of it. What it needs is to tighten its algorithm to get more relevant results and beef up its infrastructure to have more stability.

Keep an eye out for this one. Amongst all the previous Google killers that came up, this one shows the most promise.

Eclipse’s edge eclipsed? Problems with Eclipse Ganymede on Kubuntu Linux

The other day I started working on a JEE project of mine. Usually I use Eclipse JEE distribution on Linux for this. I setup a spanking new box with Kubuntu 8.0.4, Eclipse 3.4 Ganymede and JBoss AS 4x. The installations went fine and I fired up Eclipse to write the first line of code. Initial impressions were pretty good. I checked for updates to the Ganymede distribution through the Eclipse UI, found some and installed them. Thus I ended my first day with the peace of mind of having an environment that has been setup properly. 

I was in for a nasty surprise the next day. I started Eclipse and the thing would just not startup. Even the dialog box to select the workspace did not appear. A message about some error and a location of the error log was all that I got. I looked for the log and the file did not exist. I tried the -clean and the -debug flags and saw a permission denied error being reported. Why was this error not reported the previous day? Anyway, I corrected the permissions and fired up the IDE once more, and got an error saying too many open files. The max number of open file handles is an OS parameter on UNIX. I looked up the Kubuntu forums, figured out how to increase this limit, increased it and restarted Eclipse. The same damn error!!!

Frustrated I removed the previous installation, and along with it all the other plugins like Subclipse and JBoss Tools, and installed the software afresh. Installed the relevant plugins and fired it up. It worked like a charm. This time I did not update the distribution. Anyway, it was getting pretty late. So that was all for the second day.

The third day was more promising. Eclipse started up without any problems. I did some reverse engineering with JBoss Seam through the JBoss Seam perspective. Seam generates a JEE app out of a data model. Once that was done I switched to the JEE perspective. I double clicked on a .xhtml file and it opened with the default JBoss JSP editor. Now this editor has a code section at the top and a preview section at the bottom. Somehow I prefer a simpler editor. But once you install the JBoss Tools, the default editor switches to the JBoss one. In order to open a file with any other editor usually you have to right click and then select an option in the Open With context menu. When I tried to do so no context menu appeared. I tried right-clicking on some of the other artifacts and it showed up for some and not for the majority. Now this was FRUSTRATING!!! I switched to the Seam perpective and surprisingly the context menu was working without any hassles. What went wrong? Short and sweet, I do not know. Setting the default JSP editor in the Prefrences screen did not solve the problem either. Uninstalling the plugins is not a solution that I would accept since I need those.

So right now I am stuck with an Eclipse IDE that works partially; I would really prefer to have something that works better.

I have been a long time Eclipse user and thought very highly of this IDE till recently. The last two releases, Europa and Ganymede, have been big disappointments. I have used them on Linux and Windows systems. The IDE has become more prone to crashes and freeze-ups. I hope the Eclipse team does something about these problems.

This frustration with Eclipse, led me to try out the NetBean 6.1 IDE. Well, I have to admit that its a very very good IDE and in certain areas actually scores more than Eclipse. NetBeans had always been a very heavy and slow to respond software (because of Swing probably). Can’t say that anymore for the 6.1 version. It was snappy and worked smoothly.

Would I shift to it? Probably not immediately, but if Eclipse does not clean up its act, there would be quite a few converts for sure.