Updated specifications for 13 inch MacBook Pros (old configuration in brackets)

Updated specifications for 15 inch MacBook Pros (old configuration in brackets)

As you can see from the configuration above,  anyone who recently bought the 13" 2.26GHz MacBook Pro must be cursing themselves.

It is interesting to note that even the high end MacBook Pros have the total memory capped at 4GB. In fact now all configurations carry 4GB. There are two reasons – 4GB RAM is usually good enough for most users and secondly DDR3 RAM is not cheap. If you really want to upgrade to 8GB,  you need to spend a whopping $400 extra!

Ever wondered what is the most expensive MacBook Pro configuration you can buy? Well, you can get the best of the best for a low price of $7,505. This configuration includes a 24 inch LED cinema display and a 512GB solid state drive. Obviously many of us would have to mortgage our house to get one of these!

Most Expensive MacBook Pro (Configuration)

The advantages of Jailbreaking iPad

Jailbreaking iPad opens up more possibilities than on an iPhone. iPad has better hardware and bigger screen and hence it can replace a netbook with much better battery backup. We may see custom Linux distributions capable of running on iPad. The immediate benefit of course is that iPad users can download applications rejected by Apple from other app directories such as Cydia. It also enables complete customization of iPad OS. Who knows, someday someone may even find a way to run Windows XP on iPad!

However the lack of some of the essential hardware on iPad makes jailbreaking less appealing. For example, lack of an integrated camera.

How blackra1n works on iPhone?

Blackra1n works on a specific exploit on iPhone/iPad devices known as the USB Control Message exploit. By manipulating the call to function usb_control_msg it is possible to overwrite sections of memory.  See this page for the details of this. However in the case of iPad, the precise exploit that Geohot used is not yet known.

Apple is refusing to allow flash in its devices claiming that its implementation is buggy and it is not suitable for its devices. So the only option to develop software for iPhone, iPod Touch and iPad was to use the Xcode and iPhone SDK on Mac machines. It also meant that developers need to learn Objective C to build iPhone apps. Adobe thought that they can work around these limitations by providing a cross compiler which will convert Flash programs directly into iPhone’s ARM binary code. Till now Apple approved any app submitted to App store provided it met all its guidelines and hence a number of apps built with a preview release of Flash CS5 got into app store.

With iPhone OS 4.0 onwards this whole cross compiler process becomes illegal. Adobe may  probably release iPhone cross compiler as part of Flash Processional CS5. But any developer building an iPhone/iPad application on it will be taking a huge risk since Apple may reject it during App Store submission. Technically there is nothing that can differentiate between the ARM code generated by Xcode and Flash Professional. Apple can insert a unique signature to code, but the same can be easily inserted by Flash professional. But even releasing the cross compiler may be illegal.

Now I think Adobe deserves what it got. They could have directly contacted Apple before going ahead and developing a cross compiler. Things like these should teach giant companies like Adobe that they can’t be arrogant all the time. One day Apple may also get a taste of its own medicine. 

Obviously what Apple did makes a lot of business sense. They want people to use their own platform to build apps. Proliferation of cross compilers is not a good thing in long run. It can lead to inefficient code(?) as well. But the real reason seems to be that Apple wants complete control over its platform. They don’t want somebody to create a better development platform for building apps in iPhone and iPad. It also shows how easy it is to decimate the competition when you put restrictions on a platform. I now wonder why Microsoft was harassed by various governments all these years. Compared to Apple, they are angels

There is one more interesting aspect to the Apple/Adobe tussle. There are a lot of Flash game developers out there. It is not easy to compete with them if you are a pure Objective C developer building games for iPhone. So this is a good news for all the Objective C developers out there!

Unlike traditional search engine results, the Wolfram output is structured. This means that processing output from it is easy and is extremely useful in large scan data analysis. So an API interface from Wolfram Alpha was eagerly awaited and it was released last week. But the pricing of the API is one of the most ridiculous I have seen in a while. The following is the pricing for personal use (for hobbyists etc.).

What?  $60 dollars for accessing Wolfram API for 1000 times?  I can’t believe that someone in Wolfram could be so retarded to charge almost one dollar for 10  hobbyist requests for the API.  Then you look at companies like Google and think – Man, these are angels from heaven!

Now checkout the terms of use section of API which is worse than many of the free online services(you thought paying through your noise would give you better TOS?). Some examples,

• Wolfram|Alpha reserves the right to terminate this Agreement or discontinue the Wolfram|Alpha API or any portion or feature thereof for any or no reason and at any time without liability to You.
• You are required to provide a hyperlink to www.WolframAlpha.com on every page with Results. All hyperlinks associated with data provided by Wolfram|Alpha must navigate directly to Wolfram|Alpha.
• utilize the Wolfram|Alpha API in association with, or as a component of, any website that in the sole discretion of Wolfram|Alpha is determined to be obscene or otherwise inappropriate.

Here is what I think is reasonable pricing,

• Up to 5000 requests – free
• Up to 25,000 requests – $50
• Up to 100,000 requests – $150
• Up to 500,000 requests – $500
• Additional request pricing – Below $2 per 1000 requests

So here is the one liner verdict – Wolfram Alpha API pricing sucks.

The Google Chrome initial installer is under 500kb and rest of the installation is done by downloading other components from the initial installer. On my 2mbps broadband line it took under 5 minutes for complete installation. The configuration took just the following step,

The Chrome user interface is simple, elegant and extremely fast. It is very unobtrusive, just what an ideal browser should look like.

Google Chrome is built from ground up and hence is able to address a lot of issues plaguing current browsers such as Internet Explorer, Firefox and Opera. Here is a summary of major Chrome features,

1. Single text field for URL and Search – Type in the URL if you know it, else search the Web. The default search engine obviously is Google, but you can easily change it.

2. Browser tabs as separate processes – Instead of threads, Google Chrome uses separate processes and hence provides greater control of each tab and preventing possible crash of the entire browser due to a problem in a Web page in a single tab.

3. Enhanced developer features – This includes JavaScript debugging and control features, a task manager and a JavaScript console. The JavaScript console even provides auto completion. I guess this will be an area where we can find tremendous improvements in coming versions. Firefox Firebug, beware!

Why is Google going after the Browser market?

The majority of Google revenue comes from search and there is a risk that browsers such as Internet Explorer will try to divert users from it. Already IE sets Windows live as default search engine. The best way to fight this is to fight Internet Explorer itself!

Secondly an advanced browser provides an ability to roll out advanced features to Google products. Google can provide a degraded behavior to other browser users for their advanced product features. This enables Google spreadsheet and document editor applications to compete with Microsoft Office etc.  Hence Microsoft can expect intense competition in this area.

My Verdict

I think it is too early to say how Google Chrome will evolve. But one thing is sure, it will soon take up a substantial pie of the browser market. Also I think in the long run, Chrome has the potential to host complex Web applications taking over a lot functionality currently provided by operating system and desktop applications. And for the Web developers, it may become a very good development tool for advanced Ajax based applications.

Further Resources

• Download Google Chrome
• Get Google Chrome Comic (Very Informative)
• Complete Google Chrome feature list

Websense is reporting that spammers have finally managed to write captcha decoding programs for Hotmail which takes only few seconds to decode the characters! The success rate is over 10% and is good enough for spammers.  Here is how spammers make money from this,

1. Spammer releases the automatic signup bot as a virus.

2. An unprotected machine on the Internet gets infected by this bot virus.

3. Bot creates multiple hotmail accounts from victim’s machine using captcha decoding logic.

4. Bot sends advertising messages to multiple mail addresses using the newly created hotmail account.

5. When millions for spam messages are sent, fairly good percentage of them brings business and spammer makes money through affiliate cut.

Whoever wrote this must be pretty good since the Hotmail captcha is pretty difficult interpret even for a human. For example, consider the captach (displayed on the right side) I got when I tried to signup to Hotmail!

Weak captchas have caused havoc earlier. For example, PhpBB forum software captcha was so weak that was hacked in a few days. So if you have a popular forum which is running on PhpBB 2.0, you will soon find majority of new users are from automated signups! It was complicated by the fact that a signed up account can have a live link in the profile.

I think one way to handle the bots will be to have very big set of different captcha styles which will reduce the success rate to below 0.01%. Another way will be to introduce new type of captcha every day so that any decoding program will become obsolete in a day!

A machine gets infected when the user of the system accesses Websites intended to spread the virus (such as warez or illegal downloads). Mebroot uses Internet Explorer vulnerability to write directly to the Master Boot Record of the machine! The trojan itself is around 450kb in size and is stored in the last sectors of the harddisk. It then creates a backdoor on the machine.

Once the backdoor has been established, the program looks for any user access to internet banking sites. It then sends the captured banking userid/password etc. to a thirdparty site! Pretty impressive, eh?

It is estimated that over 5000 machines are affected by this virus.

The easiest way to remove this virus is to run “fixmbr” command from Windows recovery console. This overwrites the virus entry on MBR. Also some of the latest BIOS setting allows you to make MBR readonly. Any modification to MBR will throw a BIOS warning! So enable MBR protection today.

Lastly never visit any Website which offers warez, cracks, serials or free downloads. The real purpose of most of these sites are to spread keyloggers and other types of viruses. If you really want to check those sites, create a virtual pc using vmware or windows virtual pc exclusively for that purpose.

HECToR contains over 5000 AMD 2.8GHZ dual core Opteron processors the peak performance is 63Tflops. The total memory of HECToR is 33TB (6GB per processor)!! The storage used is high performance RAID disks with a total capacity of 576TB. HECToR uses Unicos/lc as the operating system and has two components – compute node linux for compute nodes and linux distro(SUSE Linux) for service nodes. The communication network uses 6 sets of Seastar2 chip based system and each link offers bi-directional bandwidth of 7.6GB/sec!

HECToR (High-End Computing Terascale Resource) is a project funded by research councils in UK and this facility is available to researchers in UK Universities. The main use of this supercomputer will be in simulation of complex natural phenomena such as climate change prediction.

HECToR is composed of 66 individual units as shown below and each of the unit is as big as a large shelf!

You can see the current status of HECToR supercomputer here!

Google has released 24 language translator bot in Google Talk. These are named in using language codes – for example a contact named en2fr will translate from English to French. You just need to add en2fr as a friend and then anything you type to en2fr will be translated to french instantly!

Since this service uses XMPP open protocol, you can write different applications to make use of the underlying engine. For example, you can enable this chat translation facility on your own site using the Google talk engine.

Machine translation comes with its own risks, so don’t try it on your date

Another advantage is that your URL can be automatically inserted in the comment. For this to happen, use your blog or site as your OpenID identity. This doesn’t mean that you have to have your own OpenID provider. You can use other providers such as Verisign or Technorati and then delegate to it from your blog! This means that you can change your provider without affecting your identity (your blog).

Now it is not necessary that you should use a third party OpenID provider. You can be your own OpenID provider! For details, check Setting up OpenID.

How does OpenID work?

When you use OpenID, there are 3 players. You (Y), OpenID service provider(P) and the site you are trying to login(T) using OpenID. When you indicate your OpenID identity at the T, it will handshake with P to get a secret key. Then you are re-directed to the P’s Web site to login. After login, the same secret key is prepared using your password by P and it returns it to T. Now T can compare this with the original secret key. If they are same, that means you did authenticate at the P!  Simple isn’t it?

Solid state drives are here (which don’t have any moving parts) which eliminates this problem! Today I came across an Alienware laptop advertisement which features a 64GB solid state drive! The good thing is that it consumes 50% less power and is supposed to be faster than an ordinary harddrive!!

Yeah, this thing comes with a price. The Alienware configuration section gives an additional $1100 for a 64GB solid state drive ($500 for 32GB). I think the best option will be to use 32GB solid state drive with a portable harddisk for your long term storage needs.

The main advantages of solid state drives are,

1. No moving parts and hence highly reliable.
2. Faster booting time (2x faster than fast harddisks)
3. Power efficient and increases battery life.
4. Quiet operation.

I think we can soon see other brands (dell, thoshiba etc.) coming up with solid state harddisks. Well, I won’t be buying this until the extra price comes below $100.

This means that Google has such good infrastructure that they can do almost realtime crawling of Websites! But how do they do that? Obviously they can’t crawl all the sites every minute (which will take a lot of bandwidth from individual sites).

The secret probably is with blog pings and sitemaps. Already google blog search, technorati etc. use the pings to update the content. This means that only incremental data is crawled. Another feature that can be used is sitemaps which also will help google in doing incremental crawling. You can also add your blog/website sitemap in google by creating an account at Google webmaster tools.

Let us do some math here. According to Technorati there are over 100 million blogs. Now if we assume 1 post per day and an average size of 10KB per post we have,

Total content = 100 million x 10KB = 1 TeraByte of transfer!

This means that just to crawl blogs using incremental techniques, Google search engine will be using a minimum of 1 Terabyte per day of bandwidth! Here we assumed that only new posts are crawled. In reality, Google has to crawl old posts also to see whether they have changed! In Sitemaps, you can define that as daily/weekly/monthly/yearly crawl.

Check out my sitemap console screenshot below. As you can see, it is possible to reduce the crawl speed to save on the bandwidth. Now in that case you content will not be near “realtime”.

Now they are providing FeedBurner integration with BlogSpot blogs! This means that if you have a BlogSpot blog, you can redirect the default feed to FeedBurner. This allows for a single feed stream and tracking your feed visitors is easier!

To configure all feed traffic to go through FeedBurner, add the FeedBurner URL in Site Settings -> Site Feed -> Post Feed Redirect URL field. This ensures that any feed access is redirected to FeedBurner.

What happens when FeedBurner tries to access feed? Well, BlogSpot can detect this from user agent and disable redirection only for FeedBurner server access!

Now the next thing I expect is the integration of FeedBurner Ad Network with Adsense. What they can do is to roll out FAN to all the existing Adsense publishers. Currently FAN is open to only select few who has more than 1000 or so feed subscribers.

Here are some of the key new features in Pligg 9.5,

1. Un-vote – Now users can un-vote stories they have voted for

2. Fixes for IIS – Fixes for the IIS cookie redirect bug.

3. Image upload module – Upload images and use the thickbox javascript library to view them.

4. Reporting/Burying Stories – Users can bury stories that are spam, inaccurate etc etc. Burying a story grays it out to signify you have buried it.

5. Major database query enhancements – A quick estimation shows that the database calls in version 9.1 was 170+ and now it’s 50-60. This should substantially increase the speed of your Pligg site.

6. Story page tabs – Each story page now uses tabs. There are 3 tabs, comments, who voted, and a new feature, related stories.

7. Upgrade to Template Lite 2.1 – Upgraded Pligg to the current version of Template Lite.

8. Upgrade to ezSQL 2.03 – upgraded to the latest version of ezSQL for better mysql query handling.

Now the only thing Google Adsense need to do in India is to enable direct bank payments. This will be a boon for Indian Adsense publishers since current courier fee of $25 is too much in Indian context. Especially when your Adsense income hovers around the $100 mark!