Funny story about Microsoft Recruiting

I was approached by MS a total of 4 times in the past month and a half ... each time they forgot that they've spoken to me before :)

First time was by an outside recruiter for Microsoft who found my resume and made big promises to get me interviews with a couple of interesting groups: Core File Systems and some high profile Distributed Systems group. Of course he didn't realize I was still a university student and based on some MS rules he wasn't allowed to talk to me ... so that whole thing was off.

Then a few days later I was contacted by another recruiter, this one from MS, and was invited for an interview; Considering I had an offer for a good position at Amazon, at the time I decided to decline that invitation ... it seemed that this would be the last I would hear from them for a while.

Not to be. I was contacted again by another MS recruiter a couple of weeks later ... it was the same story: they found my resume in their database, or something like that. So I thought to myself - wow, that's pretty weird - I thought about it and decided what the hell, just for the heck of it I would interview with them, maybe get some interesting puzzles, but only for those positions that I was offered by the first recruiter ... I knew those would be nearly impossible to get: given that I'm a new graduate with no previous experience at Microsoft, and acording to a friend of mine who tried to get into one of those groups, there were no positions available for that group.

Anyways, I interviewed with the HR person; she asked me a lot of questions, including why I chose not to interview with them earlier, etc ... etc ... To make a long story short, in a few days I was told that they weren't really interested ... no big surprises here.
It was funny though how in the email she said that she 'would like' to provide further details on their reasons, but 'unfortunately' they have 'a policy' of not telling people that information :)

Anyways, I thought this would be it ... the last I've heard from them for a while ... but, Noooo ... just today, to a shocking surprise, I got another email from another MS recruiter: who seemed to have found my resume on the web. Here is a little, rather funny, quote from the email with some highlighting :)

During a recent search on the web, I found your resume and immediately shared it with some of my managers. I am a Recruiter at Microsoft and your background is similar to candidates we have been interviewing ...

This one even came with no Subject line ... like a spam ;)

I'm not sure what system they are using over there ... but this is the 2nd time they forgot that they've spoken to me before ... doesn't seem too reliable :P

Anyways, me and my roomates had a good laugh over the whole thing ... so I decided to share this story.

Why Google Movies makes sense for Google

By providing this feature Google will allow users, who are interested in finding out information about movies, easily use their site to search instead of going to websites like rottentomatoes.com, imdb or local movie sites and movie theater sites; of course a clean interface really helps them here again.

Therefore they generate more traffic to their site and keep the Google search brand in the user's mind whenever he/she want to search for anything; this is very important especially now that they have fierce competition in the search market. In addition they will have a later option of possibly adding advertisements or finding other means to make money of this new traffic.

What could be next?
It seems logical that along the same path, they could integrate general tickets information (ie. music concerts, shows, events ... ) into their search: informing people of different local events going on.

Controversy over Google AutoLink Feature

Yesterday, Google released a new version of its Toolbar for the IE browser.

There seems to be some controversy over a feature in that toolbar - the 'AutoLink' feature .
Apparently it is very similar to what Microsoft introduced in the past called "SmartTags" which were rejected by the community and forced out. The argument against it included the fact that the webpage content was being manipulated, may have been forced on users(?) and certain sites were favored.

This feature, if enabled - Google Toolbar seems to provide that option -, changes some content on webpages you browse to point to certain links like Google Maps for location, Amazon for ISBN information, etc ... some in the blogosphere have claimed this to be an anti-competitive practice that harms other companies that are left out (eg. competitors like Mapquest, other book publishers).

Updated 20/02/05: This story has now been slashdoted and there is an interesting discussion on potential patent problems given that MS patented SmartTags in 2000. Also Scobles has some interesting comments on this.

SHA-1 is Broken!

Wow! Another secure hash function cryptographic scheme bites the dust!
According to this, a team from China has broken SHA-1 in 2^69 operations.
Only a few months ago, the same team has broken MD4, MD5, HAVAL-128 and RIPEMD: a huge accomplishment at the time.

Until now SHA-1 was believed to be secure, unlike its close predecessor SHA: which was previously found to be insecure. This result will have a major effect on digital signature schemes that were using SHA-1 for signing messages.

Updated 20/02/05: Bruce Schneier, an expert in cryptography, wrote an interesting post about cryptographic hash functions, some more details about consequences of this result, what's being done about it in the industry and potential replacements for SHA-1

Folksonomies and internet self-categorization

I've been running into folksonomies a lot lately and it seems to be a common, new old, trend on the internet these days. There is a good article at Salon that summarizes where this (a.k.a tagging) is being used on the internet today and what great ideas have come as the result.

Flikr has been using them to categorize the images and improve searching. Speaking of which, one amazing related idea was by a creator of a game called ESP Game, which pairs users and presents them with a random image from the net and the users then try to describe this image in words. Upon a matching keyword you get points and move on to the next image; this allows to accurately categorize images; there is an experimental image search engine on that site ... quite amazing!

Finally another great idea is the 43 things social network site. It uses folksonomies to group people based on their goals and thus find people who have goals in common. This presents with a unique opportunity to work together and help each other to reach those goals. The idea is so incredible, I couldn't resist and signed up for it myself :)

This brings me back to the internet, there have been proposals in the research community, and to some extent examples can be already found on the net, for a the next web - Semantic Web - that organizes, categorizes and intelligently interlinks content together. This is supposed to be done externally by following specific standards; however, I wonder if there is a way through these folksonomies to somehow at least partially accomplish this internally through joint user efforts the like we've been seeing recently.

Yahoo Toolbar for Firefox

I've decided to try out the newly released Yahoo toolbar for the Firefox browser yesterday.
Apparently there was also a version for IE before but I guess I haven't used IE in a while so wasn't aware of it.

In any case I really like what they did with the toolbar. It is highly customizable; for instance you can setup the services to be icons or short links w/ icons, you can choose which services you want to show up on the toolbar; the list is pretty long and you can even add your own buttons (unfortunately only two though). Here is a short list of services: yahoo search, yahoo email, translation tool, Stock quotes, local search, maps, yellow pages, etc.

I think I'll keep it as long as it won't hinder the performance of the browser on my slow comp.

How To Make Google News Profitable

I ran into an acquaintance today and our conversation steered into a discussion about Google News and how according to some interviews/articles he has read, Google wasn't sure how to make money off of their Google News product.

I thought about our discussion and came up with some ideas that may work.
The basic idea is to get news sites/agencies to pay some small amount per click (ie. like Adwords) in a new sponsored links section.

Why would they pay?
Because these days there are a lot of news agencies reporting about the same story and the competition for luring the reader seems fierce between them. Each agency wants to generate as much traffic on their site as possible to make their own money off of advertisements, keep advertisers happy and increase/maintain popularity of their news site/product. By advertising with Google News or Yahoo News through sponsored links section they can achieve this.

How can this be integrated into the site without introducing media bias?
This can be done by distinguishing regular links and sponsored links: as done with Search. For instance, right now there is a green line of news sources under each article in Google News. One can add another line (eg. purple) containing a few sponsored links just below. I'm sure the UI experts at Google and Yahoo can find a neater way to integrate those links, but that's the gist of the idea.

How to make the sponsored links dynamic and avoid duplicates?
First it might be possible that the same link would come up in the 'green' and 'purple' sections; there are many ways to fix this, if at all (eg. one way is to remove the sponsored link or leave it and not charge for having it there).
Secondly to make the sponsored, 'purple', links section dynamic, so that there will be a point for different news agencies to compete and pay for those links, one can integrate the original article relevance algorithm (ie. presumably one is used to sort the 'green results') with the rankings of the advertised links.

Its possible I haven't covered some aspects of the problem, but essentially this is the idea ... would this work?

Distributed Apps and Significance of Google Maps

With every passing day it seems more clear that distributed applications and especially web applications are the future of the application world. Only yesterday, Google introduced its Maps product, which once again has a great look and feel to it and pushes the limits of DHTML.

By pushing these boundaries, especially with their web GUIs, and reducing response time Google is setting a standard for web applications. By doing so, they are helping push HTML as the API of choice for many applications. Certainly those applications that have a self contained data storage component, which is shared by the users falls in this category (eg. Maps, book lookup, what about an RSS reader?, etc), as well, some of the applications that used to be thought of as desktop only apps, have now been turning to the HTML world (eg. email client, bug-tracking ... ).

This trend is changing the desktop world as we knew it in the past. This is also one of big reasons why Google is such a threat to Microsoft and potentially other desktop reliant companies; since by pushing web technology, Google is inevitably pushing its development and improvement by creating the demand for richer web technologies; not to mention causing more businesses to consider the web as the medium of choice.

References
-----------
Joel Spolsky - How Microsoft Lost the API war

Google Maps the World :)

Google has released Google Maps (beta) today.

This is a competing product to mapquest, which allows you to get driving directions, map any location and lookup any businesses in the area (eg. local pizza places) amoung other things.

I have to say I am very impressed by this product; the GUI is absolutely amazing and equally important, it is very fast (ie. doesn't refresh the page each change you make) . The map feature on the page is a really good idea and seems to be done with DHTML: impressive! The product is also very well integrated with search and has tons of good features.

It seems to be a bit buggy for now: so not always reliable for fastest directions, but I can't wait until its been around for a while and most bugs have been fixed.

Update: Here is a good link to find out some details on how the whole thing is done in DHTML

Back in Waterloo land

I am back in Waterloo this term for my last semester at UW.
I'm soon to be officially a Computer Science major; I can't believe its been 4.5 years already! %) ... I love this university and it will be hard to leave it behind :( ... but I guess its time to move on.

This term I'm living with two good friends; Ilyia and Sasha (see below). So far its been a lot of fun 8)