Manas Tungare

My Blog

A Meeting with the Father of the Internet

It's not often that one gets to be in a meeting with Vinton Cerf — who's credited as the "Father of the Internet", and holds the official job title of "Chief Internet Evangelist" at Google. (No, I'm not kidding.)

Vinton Cerf, Father of the Internet

So when I was invited to a research meeting with him, my mentor Bill Schilit, and others at Google, I was totally in awe. Of course I can't discuss what we talked, but the little kid in me was awe-struck enough to want to write a blog post simply mentioning it! ;)

I remember having seen him first about 9 years ago. I was a sophomore at Bombay, and I heard from the ACM community that Vint Cerf was to give a talk at SNDT, Churchgate. It was examination time, and as hard as I tried, I couldn't get anyone else enthused enough to make the hour-long journey to listen to Vint Cerf's talk. In an engineering school with over 800 students, I had expected to find at least a few takers. None. Nada. Zilch. Everyone was too concerned about their examinations to find time to listen to the Father of the Internet. I gave up, took the train, and went to the talk, all alone. It was totally worth it, I recollect his ambitious Interplanetary Internet project back when he was at MCI.

I had never imagined that 9 years later, I would be attending a meeting with him. It's not a dream come true, because I had never even dreamt it would be possible to share a table with Vinton Cerf.

Read More

A Proposal to Integrate Site-Specific Search Boxes into Browser Chrome

Why do I have to search for the search box on any site I visit, before I can type my query into it? Given that almost every well-designed site has a search field, and it has been recommended as a good usability practice since 2001, why is it sometimes hidden deep inside the layout? Here is a suggestion for a change in the browser UI that will enable users to find the search box faster. Even faster than other suggestions so far. It only involves a little semantic markup on part of page authors and some redesign on part of browser makers.

The search box within the browser is underused.

The search box is one of the few basic design patterns omnipresent on the Web. It is also a de facto usability practice to place this search box towards the top right corner of the page. Yet, every site has it at a slightly different location (and some, even towards the bottom.) The user needs to search visually for the search box, or at least glance around until she finds it.

During this entire time, the search box within the browser chrome typically lies unused. It has a default search engine defined where it directs all queries typed into it. Why can't the currently loaded website take advantage of this in-built search field for its own site-restricted search?

How it would work:

When a user is browsing a site that supports this feature, any searches conducted using the browser search box will send the queries to the site in question, and be able to display search results directly. When no site is loaded, the queries are directed to the user's default search engine, just as it is now.

This proposal does not require any significant changes to any markup language -- all that is needed is to enhance the markup with semantic knowledge, and microformats are just the answer! Simply marking up a <form> element with the CSS class "search" should be enough to tell the browser that this is a search form that should be promoted to the browser's search box chrome.

Prior work on similar problems

HTML 3.2 (yes, 3.2) defines a link relationship type for search pages. Adding <a href="/search" rel="search"> indicates that the outgoing link is to a search page. Browsers currently don't do much with this information (please correct me if you are aware of a browser that does something intelligent with this information). OpenSearch is (in their words) a simple format for the sharing of search results. Useful as it is, OpenSearch is more geared towards large-scale general search engines, and browser makers are adopting it as a standard for letting their users pick and install search plugins in browsers.

However, being able to customize the search field on a per-site basis with zero configuration on part of the user is not addressed by any of these proposals.

Addressing Potential Criticism

A few critics might argue that such usage dilutes the purpose of the search field ("It's meant for searching via a search engine, not on a per-site basis") or be concerned about possible user confusion ("Is my query being sent to a search engine or to the site I'm now visiting?").

Considering an intentionality-driven approach to design, this UI is perfectly aligned with the user's intentions. If a page has been rendered in a browser window, the user's intention is likely to search within that site. If the user wanted to perform a search using a search engine, there is always the possibility of loading a new tab and then searching via the same field.

Issues of Mode

This also brings up the question of whether such a UI is inherently mode-based. (Modes in a User Interface are said to exist when a single input can result in two or more possible outcomes, depending upon the state of the UI at that point. It is generally considered bad design to employ modes in a UI because it invariably leads to user confusion.) In this case, it is arguable whether or not this UI employs modes.

The user task is "to search", and the current site can be considered a specialized search engine (the specialization is that they only search within their own site). Given that a lot of sites employ site-limited versions of generalized search engines (e.g. Google's Custom Search Engines), this notion is not very hard to think of. Hence, I argue that these are not two different modes, but two different search providers for the same box, just as current implementations offer users a choice among Google, Yahoo, Altavista and others.

Indicating the currently active search providers in browser chrome

It is also easy to indicate the destination of the search queries in a visually accessible format. Safari (for example) displays the word "Google" in the search field. When a site-search box is displayed in its place, it is trivial to display the name of the site instead of the word "Google". It is also trivial to reuse the favicon for the same purpose.

Why not?

What do you think about this idea? Comments, suggestions, enhancements appreciated!

Read More

Virginia Tech Shootout

A terrible piece of news was announced to us this morning while writing a test. An unidentified gunman opened fire on two separate occasions today, killing one person at Ambler Johnston and several more at Norris Hall. The current fatality count stands at 22 — now updated to 32 — this is absolutely, positively horrifying, especially for a rural campus like Virginia Tech.

Two bomb threats, last one two days ago

The campus has been plagued by law-and-order trouble most of last year; we had two bomb threats for buildings on campus, on April 2 and April 13 (just two days before today's shootout!). On both occasions, the respective buildings were immediately evacuated and all events suspended until police could conduct a more thorough inspection of each building. Following the last threat on Friday, the three buildings were closed until Sunday night. The university sent a campus-wide email Sunday evening that those buildings would open at 7:00am Monday and all events scheduled for today would resume.

Second shooting led to more fatalities

So we went off at 9:00am, right after the first shooting. At that time, if we had known about the first incident that occurred at 7:15am, we sure as hell wouldn't have ventured out. We only heard about it later, some time around 9:26am, via email from the University. While we were on the bus to campus, it was probably exactly that time that the killer moved from Ambler Johnston Hall to Norris Hall, where the second shoutout occurred. Most of the fatalities occurred at the second shooting. If that could have been prevented — oh well, hindsight is always 20/20.

First, curfew; then, evacuation

The first email from the University asked us to stay indoors, lock the doors, and away from windows. The second one told us that the campus was being evacuated, and we should leave the building as soon as possible. We left at around 11:30am from Whittemore hall, and a classmate offered to drop us back home. The status of the Blacksburg Transit bus service was not known.

Shock at home

The real shock came when we reached home, because accurate information about the happenings of the morning was not available until after President Steger's statement released at noon. The sheer number of fatalities and injuries — 22 dead, 21 injuries — was mind-boggling. It is expected that the death toll may rise as more information is available.

Is anyone going to do something about this?

Or should we just expect the National Rifle Association to convene at Blacksburg later this month? If you don't know what I'm talking about, please see Michael Moore's documentary, Bowling for Columbine. I've written about this topic before: it's high time someone does something about this.

22 32 innocent lives lost, not to mention all the victims over the years, is reason enough to do something about the root of the problems, the easy availability of guns.


This blog entry was posted before we learned that Minal Panchal was missing. After we heard, most of us have been on campus or at Montgomery Regional Hospital trying to locate her.

Read More

The Letter and Intent of Creative Commons Licensing

I subscribe to Seth Godin's blog, and I find his opinions very thought-provoking, I might add. I especially like his rants on usability and good versus bad experience design.

This morning, I read a post by him about one of "his" books being sold on Amazon. To explain why the "his" is in quotes, here's what happened: Seth wrote the book in 2005 and licensed it under a specific Creative Commons (CC) license1. The book was and is still available for free from Seth's website as an unlocked PDF. A book publisher, who had nothing to do with Seth directly, went ahead and printed the book which is now available for $9.99 at Amazon.com2. Seth is now pissed off at someone doing something like this, and is encouraging the readers of his blog not to buy that book.

I think that the publisher's action is not only within the letter of the law, but also within the intent of the Creative Commons license Seth used. There are more than one CC licenses, and the specific one that Seth chose allowed free copying of the book, as long as authorship was properly attributed. Although there exists a Creative Commons license that disallows commercial usage, Seth chose not to apply that clause (which, by the way, he now considers was a mistake back in 2005.) This leads me to conclude that the publisher was offered those rights by Seth himself.

I can understand Seth's getting pissed off because someone else was making money off his effort, but at $9.99, I think it nicely covers printing costs and perhaps makes a little profit for the printer3. If I already had the PDF eBook and still wanted a paper copy, I'd be super-willing to pay $9.99 for simply the printing, binding, cover, etc. I see nothing wrong with the printing of the book.

Although I admire Seth's decision to license his work under a CC license, I feel he is going against the intent of the license by exhorting his readers not to buy a work that was permitted expressly because of that license. If he really wanted to follow the spirit of the Creative Commons, he should have provided a link to the book on Amazon and encouraged his readers to buy a paper copy in addition to the free eBook they might already have downloaded. His current actions undermine the spirit of openness that the original grant of the license had fostered.

1. I also use a Creative Commons license for this website and for all my non-academic writing.
2. I do not know the publisher and I do not earn any money as commission or from Amazon referrals. Just to make it clear, you know.
3. Maybe more. I don't know much about the printing industry.

Read More

From the Desktop to the Phone ... Seamlessly

Google just announced a new feature in Google Maps: Click to Call. When you find a business on Google Maps, you can ask to be connected directly. Google then calls you on the number you provide, and places a call to the business at the other end.

This is yet another example of seamless task migration. The user's ultimate goal in locating a business is to get in touch with them. The most common way to do this today is to call using a phone (at least as long as Voice-over-IP is not as ubiquitous as cellphones and land-lines). Lo, Google bridged the gap. End-to-end support for a user's tasks using multiple devices is a challenge that's getting its due attention only recently.

Hopefully, we will soon be able to do the same with phone numbers all over the Web. Imagine a button on my website that says, "click to call me". Or, a button on my photo albums page that says, "view as a slideshow on the living room TV". Or being able to press a button on your car radio to "read more about the currently-advertised product once I'm back home".

Read More