12.15.06
Yay multipoint routing!
Google Maps now has multipoint routing! For example, see this map from LA to Chicago to Minneapolis to Miami to Houston.
I didn’t work on multipoint routing, but my cubemate did. 🙂
Permalink Comments off
11.16.06
Google click-to-call
Woot!
Click-to-call is working on Google Maps!
While I didn’t do any work on the click-to-call, I sat across from someone who was working on it, so I’m getting a bit of a vicarious thrill. 🙂
Permalink Comments off
11.13.06
WWW is sixteen years old today
Apparently, the oldest web page that the W3C knows about is sixteen years old today.
Most of the credit is ascribed to Tim Berners-Lee and Mark Andreessen. I think that the deserve all the credit that they are given — their contributions were very important. However, there are some people and institutions who tend to get left out who were also pretty important.
Robert Cailleau worked with Tim Berners-Lee on the initial WWW, and Eric Bina worked with Mark Andreessen on Mosaic. You haven’t heard of Eric Bina because he’s quite shy, but I bet he did a large share of the work on Mosaic. Also, he didn’t leave Champaign-Urbana for the bright lights (and fame) of California because his wife is a professor at University of Illinois at Urbana-Champaign.
I think that the importance of CERN and the National Center for Supercomputing Applications has been understated.
CERN gave Tim Berners-Lee both the funding to work on the project and a public testbed for it. If the Web had been developed at someplace like IBM, it probably would have just been an intellectual curiosity. The Web infrastructure was not all that technically difficult: the hard part was getting people to generate content.
NCSA not only employed Marc Andreessen and Eric Bina, but they provided a whole host of support services to the world. While the stories I heard implied that Marc and Eric worked on Mosaic with out permission (perhaps even against orders), I fully believe that without NCSA’s (eventual) institutional support, Mosaic would not have been a success. Mosaic wasn’t the first graphical Web browser — ViolaWWW, developed by Pei-Yuan Wei — was. But NCSA gave technical support for Mosaic, provided documentation, hosted the download traffic, and funded the development of a number of infrastructure improvements (like the httpd Web server and its CGI extension), and so it succeeded.
Looking back a little farther, there were a number of factors that the Web depended on. The Web couldn’t have gone anywhere if it weren’t for widespread use of the Internet, and that really depended upon cheap computers and email. If IBM had not made the hardware design essentially open source, computers wouldn’t have been so cheap and therefore common.
The Web also depended upon networking. It was only because computers were already everywhere and already networked that the Web could take off. The incremental cost of running a Web server on a machine that was already there and networked was trivial, so why not toss one on? 3Com and Cisco can take some of the credit for that, though there were other networking companies as well.
Even Apple played a role. While I found their HyperCard technology slow, clunky, and uninteresting when I saw it in the late 80s, it popularized the idea of hypertext.
So while I think we should give respect to Tim Berners-Lee and Marc Andreessen, I think they also had the extraordinary good fortune to arrive on the scene at a time when cheap, networked computers were widespread and to work for public institutions that provided them with a great deal of support.
Permalink Comments off
mini-review: Regan Mandryk's stuff
Regan Mandryk did a way cool PhD where she figured out how to identify various emotions/mental states — fun, excitement, boredom, challenge, and frustration — from physiological measurements.
She stuck sensors on guys playing video games to measure things like heart rate, galvanic skin response, etc. She asked them to report on their feelings during the game play, and with a small application of fuzzy logic, was able to tease out the correlations between the sensor data and emotions. Cool stuff. (See Regan’s thesis or a terser summary of her work.)
It might be interesting to stick electrodes on people coding; there is all kinds of interesting stuff you could pull out of that. Combine the sensor traces with traces of interactions with an IDE, and see where the IDE is frustrating the coders. See if you can identify “flow”. See what happens when you interrupt flow. See how long it takes to get back into flow from various types of interruption. Figure out how to get people into flow.
Permalink Comments off
10.28.06
What working at Google is like
Steve Yegge has a long blog post that is very funny and probably helpful; in the middle of it, he talks about what it’s like to code at Google. It is absolutely dead-bang consistent with what I saw.
It’s been a little frustrating to me to try to express just what it is about working at Google that is so cool. It’s easiest to talk about the food, the perks, as that’s the most obvious, smack-you-in-the-face-can’t-miss-it. I tried telling people about how the whole culture is focused on making life easy for developers. Somehow, it didn’t seem to really capture it in its fullest.
I would make only a few additions and qualifications.
We did have schedules and target dates, they were just (for the most part) very casual. If we made them, fine. If we slipped, darn, let’s move on and get it out next week. There were, however, exceptions. There were a whole boatload of things that were announced at the Google Geo Developer Day a week after I started. While I wasn’t really involved, I believe that there was a lot of scurrying around to pull that off.
In general, things happen incrementally when they are ready. I don’t remember there being a hard drop-dead date for the continuous zoom, for example.
There are three reasons that Google can be more casual about schedules.
First, and IMHO most importantly, It’s all Web-based, so they are not a slave to big monumental releases. It’s not a shrink-wrapped product where you get one shot to get your feature in, and it better be what the customer wanted, and it better work flawlessly. If Google puts a feature in, and it breaks stuff or people hate it, they can pull it out quickly. (This summer there was at least one feature that got pulled because end users and/or external API users didn’t like it.)
Second, they have a whole pile of money. You can be more casual when you aren’t worrying about your company folding if something doesn’t ship on time.
Third, they work very very hard at not hiring the wrong people. They have a stated policy that they would rather not hire someone good than hire someone bad. I fully agree with and endorse that policy, even though I might have been adversely affected by the policy. (They recently declined to offer me a full-time job.)
10.15.06
Long slow farewell to Eudora
Yesterday, Qualcomm announced that they were going to cease selling Eudora. I have really mixed feelings.
On the one hand, it’s great that Thunderbird is going to get the Eudora team’s help to add features and overhaul stuff. I really like Eudora, so it will be nice to see some of Eudora’s features coming Thunderbird’s way.
On the other hand, I am very sad to see the passing of yet another email client. There were precious few consumer-grade email clients as it was — Outlook, Lotus, GMail, Yahoo Mail, Eudora, Apple Mail, and Thunderbird* — and now there is one fewer. I think that the existing email user interfaces have a lot of room for improvement, and worry that the field will lose some vitality with the loss of Eudora.
I hope I’m wrong. I hope the movement of Eudora users to Thunderbird will help drive Thunderbird to even wider acceptance.
*I have heard people raving about The Bat, but I’m not a Windows user, so don’t have an opinion on it. Mulberry had good intentions, but I think it fell short in some significant ways. (Other people did as well — its parent company went bankrupt and Mulberry’s future is somewhat sketchy.) OSAF might someday release Chandler and it might someday be cool, but it’s still not ready.
Permalink Comments off
08.27.06
Microsoft should fear Google, part 4
I blogged before that I thought Google should bundle mail and calendar together and offer it as a stand-alone product. As I noted before, Google had started to offer mail as a hosted application; now they are offering mail, calendar, IM, and Web page design/hosting as a hosted application.
Apparently it’s not necessary for it to be a hosted application for it to be an interesting value proposition, but I think even more businesses would be interested if they could get the machines behind their firewalls.
Permalink Comments off
06.29.06
Life as a Googler
I have been a summer intern at Google for a month now, so I feel somewhat obliged to talk about what that’s like.
Secrecy
Much of what I could talk about I, well, can’t talk about. Google is extremely paranoid about loose lips. If you go searching around for what employees say about Google, you’ll hear a lot about the food. While the food is really good and a great benefit, part of why you hear about that so much is that it is a completely safe topic of conversation.
Google is almost as close-mouthed at Interval Research Corporation, partly for similar reasons. In both cases, there was a sense that protecting intellectual property was so important that it trumped almost everything.
I am certain that secrecy hurt Interval, and suspect that it’s not as important as Google thinks it is. At Interval, it was extremely hard to recruit people, and that was partly due to absolutely nobody knowing what we did. At Google, while everybody seems to think that the ideas are important, the longer I’m there, the more I think that its bigger competency is in execution. Sure, Google people have good ideas… but ideas are no good if you can’t put them into practice.
Most research-y type companies make their money on low-volume but very expensive products. They fund their research by putting a big markup on their products. Google, on the other hand, is the epitome of high-volume, low cost products. They fund their research by pumping out a huge volume of product.
The other part of why Google is so secretive is legitimate: it’s that everybody is watching very very closely. Anything a Googler says might end up on the cover of the New York Times tomorrow. This definitely constrains me. There is a Google Maps mailing list, for example, that I pay some attention to. There are questions that I know how to answer, and want to answer, but I’m nervous that the people will think I speak for Google. If I say “There are currently seven froomblents”, will people interpret that to be a veiled admission that Google intends to change the number of froomblents they provide? (No, don’t go look it up. I searched for a word that had no Google search results, specifically so people wouldn’t start imputing meaning to that nonsense phrase.)
Personality
Google tries really hard to hire only the best and the brightest. I knew that before I started, and it made me a bit nervous. I once worked for a different company that worked hard at hiring only the best and the brightest, and they had horrible conflicts. Everybody there was used to being a superstar; everyone was used to always being more right than the people around them, and getting their own way. This meant they fought all the time.
I’ve been very pleasantly surprised at Google at how nice everybody is, and how there really aren’t big fights about technical direction, focus, implementation, or turf. Maybe I just happen to be in a particularly sweet group, but I don’t think so. I think Googlers are genuinely humble and nice.
That might be the Stanford influence: everyone at Stanford knows someone who is better than them at something. One of my Stanford classmates once said that he had three International Math Olympiad finalists in his calculus section. My husband tells of how in his Music Theory class, there was someone who was a varsity football player and a dancer and played violin. (“And was good looking. It wasn’t fair.”)
I suspect that more of it is that they worry about people being smart and pleasant to work with.
There’s more, but I can’t think of anything more right now that I am allowed to talk about.
Permalink Comments off
05.27.06
company perks
Google is famous for its perks, and rightfully so. They have the best perks of any place I’ve ever worked, even better than SGI in its heyday. It seems like Google really wants to eliminate anything that might distract people from doing great things.
It’s also possible that they just want to be nice to their employees, and I found an argument for that: heated toilet seats.
While I haven’t done a rigorous statistical analysis to determine distribution, almost all of the toilets at the Googleplex that I have sampled are high-tech megafunction toilets: the kind that can spray your privates clean and then blow them dry.
The first time I walked into a stall, I rolled my eyes at how over-the-top the toilet was. I mean, how necessary is it to have megafunction toilets?
Then I sat down and discovered the seat was heated, and to my surprise, I found that I had a very visceral response. It was comforting. I suspect that the seat is set to body temperature, and I bet that I have very strong associations of comfort attached to body heat on my butt. I wanted to just sit on that nice warm seat for hours. This argues against the perks being there to improve productivity.
I haven’t tried the toilets’ wash and dry cycle yet — I’m afraid to. After all, I do need to get some work done.
Permalink Comments off
05.22.06
to-do manager, addendum
I forgot to mention last night — of course it would be great to allow people to share to-do items, in much the same way that people can share calendars with Gmail. For example, it would be great to be able to put “buy new laptop for Ducky” on my husband’s to-do list.
Permalink Comments off
« Previous Page — « Previous entries « Previous Page · Next Page » Next entries » — Next Page »
