Vamshidhar Kommineni

August 17, 2007

Peter Spiro on “Behind The Code”

Filed under: Technology — Vamshi @ 8:32 pm

Microsoft recently started doing this interview series called Behind the Code where they interview key architects at Microsoft in a more free-wheeling fashion about their lives, how they got to working on computers, what drives them, etc. Kind of like Charlie Rose interviewing techie people. They’re up to about 8 interviews now. Their official blurb is:

Occasionally the office of the CTO, David Vaskevitch, sits down with some of Microsoft’s most influential technical employees to capture their stories. Instead of examining specific technologies, BTC takes a closer look at the person, the career and what it takes to produce world-class software.

Anyway, I recently watched a talk online on Peter Spiro (one of the architects behind WinFS and considered to be a very colorful personality):

http://channel9.msdn.com/ShowPost.aspx?PostID=333294 OR
http://www.researchchannel.org/prog/displayevent.aspx?rID=19278&fID=345

Peter is a Technical Fellow at Microsoft, and it was pretty interesting to listen to him. Particularly about his early years with a degree in forestry and "not getting a real job until 30". It’s also interesting to see him talking about the early days of SQL development at Microsoft. Things like having a small team with very little process and the importance of building teams of people in order to execute on software projects. The talk is well worth watching for his insights into people and team building alone. Other gems include his point about long lived software needing to be a set of subsystems and providing enough infrastructure and contracts between subsystems to allow swapping them out over releases.

In the Q & A afterwards, he made some very relevant points as well:

  • It’s all about the people. People skills are very important in technical organizations and will continue to serve people well
  • Being someone who fights against "the system" and tries to induce change is important
  • Finding the right technical leaders who then influence all the other people that they manage is one of the hardest things about building a software company
  • We don’t train new graduates well enough in the software business

The only weird thing about the video is the jarring edits to keep it under time. There are some abrupt transition points that could have been handled better.

Funnily enough, he reminds me of a Professor from UIUC, Steve Lumetta, both in the way he talks and the way he looks (People who know Steve will know what I mean :). Steve would like Peter too, they probably would get along famously on their shared admiration for DEC. The amount of joy in their work and the technical acumen they bring to the table inspires me to continue learning and become better at what I do.

Technorati tags: , ,

April 25, 2007

“RIM: Not Enterprise/Government Ready”

Filed under: Technology — Vamshi @ 12:56 am

Last week was the second time RIM put its customers at risk. This time, however, these customers found themselves disconnected without explanation for an extended period of time. The lack of notification, the lack of accountability, and the lack of any real evidence the core problems have been addressed make RIM unacceptable as a large enterprise, […]

Source: RIM: Not Enterprise/Government Ready

Another good post about the RIM/Blackberry outage that explicitly addresses the issue of availability and the trust that users, corporations and governments have put into RIM. I talked about this in an earlier post as well

Technorati tags: , ,

Post to Windows Live Writer from Sharp Reader

Filed under: Technology — Vamshi @ 12:50 am

Windows Live Writer (WLW) is definitely my favorite blogging client and I’ve moved back to Sharp Reader from Outlook (stopping at Google Reader and RSS Bandit on the way, but more on that in a different post) as my main RSS reader. I was looking around for a way to post directly from Sharp Reader to WLW when I found this page about Sharp Reader plugins:

http://www.sharpreader.net/plugins.html

Which led me to the IBlogExtension standard:

http://www.pocketsoap.com/weblog/stories/2003/04/0023.html

which led me here and finally to here: 

http://www.codeplex.com/Wiki/View.aspx?ProjectName=rssbandit2wlw

Though developed for RSS Bandit, the plugin works like a charm with Sharp Reader saving me a bunch of time when I want to post about something I’ve read. A few nifty things here:

  • The IBlogExtension standard that different RSS readers including Sharp Reader chose to respect
  • Steve Kaschimer for implementing the plugin and making the code available so I muck about with a private copy if I wanted to
  • Actually being able to find this stuff because of persisted blog posts (one of the main uses of blog posts in my opinion is the creation and persistence of user knowledge)
Technorati tags: , ,

April 18, 2007

Scalability of web services

Filed under: Technology — Vamshi @ 6:55 am

In a very timely fashion, here are two Slashdot posts about the fragility of our networked existence:

Blackberry network is down

Turbo Tax melts down on Tax day

I’ve never used Blackberry but have always wondered about the scalability given that corporate email gets routed through their central servers to cell phone providers. I think there is still communication with a centralized environment in the case of their BES servers. A costly lesson learned about keeping your environment decentralized and scalable. I’m not an expert, but I think getting mail on your Windows Mobile phone from Exchange server doesn’t suffer from the same centralization problem (I could be wrong, I’m just an end user and have no idea about the architecture).

The other one is going to cost Turbo Tax a lot of customers and result in a bunch of negative PR (and probably some fun postmortem meetings behind the scenes). Don’t get me wrong; I love Turbo Tax. I’ve used it to do my taxes since 1999 and can’t imagine the dark days when people were forced to navigate tax forms before they came along and simplified it. Sure, you can blame the customers who waited till the last evening to submit their taxes, but customers will be customers (and they’re always right). Turbo Tax is a very state heavy web application (lots of data going back and forth from a central data store), so depending on how it is architected, there are going to be some painful choke points. That’s one system I’d love to learn more about.

Failures like these are why good design is critical in the "Web 2.0" world. Designing a web service for fault tolerance and scalability are hard problems. It’s also why services like Amazon’s S3 and EC2 make for a very compelling scenario. Imagine infrastructure, fault tolerance and scalability on the cheap with all the heavy lifting done by teams of people that understand how to design these systems without everyone putting up a web property having to provision and deal with the problems themselves.

Technorati tags: , , ,

March 26, 2007

Scoping Software to make sure it ships

Filed under: Technology — Vamshi @ 10:25 am

Raymond Chen (one of a small set of Microsoft bloggers I follow regularly and someone who I have a great deal of respect for as a dev) wrote this article a few weeks ago:

You don’t know what you do until you know what you don’t do

I do believe this is an essential part of software development, right from an individual dev’s commitments all the way to large projects (like Windows).

As a dev, the more experienced (and hopefully wiser) you get, the more you realize that achieving success and satisfaction is as much about realizing what you can do in the given timeframe, and firmly saying "No" to every other feature request (and trust me, everyone always comes up with feature requests, particularly at feature review meetings after you’ve finished the feature).

For a startup or a small dev team, success on a new product/feature depends a lot on restricting feature creep and resisting the temptation to solve world hunger ("We’re not just building a simplified code editor, we’re going to build a platform to replace Word…"). I think OneNote and Windows Live Writer are two relatively recent MS products that I’ve used that have consciously evaded this trap and delivered elegant, focused and usable pieces of software.

For a large team, lacking a central vision and concrete customer scenarios can lead to lengthy delays and products that don’t necessarily have a coherent story at release time. This isn’t to say that there aren’t benefits of bottom up feature teams, but they need to be tied together with a central feature plan. And you can’t do that unless you reject some features (postpone to future releases, cut, etc.) with a "No".

I’ve seen the first and third above in operation personally. While I haven’t experienced the case of a startup team yet, I suspect it applies just as much there.

Technorati tags: ,

December 20, 2006

IE Inline Search

Filed under: Technology — Vamshi @ 12:22 am

Via Omar’s blog:

Very cool and simple plugin that improves greatly on the Find in page (Ctrl+F) functionality built into IE. Definitely one of the must have IE7 plugins. Straightforward find next and find previous without a modal dialog and the ability to highlight all occurrences in the page.

Inline Search Plugin Page

It’s also on IE’s recommended plugins list

Technorati tags: , ,

November 22, 2006

SOAP and WSDL explained

Filed under: Technology — Vamshi @ 3:45 pm

On Pete Lacey’s blog via Dare Obasanjo:

The S stands for Simple

This is a great article that lays out the weirdness behind SOAP, WSDL and other "simple" web service APIs in a very funny dialogue between a dev and a "SOAP guy". My head hurts every time I look at the weirdness of SOAP. Thankfully, I don’t have to do this for my day job, just when I’m poking around at various web APIs (Amazon, Smugmug, etc.) for personal coding projects.

Websites and outage for maintenance

Filed under: Technology — Vamshi @ 12:57 am

Went to visit Netflix to add a movie to my queue before it slipped out of my mind again and encountered this at 12:50 am:

The Netflix web site is temporarily unavailable due to scheduled maintenance.

It is anticipated that the site will be available again at 1:30 AM Pacific time.

We apologize for any inconvenience this causes you.

Please visit us again soon.

It’s amazing that large websites like Netflix need to and can take their sites down for maintenance. Netflix can probably afford to do this because they are a US only site (I think, but their site is down, so I can’t verify :) and 12:50 am is pretty much dead of the night across the US (except Hawaii, but no one ever takes poor Hawaii into account). To me, it indicates how far we still have to go in providing real high availability systems to everyone.

Technorati tags: ,

November 21, 2006

Yahoo hates Peanut Butter?

Filed under: Technology — Vamshi @ 12:17 am

Brad Garlinghouse, a VP over at Yahoo, published an internal document that’s made it onto the Internet about the ills of Yahoo and what can be done to correct them. The peanut butter reference is about being spread too thin across too many productsand lacking focus in them.

The memo is a great read and well laid out. I think every tech company, once it finishes going through it’s growing phase faces this issue. Having a coherent and focused product strategy with full employee engagement is the key to success at any company including Microsoft.

However, I’m not so sure about the solutions he proposes. Headcount reductions can be very, very bad for morale of the rest of the people, and you’re never really sure whether you’re getting rid of the right people (not mention the collateral damage of your best people leaving rather than being subjected to later layoffs)

Full text of the memo (WSJ)

TechMeme on the memo

Technorati tags: , ,

October 9, 2006

Terry Semel on Yahoo's role

Filed under: Technology — Vamshi @ 12:15 am

Saw an interesting interview with Terry Semel on Charlie Rose. Aired on 9/21 (I don’t watch much TV, so even though I save only the most interesting Charlie Rose interviews, my watching tends to be way behind schedule).

One of the things that struck me about this interview was how much he talked about monetization of search vs quality of the algorithms behind search. He believes that Yahoo is on par with Google in terms of quality of search but Google does a much better of monetizing each search query. I believe this is mostly true. For probably 90% of all searches, I suspect live.com (Microsoft’s re-branded search portal; it’s kind of sad that I even have to explain that) or yahoo.com will give similar searches to google. I do believe that Google’s algorithms still do a better job on the long tail search queries. This is purely from anecdotal observation of course (I switched over to using search.msn.com and then live.com about a year ago, and only revert to Google to compare search results, or when I think results from live.com are particularly poor).

Where Google does take it to the competition is in how good their Ad Words program is. I’ve read nothing but good things about how flexible and usable it is for advertisers. From a consumer’s perspective, I’ve been impressed by how easy they’ve made it for third party sites to earn revenue based on including ads from their platforms. I would argue that the real platform they’ve managed to build and innovate on is the ad platform and not the search platform (of course I wouldn’t deny that they innovated on search as well, but the competitors are catching up in this space). Hopefully, MSN’s Ad Center and Yahoo’s equivalent platform will even the playing field a bit more.

Terry also talked about how Search represents only 5% of the time people spend online (he says people spend 40% on information and entertainment, 40% on communication like IM and email and the rest 15% on commerce), and it’s in the other fields that Yahoo has an opportunity to do better than it’s major competitors. He also talks about the importance of global reach and being a media company as well a technology company, the latter of which is expected given his background.

I think Yahoo is a fascinating second runner (maybe not in reality but certainly in perception) to Google’s limelight. They are no longer as relevant in many ways as they used to be 5 or 10 years ago. I remember the days when Yahoo was pretty much the default portal for everyone for almost everything, and that clearly is not the case anymore, particularly among the younger demographic. I personally haven’t been to Yahoo’s front page in years and only use them for their movie places/times portal (not because it’s superior, but due to force of habit). At the same time, they have a very credible set of services in email, IM (both of which have way bigger customer bases than Google’s offerings) and other things. They are also probably the best one stop portal for the casual consumer bar none (sorry, MSN). And they are a company that gets media and technology at the same time. It will be interesting to see how they fare over the next 5-10 years.

tags: ,

October 7, 2006

Crossing the Chasm: Technology whose time has come

Filed under: Technology — Vamshi @ 7:33 pm

The title of the post comes from this book (see more on Wikipedia) which in turn is based on the earlier Diffusion of Innovation theory.

The book argues that there isn’t a smooth adoption curve for technology products as the diffusion of innovation theory suggests. In particular, the claim is that there is a gap (chasm, I guess) between early adopters and the early majority (see above Wikipedia articles for an explanation of the terms).

I believe in the theory. I think it has been illustrated in the past with products from cell phones to DVD players to LCDs and HDTVs. I still remember the weird looks I would get when I owned a cell phone in college in early 1999 (the friends/acquaintances who did comment to me said something along the lines of "Do you really think you’re so busy that you need the cell phone and need to pay $30/mo for the convenience of being reached anywhere?", and that was the polite ones). Less than 2 years later, every college student in Engineering had a cellphone and today I don’t think college freshmen can imagine a world before cellphones.

I think it is interesting watch the cross-over from diffusion among early adopters (techies like myself) to the population in general. I’m going to comment on a few products/devices to illustrate what I think is this shift in it’s early phases (I may be completely off base of course). It was watching the various articles/posts on these products that led me in search of a unifying theme to tie them together.

tags: ,

Create a free website or blog at WordPress.com.