speed saves

So thinking about web projects, I'm curious what people actually do with their hardware. From a development point of view, most developers kinda throw it over the wall to the operations folks to make it go fast. Most of the architects I know give some basic consideration to making a solution scalable, but few rarely look at it as carving more cycles out of the existing hardware vs throwing more money and hardware at the situation. Is this endemic of all large companies? By comparison reading about different Web companies, who have grown out of very, very cheap servers with some great strategies for increasing capacity, decreasing latency and keeping costs down. Google's tech talk on Dapper, their RPC monitoring unit shined a light on how their site has grown over time. From a set of borrowed machines in 1997 to warehouses of servers today.

At work our standard is a sub-6 second response time for most apps. Google would not discuss what their threshold was for response time, but the impression is that data needed to leave the datacenter in a sub 1 second response time, and all of their development was geared towards that goal.

I think I need to spend a bit more time figuring out guidelines for better performance. I should restate that. I need to study the existing performance guidelines and learn to apply them better to my work.

deadlines deadlines deadlines!

I have a lot to do in a little time. Project A from work is eating a ton of time at this point. Way more hours that I was hoping to work. I have a WSTC Lab 1st draft due by the 21st, my Firefox plugin has to be updated by the 18th, and I have a WSTC presentation due in early April.

Of course it doesn't help that my lack of time is making me look at other projects as well. I want to try and build that market driven development engine using Project Zero. I also want to wrangle a bit with Google Gears. After hearing about the App Server that runs javascript on the server and client, I wonder if that's a better solution for Gears like functionality. Of course gears has the requirement to work on all browsers. For my curiosity I could build a Firefox plugin only version to run my own apps both from a server or in a connected -less mode. Perhaps Python might be an other accessible choice... there was code samples available for compling python into a plugin. ..or perhaps not, this note from Brandon Eich suggests this might not be a great idea.

Goodbye GG

Gary Gygax passed away. The father of Dungeons and Dragons has moved on to other planes. I haven't played a game in forever, but I still love it had have a good number of happy memories connected to the game and the other games that have spawned off it. The game has moved on in different directions, and allowed a lot of geeks a chance to spread our minds to other worlds as well.

Thanks Gary.

PghGooglePlex

Spent the evening hanging out at the Pittsburgh Google Plex for a talk on RPC tracing with Dapper.  Dapper is the google tool for measuring how much time RPC calls are spending at different layers of the Google stack.  Interesting from the pov of what does it take to make google go and some of the assumptions of what they can expect from their infrastructure.  The talk wrapped up with a tour of the facilities.  Very cool overall.

Also interesting to see the tons of smart folk out to hear a bit more of what Google is up to and the side conversations they were having.

Hats

Ah in a misspent youth I had a flair for hats. I had a fedora, a jester's cap a couple of berets, a Cat in the Hat Tophat and a few caps as well. Never really went for the baseball hat, but I found the ones I like best. These days I wear the occasional winter cap, for the cold, but that's about it.

On a whim I hit hats.com, and I must admit, the old jonesin for a hat came back. the boris. the porkpie, an agean greek fisherman's hat. So many wonderful hats.

I'd look ridiculous in most, but man, I do admire a really nice hat. A clean style. A well made hat is a wonderful thing. Styles a thing I appreciate even if I'm not one to invoke, and while I won't be mistaken as stylish or in vogue, I will say, I do appreciate a true timeless style. There is something very classic about hats. What would Indy be without the fedora? Or Steed without his bowler? An elegant artifact of a more civilized era... Perhaps their gone, perhaps they'll make a comeback. I'd never say for sure.

All I know is, I think I need a few new hats.

... Now with Gaming!

So originally I looked at Brainstorm from Ubuntu, and digging a little deeper I can see a similar idea from Dell in their IdeaStorm site. These sites are good for taking the pulse of an audience, finding what users like and dislike, and what they'd like to see more of, but what about for owners? When we have a team of people who are shareholders in a project. Would a market based approach work out better? Give each a stack of 'dev points' and let them spend by priority? What is the payoff? How do you measure the 'win'.

Predictive Markets seem like a good source of material to explore in this vein.

I keep looking for new projects to undertake in new languages. Learning the nuances of new languages gives me a new headspace to think about things. The patterns that a language provides gives some new insight into old problems. Perl, Java, PHP all have very different intents, strengths and weaknesses. Each makes a subset of problems very easy to deal with. Perl is whiz at system administration, Java for making standardized modular code and PHP for whipping out web apps. It can be an interesting challenge to push those languages to other domains to see how they swim or sink as the occasion arises.

On the other hand, there is something to be said for trying out new frameworks for old languages. There are subtle differences in execution that can make all the difference. For instance plumbing together WebSphere Applications on a J2EE stack has a completely different model than running some PHP apps or a Ruby On Rails application. But something like Project Zero gives a fusion blend of Java with PHP in Rails like framework. Still very course and working it's way to maturation it does have a few possibilities to play with. I'm looking to build a few apps in the next few days. This may be a good time to brush off a few project notes and find a good language or framework to measure against these apps.