Hypertext Weekly #2

It’s that time again : when, running out of ideas and/or time, Heavy Mental multi-twitters.

On the (pretty geeky) menu : high scalability, graduating Vs startuping, Web 2.0 Noise, OS/360 Tar Pit and a rock solid business model there for the taking : God Bless HTTP !

Kill the HIPPO

Both SignalVsNoise and Fred recommend the link, that’s just enough to make it a mandatory read : a long article by high scalability on Amazon Architecture and lessons learnt while building the system. Some favorites :

• Start with a press release of what features the user will see and work backwards to check that you are building something valuable.
  
• Cut technology to the minimum you need to solve the problem you have. It doesn’t help the company to create artificial and unneeded layers of complexity.
• Use measurement and objective debate to separate the good from the bad. Avinash
  Kaushik calls this getting rid of the influence of the HiPPO’s, the highest paid people in the room
• Look for three things in interviews: enthusiasm, creativity, competence. The single biggest predictor of success at Amazon.com was enthusiasm.
• Amazon doesn’t like middleware because it tends to be framework and not a tool. If you use a middleware package you get lock-in around the software patterns they have chosen. (…). You’re stuck.

College of reason

www star Paul Graham essay of the month is a controversial one on good colleges. For those who dont know him, he’s a software programmer and a Venture Capitalist investing in Startups.

Here’s the pitch : Startup is a fantastic lab to see how good professional people are since there only are two outcomes : success or failure. Paul didn’t notice any particular success rate for young entrepreneurs coming out of good colleges, while comparing with other graduates.

• Success is decided by the market: you only succeed if users like what you’ve built. And users don’t care where you went to college.
• Y Combinator has been an unprecedented opportunity for learning how to pick winners. One of the most surprising things we’ve learned is how little it matters where people went to college.
• Y Combinator showed us we were still overestimating people who’d been to elite colleges. We’d interview people from MIT or Harvard or Stanford and sometimes find ourselves thinking: they must be smarter than they seem. It took us a few iterations to learn to trust our senses.
• There used to be a saying in the corporate world: “No one ever got fired for buying IBM.” (…) A recruiter at a big company is in much the same position as someone buying technology for one.
• We can learn more about someone in the first minute of talking to them than by knowing where they went to school.
• There is only one real advantage to being a member of most exclusive clubs: you know you wouldn’t be missing much if you weren’t

Usual argument against this type of statement : the guy has succeeded but he still is bitter because he didn’t graduate from a good college. Nice try but unlucky : he is a PHD from Harvard, as he reminds it :

• There’s nothing like going to grad school at Harvard to cure you of any illusions you might have about the average Harvard undergrad

Even such A-lister as Adriana Lukas is quite uncomfortable with this when referring to Paul’s essay :

It is true that for the purposes of business, a university degree is a signal. A signal that (…) you can absorb, regurgitate and communicate large amounts of data and knowledge by dead white men (…) Not much about innovation, irreverence, disruption and not caring about consensus. Unless you, of course, went to Oxford.

Man ! Poor Paul would choke with disbelief in France : if only 8% of CEO of American companies are graduate from so called good colleges, the figure is ten times higher in France !

Web Of Noise

Excellent post by Kari Silvennoinen, a fellow blogger of techiteasy : The Noise of Web 2.0 .

Google Reader told me that “From your 30 subscriptions, over the last 30 days you read 1,023 items, starred 3 items, shared 10 items, and emailed 1 items.” Over thousand items? (…) That number was staggering. It can’t be all signal and I tend to skim most of the headlines and click “Mark all as read” on some of those high traffic news feeds, but that makes it worse actually – it means it’s just a lot of noise. Twitter? Jaiku? All noise. Facebook. IM. SMS. Del.icio.us. Noise. Noise. Noise. N.oi.se.

I guess this about time we realise that when we get news/twitter/info pushed to us, this is a source of distraction and stress : we just feel guilty not going through all of those.

Tar Baby

Joseph Ottinger has started reading the Mythical Man Month by Frederic P. Brooks, a 30 years old best seller. He talks about it on the Server Side. This is the book about the massive project of writing a brand new IBM Mainframe Operating System (OS/360) in the early 60s.

Joseph focusses in this review on Chapter 1 (The Tar Pit) and the perpective it brings to our today’s problem while building large scale enterprise applications.

If you’re unfamiliar with what a tar pit is, consider the La Brea Tar Pits, considered one of the world’s best repositories for fossilized vertebrates. Animals would wander on top of a tar pit (which was hidden through dust, leaves, or other such detritus), and become unable to escape. Predators would be attracted by the struggles of a trapped animal, and become ensnared themselves.

That’s very much like common architectures today, where most of the application structure is trivial, and the last 20% of the application takes forever to work through (or escape.) In many ways, the “tar pit” describes much about our field, and symbolizes where the endless drives for componentization, software-as-services, enterprise service buses, service this, service that, service the-other.

Groundhog Web

If you’re feeling savvy having read the tar pit, knowing how to get rid of Web 2.0 Noise and you wish you had a great business model to offer Paul Graham, do not worry : Joel brings it to you on a silver plate.

Funny yet long article on a spotted software business opportunity. Joel Polsky goes first into software history going back to OS/360 (again !), PCs evolution and Moore’s law, with that hilarious and casual style :

As a programmer, thanks to plummeting memory prices, and CPU speeds doubling every year, you had a choice. You could spend six months rewriting your inner loops in Assembler, or take six months off to play drums in a rock and roll band, and in either case, your program would run faster. Assembler programmers don’t have groupies.

And then his theory : “History is repeating itself, in three different ways, and the smart strategy is to bet on the same results”. Web is only at his early age and “HTML is CICS with fonts” (I guess he meant CICS forms since CICS is a whole transaction system).

Anyway : the point is

We don’t care about performance or optimization much anymore. Except in one place: JavaScript running on browsers in AJAX applications. And since that’s the direction almost all software development is moving, that’s a big deal. (…) This time, it’s not the RAM or CPU cycles that are scarce: it’s the download bandwidth and the compile time. Either way, you really have to squeeze to get complex AJAX apps to perform well.

However, the main issue with AJAX is not the the fact that is not optimized. Because the Moore’s law will somehow apply to bandwidth

History, though, is repeating itself. Bandwidth is getting cheaper. People are figuring out how to precompile JavaScript. The developers who put a lot of effort into optimizing things and making them tight and fast will wake up to discover that effort was, more or less, wasted, or, at the very least, you could say that it “conferred no long term competitive advantage,” if you’re the kind of person who talks like an economist.

No, no, no : here is the winning approach :

The winners are going to do what worked at Bell Labs in 1978: build a programming language, like C, that’s portable and efficient. It should compile down to “native” code (native code being JavaScript and DOMs) with different backends for different target platforms, where the compiler writers obsess about performance so you don’t have to.