Deviant Abstraction

We all know this chart is already obsolete and somewhat irrelevant. However, reflecting upon it is always useful.

What it really shows is the mainstream view of these technology. You can see VoIP or Web 2.0 but not P2P which is an emerging system trends (in the sense that a lot of its potential is still untapped). Some points are senseless (Grid computing: what does it mean? MPI is already there and battle-tested) and mix commercial application (VoIP), applied research (sensor networks) and fundamental research (quantum computing). In a way, this is good: it lays down some frontiers the academic community has artificially created.

Via JdN

Amazon EC2(Elastic Compute Cloud) is the next logical step after S3. It is built on top of S3 and allows you to instantiate a virtual machine. You pay by the use (you also need to include S3 costs) and you can use it how you want.

This is a smart move from Amazon. Smarter than Sun’s one. The problem with the grid architecture is to parallelize a computation (whether an addition or a program; for instance how do I parallelize int i = 5 + 8; i++; the answer is it is uselss).

Amazon is creating a really amazing web services stack. Those days Google is all the rage, but Amazon is really creating something much more than their ecommerce shop.

The service is currently in beta. I am subscribed to it and plan to test it. What is sure is that it will provoke a lot of “Schumpeter destructive creation”.

Link to the documentation

This is a great interview of Reid Hoffman. A great lesson in entrepreneurship and “enterprise wisdom”.

I do wonder one point. He states: “Smart people tend to think that they can execute on a complex plan. Executing on a complex plan is generally a recipe for failure.” What is a complex plan? How would you define one compared to a simple one?

Link

In this post, John Battelle shows us a Microsoft slide showing how MS expenses outsells his competitor (mainly Google and IBM). John then states Google engineering department operates as a R&D lab.

This raises an interesting question, deeply rooted in today’s questions what is R&D? Should we separate (as we do in France) fundamental research and applied research? What is the best way to do R&D?

If we look at the past’s big names, we can see how much those questions are somewhat falsely named. For instance John Von Neumann invented both computer and game theory (and a lot of other research). Metcalfe invented Ethernet and founded 3Com… Basically, all those human beings have done everything and did what they wanted to.

On an organizational level, this question is way more complex to answer: how should R&D be handled, what is expected from it? Papers? Products? New way of understanding reality?

Link

Search was becoming boring. Google, MSN and Yahoo! have all stopped innovating in those areas. Of course, new companies are trying to redefine the game: JetEye and IntelliText.

They are trying to understand, your context and provides you with search result before you need them.

Something quite interesting and revolutionating have started.

More later…

(Via SearchBlog

This is the last part on my presentation on large scale distributed file system. Here is the report (in French) synthetizing all we have said on the subject.

This presentation was made for a CNAM class and I wish to thank all the professors and the other students for their support and help.

Link

This quite an interesting article. I have advocated for code reviews in my past companies. Sadly without success. This time I can at least say Google is doing it.

This is an article about code review from a Google developer, not from a Google point of view.

Link

For once, Linux users can rejoice. We are on the frontline in the 3D desktop. Xgl is a cool 3D desktop environment.

It seems 3D desktops are going to become the norm (Vista has it, OS X will have it,…) so let us get used to it right now.

Be careful playing with it, Xgl is still highly unstable. Here is some help. It is only my work computer (and on a separate partition).

To be continued. More on my 3D Desktop experience and experiment later.

Don Dodge has been given seven rules through which Kleiner Perkins (the VC company behind Google, Amazon, Netscape and so on) invest or not in a start-up.

(source: roland from Flickr)

• Instant Value to customers - solve a problem or create value with the first use
• Viral adoption - Pull, not push. No direct sales force required
• Minimum IT footprint, preferably none. Hosted SaaS is best.
• Simple, intuitive user experience - no training required.
• Personalized user experience - customizable.
• Easy configuration based on application or usage templates.
• Context aware - adjust to location, groups, preferences, devices, etc.

My few comments:

• It shows how much the business landscape has changed. As others have put it, no one is anymore listening to sales pitch. In the Internet age, most deals are initiated by customers and not by sales guys. Other deals are handled directly at executive levels. Making needs emerge from an interaction with a sales guy is out of the norm nowadays. Internet and new media own this role. Of course, doing so allow easy scale up for a start up.
• The KP second rule about minimal IT footprint seems obvious. You want to build a product. You need to lower as much as possible any obstacle to adoption. KP want company serving markets. Not projects based ones.
• It is also interesting to see how automated adaptation is the norm now for every product: you expect the product to be easily customizable to your need. Whether adding new levels on a videogame or a functionnality in a software, it should be easily doable. If the software is doing it by itself then even better. For instance, Chandler is trying to add it as seen here or here. There is still of values in customization and integration but only if it is automated. Customers want to be empowered. As Bruce Sterling puts it we are moving away from a “systemic” economy to a “spimes” one
• Value lies in community and mass adoption. As we have already seen, customers want to be empowered. They want to be able to manage on their own the product and have it customized to their needs. This fact opens space for mass products which are automatically tailored to one’s needs. For instance, Google, Amazon,… Why do you think they all invest in data-mining?
• Those rules summarize what is hot right now. Not what will be hot next year (next month?).

Via Don Dodge’s Blog

Update Ben Barren is also blogging about this.

After offering Office beta to download, Microsoft is offering now Windows Vista Beta 2 to download.

If even Microsoft is in the beta trend. Are we condemned to use beta for the rest of our life? Anyway it is a really good initiative both to increase the press coverage and increase Vista’s stability.

Link