April 21, 2005

Hiring is obsolete

Good writer connects with his readers in a very comfortable way, without pretending to be an author. That alone is the reason to enjoy Paul Graham's writing. His latest result of uncanny observation leads to the idea that the days of hiring are over.

Welcome to the world where you get bought over by big companies.  His advice for graduates is to start your own startup. Master of contrarian thinking will be speaking on the changing recruitment model in Berkeley.

Most CS undergrads hope to get a good job when they graduate.  But
as the age of startup founders creeps downward, I foresee an alternative path for the most ambitious: instead of going to work  for Microsoft, start a startup and make Microsoft buy it to get you.                                                                            

This change will do more than make some young hackers richer.  It  will fuse recruitment with product development.  Instead of applying  for a job and then being told what to work on, you join the company  as a complete development team, with a beta version.  Results: (a) a shift in power from companies to hackers, and (b) an increase in the rate at which new technology gets developed.

Obviously this new model will be a better deal for the best hackers.  But I think it will also be better for the Microsofts.  The few  tens of millions extra that they'll pay will be a bargain for what  they'll get.

This is a very interesting observation. Getting hired-via-acquisition is something which we will be seeing more and more. This is where candidates will go to companies with a clear value-proposition. I think you should hire me because I can do this which will  help you save this much money [or help make you more money] and I will show you how ! .

 Candidate should be able to connect his data structure mastery or machine intelligence genius to how companies save and make money ! That too in plain English. Clear articulation of skill sets and how that connects to productivity or efficient problem solving in a differentiating way is a sure shot way to get a job.

Cisco has mastered this at the mature end of the talent chain by methodically acquiring companies for their talent pool.  Other companies are also jumping into this by having their M&A execs troll university student hangouts. Starting with Slashdot and Sourceforge off-course.  This builds on  the J-Curve mentality as well:

[What is J-Curve..]  That period of time in advance of mass-confirmation of a new idea

You get this advance notification of idea by listening and following the work of this  next generation. In this age where hunt for talent is at its peak.  Candidate's work  and his beta software becomes the resume. Google hits become more important than the references in  resume.

As open source lowers the barrier to build innovation based software companies and universities put more new ideas into the  open source pool, you will see more and more students with entrepreneurial mind-set. This is something to cheer about.

Paul sums up this nicely in his book:

you need to start doing something people want. You don't need to join a company to do that. All a company is is a group of people working together to do something people want. It's doing something people want that matters, not joining the group

April 21, 2005 in Entrepreneurship, Open source | Permalink | Comments (2) | TrackBack

April 14, 2005

Coop model for the open source risk management

Avalanche , a community initiative formed by end customers have come out with a project to help member companies tools and expertise to mitigate open source risk.

The Avalanche COSS Project is building and will provide an Enterprise Class Support organization for Avalanche Members. The objective is to mitigate the risk of open source and enable our corporate members to confidently and economically implement open source solutions.

Interestingly they have a Sarbox project intiaitive as well. If you know how much companies pay for consultant's hotel stay, their invitation fee of $5000 is a steal !

Go ahead, pool your knowledge and save some money.

April 14, 2005 in Open source | Permalink | Comments (0) | TrackBack

April 13, 2005

Complexity is not solved by open sourcing it !

You dont solve complexity by dumping something into the open source community.

Developers have not used the EJB capabilities widely, in part because the technical specification is hard to learn. MacNeil said that creating an open-source tool for the popular Eclipse platform will help drive adoption of the upcoming J2EE server software

This is a lousy argument. Solve complexity by focusing on what really matters most and keep things extremely simple. J2EE is fast becoming a very complex platform to program on. Thats part of the reason why Ruby is looking so exciting and cool !

April 13, 2005 in Open source | Permalink | Comments (0) | TrackBack

Mind your cost and your consultants

Philip Greenspun sharing his own open source venture experience:

At the same time, the customer will always have a few differences from your main source code and to the extent that those features aren't rolled back into the main release, they're always going to be orphans. They're always going to have this stuff on the side that they're going to have to own and maintain, which they really don't want to do. Their business isn't software development. They don't want to have ownership. They'd much rather pay you, the software developer of that code, to make the changes and put those features that are not core to their features into the main distribution of the software so that if they ever want to upgrade to a new version, they won't have to undertake the project of recustomizing that for their special needs. So, basically, if you are in control of a piece of software, you really have the ability to charge a lot more than an ordinary IT services business to make modifications. You have the power to roll that into the next release. So that's one thing that is good about open source.

The thing that's bad about open source is its very unforgiving if your costs and your time goes up. For example, in the early days of ArsDigita, we did a lot of things sort of MIT grad school style. We took fairly young people who wanted to build their careers and professional reputations and we'd have two of them to a project. So two programmers were totally responsible for the project and they met directly with the customer to find out what was needed. They wrote the specs. They wrote the docs. They wrote the code. They tested it with the customer. They made the enhancements as requested. They showed the profit when it was done.

He is  right on money here. This will probably explain why JBoss and MySQL are so successful in making efficient money via customization and support services. Their baggage is low and light. 

Note to all open source services startups - DON'T THINK LIKE THOSE BIG SERVICES COMPANIES. Its better if you model your company assuming  every employee is a geek and a hands-on engineer who can ALSO do solution engineering, clean customer interaction, and generally keeps things agile. Any extra structure is a cost. And any extra cost is a negative for the open source value proposition.


April 13, 2005 in Open source | Permalink | Comments (0) | TrackBack

April 11, 2005

Community model

As we are digging more and more into the governance aspects of open source projects, Javaworld raises the questions around company-controlled communities (eg include JBoss, MySQL etc) versus community-controlled communities (Apache being the leader here).

Transparency, trust and reputation are the key drivers here. Whoever can come out clean on those three aspects will have a good community engagement irrespective of the specific governance type.

April 11, 2005 in Open source | Permalink | Comments (0) | TrackBack

April 10, 2005


O'Reilly and Spikesource roll out a sort of adult-version of Sourceforge.  It's name carries the usual O'Reilly's wild-life theme based naming scheme.

This raises a question (have to ask since lot of well intentioned open source advocates are involved in this project) why try to upstage sourceforge ? It does a great job and has close to 90 thousand projects. Why do they expect a developer to go ANOTHER community/code repository site to hang their codebase or to download  the code. Looks like Spikesource wants to own the mindshare around repository like CodeZoo but without owning the community governance. 

As per their site:

"CodeZoo exists to help you find high-quality, freely available, reusable components, getting you past the repetitive parts of coding, and onto the rest and the best of your projects. It’s a fast-forward button for your compiler"

It's not just the licenses and components which are going through the combinatorial explosion but also the communities, code warehouses, stacks etc.  Instead of this codeZoo they could have focused on providing DOAP wrapper to all sourceforge projects and allow XML api based query straight out of the Eclipse environment.  I am sure developers will go for that.

April 10, 2005 in Open source | Permalink | Comments (0) | TrackBack

April 08, 2005

osbc buzz

Going by all the  blog posts, looks like OSBC was lot of fun. 

Picking pieces from different blogs. Steve Mallet of OSDir talks about Jonathan Schwartz's GPL criticism.  " Country forks" and "GPL as an imperialistic tool" were hotly debated issues.  I think we will hear more about country forks in the future. Open source is not only a good hedge against vendor monopoly but also a creative wedge on the intellectual property issue. 

Its not as if Sun has suddenly found new love for the rest of the world. They see the technology world (specially the Mobile and developer's market) shifting globally. It doesn't hurt to give a sympathetic voice to the 3rd world countries.  Remember all this CDDL thing is as much aimed against GPL as against Redhat. Redhat is well established in the emerging countries.  OpenSolaris is NOT ! They are catching up here.

Considering Spikesource's PR splurge there is definitely attention from the money lords. Taking into factor  Sand Hill's famous tendency to flock together, we will see more clones of Spikesource and SugarCRM pretty soon. Offcourse those will be under stealth and they will have their own differentiators. And I am not joking ;)

Talking about Spikesource,  They should succeed. If they succeed then the legendary might of KPCB will alone put open source in the board room. It will make selling little easier for other teeny weeny startups. Though I cannot resist the comparison between what Spikesource team did in the $115 million flameout Asera (Whitepaper in doc format) and what their test harness in Spikesource claims to do. That combinatorial complexity, dependency analysis and composition all sounds very familiar. Considering component explosion on the open source side they will certainly get plenty of attention, if only they can let the technology do the talking rather than drumming the PR parade.  It will be fun to watch how this one develops.

Checkout the buzz on blogosphere

April 8, 2005 in Open source | Permalink | Comments (0) | TrackBack

April 03, 2005

Open standards first and foremost

Mat Assay [of Novell] on the role open standards play :

So, while open source is a Very Good Thing, it's not the only thing one should consider when looking at software. If you're an IT buyer, open standards should be the overriding concern, with open source a close third. Your second consideration should be whether the product meets your needs. Why standards before needs? Because your needs tomorrow may change - if you're locked into a product because you bought into a closed standard, you're bien fichu, no matter how good it was at the beginning.

I started following his blog recently. Being an insider he knows how the vendor moves are shaping the open source world.  His reiteration of this correct sequencing starting with open standards and THEN moving on to customer needs and open source will help keep things in perspective. Little pragmatism will go a long way in helping all who care for open source.

Combining open standards with open standards will prove to be a winning combination. Both aim for maximum adoption and seek volume.

April 3, 2005 in Open source | Permalink | Comments (0) | TrackBack

March 26, 2005

Xen and the art of causing MS-heartburn

Not easily to give into all the projections of MS-killer  and death of all proprietary software.  I am skeptical of any general statements about the potential or consequences of open source software. As always reality is buried in the details and segments.   Software world is a complex system and like any complex system it has evolving equilibrium points.  Between those equilibrium points one sees many game-changing possibilities.

One such possibility relates to the business goal of server  consolidation and scaling. That's where this open source virtualization software comes in - Xen. You will hear a lot about this.  This is what it does (via Zdnet - Gartner ) :

In March 2005, Intel announced the introduction of its Vanderpool hardware virtualization technology. This technology gives Xen the ability to support any operating system. Xen is small, with only 50,000 lines of code. This simplicity and its open-source license bode well for security and stability. Its availability will drive innovation across computing platforms. The product could become as common as the PC BIOS is today. XenSource s goal is to provide services and support to help companies implement Xen-based systems. The ubiquity of Xen drives this business model, so the open-source zero-revenue model becomes a benefit rather than a cost. XenSource has backing and support form Intel; IBM; HP, AMD and others

Same Gartner report mentions three  challenges this technology has to solve before its widely adopted -

  • technology can insert into target platforms without compromising performance, reliability or supportability
  • ecosystem of technology providers that plug into the Xen environment, enhancing platforms without changing one bit of the system image that runs in the platform partitions
  • Challenge from Microsoft.

Precisely to solve these challenges Xensource (commercial arm of the Xen creators) has managed to get Kleiner Perkins and Sevin Rosen on board.  In the mid-to-long term VMWare (and now EMC) should worry about this. In the short term this alone will expand the market for virtualization and may spawn new set of system level applications.   More stories on Xen is available here , here,   and download from here.

On a lighter note, staying on top of the open source buzz is like waiting for some shoe to drop. Which is happening one every day (or week depending on what you are counting - libraries or packages).

Probably both proprietary and open source world is working towards some equilibrium. How long it will take to reach there is anybody's guess.

March 26, 2005 in Open source | Permalink | Comments (0) | TrackBack

March 24, 2005

Movement to the middle

David Berlind over  at Zdnet quotes Microsoft's open shared-source honcho Matusow on Microsoft's movement-to- the- middle strategy:

The difference for us is that we felt very strongly that we knew that parts of our strategy were not going to meet open source and what the whole world was going to call open source. So instead of getting into a semantactical battle every other day, call it shared source and say "Look, we re not going to make decisions every day that the open source community is going to agree with." And truth to tell is that's not the most important issue. What's most important is that customers are having their needs met, partners are having opportunity built to go and spin off new businesses and build a greater ecosystem around the windows technologies, academics can do the work that they want to do, hobbyists can get to the interesting technologies. That's what matters to me. Not whether I can call it open or not.

Like it or not but this is where giants will eventually move.  They will create an over-arching blueprint license (license will become an engagement strategy to control platform) to create an eco-system. Sun's CDDL and Microsoft's Shared-Source program is nothing but an attempt to lay the foundation for having their own say in the open source movement. It will be interesting to see how far they will succeed and how other big software vendors formulate their own "proprietary" open source strategy. 

It will require many lawyers and many creative marketing gurus to craft a compelling firewall around their current revenue stream.

March 24, 2005 in Open source | Permalink | Comments (0) | TrackBack