InfoWorld CTO Chad Dickerson writes that "commercial software has its place, but open source and hosted solutions now dominate."

I'm finding the "open source or outsource" question at the center of my IT decision-making process these days. My team is still implementing scalable and flexible systems that I'm confident will serve the needs of our business for the foreseeable future, but I'm buying very little software these days. Custom-developed software built on reliable open source frameworks wins the day when the benefits of slick commercial software can't outweigh the licensing costs.

On the other hand, when done for the right reasons, outsourcing can save money and reduce unnecessary distractions for in-house IT teams. I have never been a fan of outsourcing as a dumping ground for IT problems, but outsourcing is becoming increasingly attractive in two particular scenarios. The first instance deals with outsourcing relatively generic functions that rely on widely used commercial software packages that require specific expertise and near-perfect uptime -- think hosted Microsoft Exchange or Lotus Notes. In these scenarios, it's difficult to run these systems noticeably better than an outsourcer.

The second, more interesting emerging scenario is the Web-based hosted service with a well-documented Web services API, an offering pioneered by Salesforce.com with their sforce platform. In this type of outsourcing arrangement, you get all the benefits of a core service without having to own software or hardware, but you retain the option to extend the system to address specific needs of your business. I think of this phenomenon as "open outsourcing" -- a service that sits outside of my datacenter with secure access to a rich API. Platforms such as sforce dispel the notion that outsourcing is about cramming a variety of square-pegged businesses into a single hosted round hole.

n areas where outsourcing doesn't make sense and IT needs to build custom functionality unavailable in commercial products, open source is an increasingly viable choice.

Brad Feld has a fascinating story of his entrepreneurial days, and writes about the 10 precepts that he held dear when building his company. They still hold true today.

We must be financially successful to be a business.
Respect.
We want our clients to love our service.
The success of the business along all dimensions is everyone's responsibility.
Our work environment must be comfortable and stimulating.
Strive to be the best.
We want to grow into new areas that we have not yet explored.
Work should be a substantive, positive part of our life experience and personal growth.
We want to be our clients' best computer technology investment.
We will periodically meet as a group to evaluate, question, and re-affirm our precepts and our mission.

Om Malik points to a Baseline story about Heinz in Europe, which "has basically replaced most of its voice contracts with a company owned IP-PBX service, which has allowed them to cut costs seriously."

Om adds: "It shows that if mega-corporations think creatively, they can become their own phone companies. They can interconnect their partners, their suppliers and manufactures on a private VoIP network and basically cut their long distance and business communications costs."

Interesting to see the projects coming around the Google Desktop. Kent Fitch writes:

If Google Desktop were accessible from places other than the machine on which it was installed, some interesting possibilities would arise:

Install it on a web site and use it to index/search/cache copies of the static pages (and unless you are very careful, the web server logs, configuration settings, scripts with embedded passwords...)

Install it on a workgroup file server and use it to index/search/cache copies the contents (including secret memos, tedious MSPowerpoint presentations, suprising job applications and resumes, commercial-in-confidence tender information...)

Install it on your PC's at work and home and your laptop and search/retrieve on any from the others (along with anyone else who uses/steals/spoofs the right IP addresses)

Use it as the basis for a distributed, maybe a P2P search service

The Java program described here allows others to search your Google Desktop.

For some reason I found myself thinking late last night about what it means when there are millions of local google HTTP servers running on millions of individual PCs, yearning to be connected. What kind of innovation might spring from such an ecology? Of course uploading your local content - just that which you want to upload, of course - to Google's index is one possibility. If you could do that, you could pretty much wipe out hosted solutions for micropublishing (ie blogs), for example. Your machine, given advances in broadband and computing power - would become your web server, just as it was in the beginning, when there was so little traffic on the web (before the hosting business took over...). GDS could also lay down the framework for some killer distributed computing hacks, stuff that Google has demonstrated an affinity for in the past (Google Compute is built into the Toolbar).

But thoughts fail me. What comes to mind from you all when you think of a world where all our hard drives can seamlessly be connected to the Mother Index/Platform?

Dana Blankenhorn sums it up in one word: Applications. He reminds us:

When my wife first got her current job, as a programmer at a local transaction processor, she said she enjoyed it because her computer "actually does something."

Fact is that operating systems, by themselves, do nothing. They are a platform on which you build programs that actually do things.

This remains the biggest challenge faced by open source.

Microsoft's big advantage is that it has an ecosystem whereby applications can reach the market. It has nothing to do with the quality of Windows. Developers know they can reach buyers with Windows applications, through established channels of distribution.

Now it's too late to expect a line of shrink-wrapped Linux applications over at Fry's or BestBuy. Today's channel consists primarily of downloads. But even here there just isn't that much, only partly because the idealists' Linux model is you don't sell software at all, but services.

Yes. I know that Google is a Linux application. I know most Web servers run Linux, so Web services will mostly be Linux-based. I also know that there are Linux analogues to all the office applications you use most often. And that there are Linux versions of many middleware programs.

But I'm still going to remain on the hunt for more open source applications, and for open source distribution. It's the fastest route to open source clients.

Following the purchase of Stata Labs (Bloomba) by Yahoo, Paul Kedrosky writes:

It is now obvious Yahoo is preparing a major thrust in the email market. Oddpost brought server-side smarts; Bloomba brings client-side search in an email context. Yahoo already had the reach, storage, and resilience to create an industrial-strength platform.

Will we see truly persistent messaging that bridges location, client type, and message store?

Do you have any advice or words of wisdom for people starting their own businesses?

Based on my experiences, there are three things that I’d like to tell people starting their own businesses: