If there is one company I'd like to see survive and thrive, it has to be Sun. But the challenges they face are daunting. They are getting squished between IBm, HP/Compaq and Microsoft, still quite unable to decide whether they should focus more on hardware or software. Sun has always surprised, but now, the odds seem stacked against them. Some advice from two commentators:

Cringely: "Sun can either find a merger partner to take the company out of its predicament or it can find its own strategy to achieve the same result. Either way, this is a time for Scott McNealy to literally bet the company...One way to do that is through a merger, but the logical merger partner isn't Apple, it is Sony."

Charles Cooper writes on the challenges facing Sun with the Linux-on-Intel combo: "With prices on Linux-Intel systems falling, the pressure is on a higher value company like Sun to justify the higher prices it charges for systems comprising proprietary Unix operating systems on RISC processors. Corporate data managers are especially anxious about reducing hardware costs. What's more, they know the migration to Linux from an existing proprietary Unix platform reuses a lot of the existing code and skills."

It is always interesting to read an interview with Andressen. He is currently CEO of Opsware. Its just about 10 years since he launched Mosaic along with his team at the NCSA. Some comments from his interview with Wired News:

On what he'd differently with Mosaic: If I had to do it over again, I'd probably show some sort of graphical representation of a tree, so you could see what path you're traveling on and could backtrack. I'd also include thumbnail renderings on the tree to show where you'd been.

On Opsware: The idea is that over the last 10 years people have bought an enormous amount of software and servers for Web applications. But while the Internet opened up all this potential power, it also created the problem of actually making all these things work. Our software gives an answer to that.

The big picture: With the Internet, we're really 10 years into what will ultimately look like a 25-year cycle from invention to full implementation.

Video Games is one market Microsoft has been trying hard to conquer with its Xbos console. But it hasn't succeed - yet. Sony's PlayStation has a commanding lead. The reason: a 2:1 advantage in games. The New York Times elaborates:

The market for consoles and video games is worth more than $9 billion a year. Even while losing on selling consoles, Microsoft could still make a lot of money from game makers. They pay the console makers about $10 for each copy of the games they manufacture — and they made well in excess of 50 million games in the United States last year. Games typically sell at retail for $50.

For now, though, there are only losses for Microsoft. It declines to say how much it loses on each console, but industry analysts estimate the figure at close to $100. Sony, by contrast, does not sell PlayStation 2's below cost.

Part of Microsoft's problem is the ambitious design of Xbox: its chipsets and other electronic components are more expensive than those of the PlayStation. Microsoft has also been unable to realize certain economies of scale because sales have not been as robust as expected.

Fortune writes about the new business venture of Red Herring founder Tony Perkins:

AlwaysOn is almost entirely the creation of its members, who express their opinions in a variety of pre-defined topic areas, including "The Wireless Device Boom," "Security in a Hacker%u2019s World," "The Real-Time Economy," and "Entertainment Goes Online." There are three categories of bloggers on the site: ordinary members, who after two weeks already number 5,500; about 100 volunteer "correspondents;" and industry celebrities, who Perkins will interview periodically. He merely asks these people to talk about their strongest opinions of the moment. He%u2019s already posted comments from Michael Dell, John Doerr, and venture capitalist Tim Draper.

"This is the eBay-ization of media," says Perkins. "We've created the arena, like eBay did. We organize the world, then invite members to come in and play." He calls the site a "super-blog," comparing it to Slashdot.org, a phenomenally successful site for serious technophiles that now claims over two million members. "While Slashdot is for techie geeks, AlwaysOn is for business geeks," he says. He will impose editorial order by continuing to fine-tune topic areas, recruiting appropriate bloggers, and contributing heavily himse

AlwaysOn is another example (like Nick Denton's ventures and Corante) of nano-publishing on a small budget.

Nick Denton unravels a few mysteries surrouding the other project he is working on, which is expected to go live in the second half of this year:

What is the Lafayette Project? It's the working title for a weblog media project involving Meg Hourihan and Nick Denton. It's called Lafayette because that's the street in Manhattan from which we're working. It's a project because it seems pompous to call a self-funded three-person team anything else.

What do you mean by weblog media? Systems such as Blogger and Movable Type have made it easy to publish to the web, but the reader experience leaves much to be desired. The more weblogs there are, the harder it is to keep track of them all. That's the problem we're addressing: turning the weblog network into accessible media.

So you're working on weblog search? No, companies such as Google already provide keyword search over weblog posts. We want to help readers browse weblogs when they *don't* know what they're looking for. A best-of-the-blogs show, if you like.

Isn't that what services such as Technorati and Blogdex already do? Yes, and we've learned a lot from the experience of Cameron Marlow at Blogdex and Dave Sifry at Technorati.

Pip Coburn of UBS Warburg writes about "Cold Technologies" (as opposed to Hot Tech):

A cold technology issue is one that commands a major portion of the agenda while having neutral revenue or even anti revenue attributes. A hot technology has the potential to generate revs. So, in 1980, whether one was a fan of the PC or not, both would agree that if the PC took off the tech pie would expand

The PC was a hot technology. Linux is a cold technology. It will shrink the pie. Cold technologies often are issues that are not product related but gain a disproportionate share of the agenda. The migration of the food chain into China is a cold technology issue.

I like the notion of "cold technologies" - this is what the emerging markets need. All the ideas that we are working on - the 5KPC with open-source software, an integrated ebusiness suite for SMEs - are cold technologies. In many ways, disruptive innovations are cold technologies - they are cheaper and simpler, even as they address new technologies.

Even with the Rs 5,000 PC (5KPC), the total cost of computer ownership may still be perceived to be high for a significant majority of the people, especially in semi-urban and rural areas. This is where telecentres (or Tech 7-11s, as I used to refer to them in earlier columns) come in. By focusing on providing access to computers rather than ownership, we can accomplish the same goal of taking computing to every family and employee.

This is akin to the way telecom access in India is available to most of the country via Public Call Offices (PCOs) even though at a national level, there are less than 40 million telephone lines (and an additional 10 million cellphones) for a population of 1 billion. A million of the telephone lines go into PCOs which serve a community rather than a family. Telecentres take a similar approach a computing.

Telecentres are computing and communication centres. They will typically have 5-40 of the 5KPC, depending on the available space. These “PC terminals” are connected to a thick server. The server cost would be about Rs 50,000 (USD 1,000). Thus, with an investment of about Rs 200,000 (USD 4,000), it should be possible to set up a 10-15 computer telecentre. In addition, the telecentre will have telecom services, via a few wired or wireless phones, perhaps an IP phone, a fax machine, a printer and a photocopying machine. It also will have a wireless (WiFi) access point, thus providing coverage to the neighbouring homes and offices. It is connected to the Internet through a broadband connection – ISDN, cable, fibre or wireless.

What makes the telecentre affordable in terms of investment is its use of the 5KPC running Linux. For an equivalent setup using new PCs and Microsoft Windows and Office, the cost would be higher by about Rs 40,000 per PC, making it unaffordable. Today in India, other than the branded chains, most cybercafes use pirated software – but that is not a sustainable model for growth. What the 5KPC provides is a solution which is not only cheaper, but also uses legitimate software. This is the only way to build a scalable, replicable and viable solution.

The telecentres perform multiple functions:

Cybercafe: First and foremost, the telecentres perform the role that cybercafes do - connect people to the Internet. They provide email, browsing and chat, and now increasingly voice-over-IP and video chat/conferencing. The one flaw in a cybercafe-only model is that its only source of income is tenuous – that of charging people by the time they use the computer. This charge has been falling steadily. Cybercafes need additional income source to survive and thrive, and this is where the myriad other services that the telecentre makes possible come in.

Computing Point: This is the real strength of the telecentre. It goes beyond just communications. In a way, by providing communications and community services, cybercafes have leapfrogged ahead – but left behind those users desiring basic computing services. The telecentre offers computing time and storage space for hire – for writing letters, making presentations, or running specialised business applications. It lets people test-drive computers before deciding if they need one at home or in their office.

Digital Library: The server in the telecentre can become a huge repository of digital content. Disk space is cheap. Data mirroring on the local server can be done by dispatching CDs regularly to the various telecentres, and copying the megabytes of data on to the hard disks. What does this is ensure that users don’t need to worry about going to the Internet over slow connections – they get “broadband” access to the content on the LAN itself in the telecentre. Anything that can be digitised can be made available locally for access. The digital content can include mirrored websites, software downloads, educational content, e-books, government forms and notifications, and more.

Tomorrow: Telecentres (continued)