April 03, 2010
The Importance of Branding in Open SourceTwitter recently announced they intend to "open source" FlockDB some time in April. FlockDB is Twitter's high performance database. It uses a heavily sharded MySQL server, with Scalla providing the functional layer. Opening up FlockDB is something Twitter can safely do now that the company is the well established leader in the market niche it occupies. If FlockDB had been released while Twitter had been in the early stages, another company could have picked it up and used it to create the same type of service. This shows the power of branding in an open source world.
February 14, 2010
Web 3.0, Social Gaming and the Culture of NarcissismYesterday, I presented a talk at the BIL2010 conference in Long Beach. It was a last minute affair, but I've been meaning to get some new ideas out in some form, so I saw the show as a good venue. My talk is titled Web 3.0, Social Gaming and the Culture of Narcissism.
My first order of business was to call for a definition of the term "Web 3.0". The term Web 2.0, and more importantly what it means, needs some extension and what better way to add those extensions than to create a new version. Web 2.0 focusses solely on the technical issues of web development. Some mention is made of user experience, but at closer look what is specified is something called "User Centered Design". The problem with user centered design as it relates to web applications, is that it only deals with the user interface as it relates to the user interface and how the user navigates the site. The users other needs are not considered. Those of you who are tired of dealing with chat spam, unnecessary collection of your personal data, automatic solicitation of people in your email lists, etc. will likely agree that the user experience encompasses more than just how easy it is to navigate the site or perform actions. I think the main distinction of Web 3.0 is a pledge to protect users privacy, not contact people in the user's email lists, not push chat spam at them, nor any other action that may be considered annoying or invasive. This specification is basically a pledge to follow a set of best practices affecting a users overall experience, and attempting to make that experience as unobnoxious as possible.
On Friday, Johannes Bhakdi gave a talk titled Web 3.0: User Generated Business. I unfortunately was unable to watch his talk, but visited his web site. Check out his well done video explaining his proposal for online business. It is basically what I've previously proposed for compensating writers for online content, only he extends the model to encompass all forms of online commerce. The idea is quite viable.
Speaking of online commerce. Part of my talk dealt with online advertising, in particular advertising from within online games. I presented a concept I've employed in a new online social game I'm working on, called Heist. Heist is a crime game, in the genre of Mafia Wars. The concept for Heist was the brainchild of Don Marti. Heist was designed to a Zen like asthetic, simple, clean and fast. A different type of gamer will likely be drawn to it. It was designed with the busy person in mind, someone who doesn't have the prodigious amount of free time required to play games like Mafia Wars. The fun of playing Heist will be in reading the scenario descriptions while viewing the accompanying art work and photographs. A Heist scenario is kind of like reading a comic in the New Yorker. The tense element to Heist will be that any member of a crew can "rat out" the other crew members, spoiling a job in progress. Back to the new marketing concept in Heist. I had the idea that items in the game can be advertising. For example, some scenarios have artwork as the object of a heist. In some cases I've included hyperlinks to the artist who created the work. This can be extended to include automobiles (click on the Ferrari and you end up on the Ferrari web site,) and so on.
The other aspect of online marketing I spoke about was true affinity marketing. I've suspected that standard affilate marketing techniques have been diminishing in effectiveness. I think the next wave in online advertising is to build a strong relationship with the viewers and present marketing to them in a way that is more closely aligned to their personalities and affinities. One way marketers have done this is through celebrity endorsements. Having a famous athlete's Twitter page background plastered with a company's logo is a form of this, but of course that is nothing new. Still, that technique does come across as more intimate than simply pushing ads at people that may be of general interest. This also comes back to my proposed definition for Web 3.0. Happy users are more inclined to look at those advertisements that are dangled in front of them. So, my proposed formula for online success is: first, make the user happy, then only show ads the user is likely to have a strong interest in and finally find some novel ways to incorporate the advertising into game items or other objects in a fun way that won't offend users. Sure enough, after my talk I was approached by a couple of people who study online marketing and they confirmed that traditional online marketing techniques have been providing continually diminished returns. I will say that the bad economy undoubtedly also comes into play (there are ways to adjust effectiveness studies to factor out macro economic effects and I don't know if the study they alluded to incorporated such techniques.)
In all, it was an interesting Saturday in Long Beach.
August 29, 2009
Nokia's TrollTech acquisition bears fruit with new Linux phoneLast January, Finnish cell phone developer Nokia acquired TrollTech. TrollTech is a small Norwegian company that produces a cross-platform application development environment called Qt. Qt is the foundation of the KDE desktop. Some notable Qt applications are Google Earth, Opera, Skype and Adobe PhotoShop Elements. I saw the acquisition as strategic for two reasons: 1. It gave Nokia an excellent cross-platform development tool for use on their phones. 2. It protected TrollTech from acquisition by a hostile entity.
The question on many people's minds has been when, if ever, Nokia would release a cell phone using the Linux operating system. Linux is being used as an embedded real time operating system in a variety of devices, such as video cameras and high-definition television sets, so it could certainly do well in a cell phone.
Now the wait is over. Nokia has finally annouced the world's first commercially available cell phone running Linux, the N900. This new tablet phone boasts impressive hardware specs and with Linux is capable of true multitasking performance. The phone is scheduled for availability in October 2009. You can view a short video of the phone in action here.
Looks like the Qt acquisition is about to pay off.
April 23, 2009
Oracle and Sun a Good MatchOracle's pending acquistion of Sun Microsystems ends a torturous chapter in Sun's history. It also begins a new era for Oracle. The Sun acquisition should transform Oracle from a database and application package software vendor into a full-fledged computer technology company. Oracle will now have a balanced and formidable product line of hardware and software solutions. They will be equipped to compete full-on with IBM and HP.
Both IBM and HP allegedly made offers for Sun. It is good for Sun, Oracle and the San Francisco Bay Area in general that Oracle prevailed. The deal is good for Oracle, for the afore mentioned reasons - Oracle now has an array of hardware, storage, operating system, database, development and applications software, making it a one-stop shop and reducing dependency on outside vendors for their customer's system needs. The deal is good for Sun, in that Oracle will be the least onerous parent - Sun's product profile is so similar to IBM's and HP's that acquisition by either of those companies would have undoubtedly resulted in massive layoffs. IBM or HP would have absorbed the technology and jettisoned all but the most critical Sun employees. Oracle, not having most of the products Sun brings them, will have to retain far more Sun employees to sustain the technology. There will be many layoffs under Oracle, but not nearly as many as with the other two suitors. This in turn is relatively good news for the San Francisco Bay Area, in the sense that not as many people will lose their jobs.
There have been rumors that Oracle will terminate Sun's new Niagara line of SPARC processors. If they do, that will be a big mistake. Having their own processors will allow Oracle to otimize their operating systems (Linux and Solaris) and will enable them to offer customers a complete system bundle. Oracle will now be able to offer customers the advantages of a one-stop shop, including pricing advantages (as in "buy all your stuff from us and we'll knock 10% off the total price.") Owning their own microprocessor is also a powerful insurance policy for Oracle, giving them independence from other computer vendors. If Niagara and the computer division can break even, or even lose a little money, Oracle can still benefit from keeping them. IBM has traditionally been able to charge customers a premium for the peace of mind that comes with having a single technology provider supporting their systems 24/7/365.
The one area where Oracle currently can't compete with IBM or HP is professional services. Indeed, one major reason Sun stagnated was their lack of virtually any professional services division. IBM is king of professional services, with everyone else playing catch up. HP finally recognized that and acquired EDS. HP should have acquired Price-Waterhouse Coopers when they had the chance - when they didn't IBM did. But Oracle does have a professional services group, to support their database and applications, so at least they have an organization upon which to build.
Regardless of the outcome, I look forward to seeing how Oracle plays their hand with the Sun acquisition. Will Oracle develop into a full-fledged computer technology company, or will they become the West Coast equivalent of Computer Associates?
May 14, 2008
HP to acquire EDS - too little too late?It has been announced that Hewlett-Packard plans to aquire EDS. If the meger actually finalizes, it will create an IT services organization that is approximately equal in revenue to IBM's Global Services division. The trade press has been ruminating over this, one of them proclaiming the acquisition "Validates IT services" (tell that to IBM, who has been quietly earning ~ $44 billion a year for quite a while.) Indeed.
I find this acquisition interesting, because it highlights a key difference between IBM and Silicon Valley technology companies, that not many have paid attention to: HP has been the only major Silicon Valley technology company with a large presence in general IT services.
I believe the lack of attention on IT services has hampered certain Valley companies, especially Sun Microsystems. Sun made an effort in the early 90s to establish a professional services group, but it didn't work out. HP came close to purchasing Price Waterhouse Coopers, but balked at the price and backed out. IBM then almost immediately purchased PWC, adding to their then insurmountable lead in the sector.
Coming back to the current merger possibility, I don't see how purchasing EDS will help HP in the long run. EDS is a mature company, busily servicing existing customers. Depending on who occupies the White House next January, EDS could experience an increase in government contracts (EDS's primary business), which would help HP's bottom line. But EDS, in my opinion, is simply not a threat to IBM in the private sector.
So, while EDS is not a bad acquisition for HP, just as Compaq wasn't, it isn't an industry-shaking event. HP should really have acquired PWC when they had the chance. IBM's Global Services may experience a downswing due to a faltering economy and EDS an upswing due to increased government spending, but that really doesn't change the fundamental dynamics of the industry. IBM seems to have always held the trust of private sector Fortune 500 companies, which I think is more valuable in the long run than a temporary artificial mini-economy created by government spending.
A quick note to the journalist who publishes articles almost identical to my weblog entries a few days after I post them: I know it has been a long time since I've posted. I haven't read your publication much in that time, so I don't know how you've been doing. I do hope my lack of posting hasn't caused any undue hardship.
November 17, 2006
Dynamic Content for Technical BooksOne of the challenges in documenting new technology is the rate at which it can change. This is especially true with some of the new software technology. This presents a problem for publishers and their audience.
Online publishing is a fairly recent development, that has potential to alleviate this problem. Some publishers are currently offering books for sale online, in PDF format (often referred to as "ebooks"). This business model is a very efficient form of distribution. Production costs are amazingly low for online books. This has enabled publishers to sell ebooks at greatly reduced prices. I suspect profit margins for ebooks are quite good. Online publishing does increase the risk of piracy, but I believe this won't be a big problem with technical books.
What I find most exciting about online publishing is the potential for creating work that can be regularly updated as the subject matter evolves. This will be very attractive to readers. The publisher can charge a small additional fee when an updated version is available. Or, the publisher can charge slightly more for the initial download and allow readers free access to updated versions for a limited time period.
O'Reilly and Associates is the leader in online publishing. O'Reilly's online publishing infrastructure is developed enough where, with some enhancement, it can support online dynamic content. O'Reilly's Safari online subscription service is very close to the concept of dynamic content.
One of the challenges of delivering dynamic content is the issue of how authors will be equitably compensated for their continued work on a title. There are several solutions to this problem. If a publisher charges a flat subscription fee for each book, the author can receive a percentage of the subscription, equal to the author's royalty percentage. In the case of a business model like Safari (where readers pay a single subscription fee to view an entire collection of work), a compensation formula could look like the following:
( # of page views of author's work / total # of page views for all work ) X ( total subscription revenue of catalog for time period, pro-rated for time period ) X ( author's royalty percentage )
The weakness with the above formula is the potential for mischief surrounding page views. Someone could create an automated bot that continually refreshes pages or downloads a lot of material, artificially inflating activity for a particular author's work. Of course, I'm not suggesting anyone would actually do such a thing, but it is a possibility. An alternative compensation formula could be:
( total # of pages of author's work in collection / total # of pages in collection) X ( total subscription revenue of catalog for time period, pro-rated for time period ) X ( author's royalty percentage )
The objection to this formula is that every author receives an amount based on percentage of content in the catalog, regardless of popularity. Still, I personally regard this method as the best, one that is immune to traffic tampering.
Another compensation model would be for the publisher to simply pay authors a flat fee every time modifications are made to a title. An editor would have to approve any such modifications and an amount agreed upon for the work.
Dynamic content is an interesting concept that promises to improve the value of technical books. We will have to wait and see if it ever actually happens.
October 30, 2006
Open Source Digital Media
The last remaining obstacle to open source operating systems gaining traction in the commercial mass market is digital media. However, digital rights management (DRM) is necessary when dealing with the recording and film industries. This has prevented open source alternatives to the audio and video codecs available for Microsoft Windows and Mac OS X.
One important fact to keep in mind about the "secure" media formats offered by closed source vendors: they are always cracked. So what is so secure about them? This is the key to an open source alternative.
What some parties in the open source community need to do is: First, develop alternatives to Quicktime, WMA, etc., along with DRM technology to complement them. Next, create one or more online music and film stores, similar to iTunes and Rhapsody. Finally, go out and cut deals with the artists, distributors and their agents. Offer them a better deal then they are currently getting. There are many parties capable of succeeding at this.
The DRM component of an open source solution may require monitoring. In this case, the monitoring would be performed by a neutral organization, following a strict set of operational guidelines designed to protect people's privacy. Even better if federal statutes could be enacted to protect the personal data, similar to the tough data privacy laws in Germany (not likely to happen here, but hey you never know.)
This is a very viable business model, which I fully expect to see materialize before too long.
July 30, 2006
IBM and Adobe Make Some Moves
IBM Alphaworks just released
version 1.1 of the Cell Broadband Engine SDK.
The kit now supports Apple Power Mac G5 and IBM PowerPC, which gives that
hardware a new lease on life. The SDK includes a Cell Broadband Engine (CBE) simulator, the SDK actually
runs on Fedora Core for now. The simulator is designed to give developers a head start
with CBE applications. However, with CBE, IBM is finally developing their own Linux distribution.
I was wondering if and when they
would actually do so. IBM may also make CBE the processor for most, if not all,
future hardware platforms. If you follow the news closely, you can catch hints of
things to come. In this case, the hint was last year when Linus Torvalds spoke favorably
of the Power 5 architecture and mentioned that he uses a Power 5 system as his development
box. IBM will likely still support other hardware and continue endorsing SuSE and Red Hat, at
least for the foreseeable future. At some point, IBM may well support only their own Linux distro.
Speaking of hints, we may be witnessing one in Adobe's move this Spring to
improve Flash on Linux.
Adobe's Flash Player Product Manager talks about this in her Web Log.
She mentions that Flash 9 for Linux is expected in early 2007. This is
interesting, because that is when KDE 4 is expected. Adobe has also provided one
of the Linux Flash engineers, identified as Mike M., with a Web Log
to interact with the Linux community regarding the project.
I've speculated that Adobe may port all their applications to Linux, the primary
impetus being the blazing graphics performance of the Cell Broadband Engine.
If you were Adobe and were thinking of this, it would make sense to first make
sure Flash runs well on Linux. If I were a venture capitalist, I would want
to know if Adobe is indeed porting the products, and if they are not I would go out
and round up some talent to start a company providing similar products for
Linux on the CBE. The performance improvement offered by the CBE presents a
rarely occuring opportunity.
July 01, 2006
MySQL Leverages Closed Source (and Vice-Versa)
MySQL running on MS Windows is nothing new. What is new is the two companies formally recognizing and supporting each others products.
Closely related to this is news of Microsoft and JBoss working together to improve interoperability.
It looks like this is what a Microsoft executive was referring to recently when he said Microsoft intends to leverage open-source software. We'll have to keep our eyes open and see if any marketing results from these partnerships.
Market dynamics are quite interesting right now. These moves could turn out to be more significant than they appear.
April 06, 2006
Digital Media as Commodity
The latest hot rumor has Google entering the online music business as an alternative to iTunes. Apparently, some entities in the music business are unhappy with the iTune distribution and pricing models.
There have actually been alternatives to iTunes for some time. Companies such as Rhapsody and Napster (now a legal for-profit concern) have been selling music over the internet for several years. There are also a slew of smaller companies offering music from lesser known artists (which is good exposure for those artists.)
Google is the perfect choice for a new music store. No company works the web better than Google. Google's huge traffic and mind-share will instantly drive customers to the store. Expect Google to, at some point, also offer movies.
The online entertainment business is wide open, with room for many players. It really isn't overly difficult to set up an online entertainment store; the critical factor is securing content.
March 05, 2006
Microprocessors Everywhere
Philips recently announced their new low-power consumption, (relatively) high-performance microprocessor. IBM just open-sourced their Power5 RISC architecture, allowing anyone to build Power5 systems. AMD and Intel have been locked in competition for some time, with consumers benefitting from the resultant performance increases and cost reductions. Transmeta offers system builders a viable alternative for low-power consumption CPUs. IBM and Sony are about to launch the most powerful mass-market CPU in history. Even tech boutique Sun Microsystems has developed a powerful, multi-core CPU featuring 32 execution units on a single chip.
The interesting aspect about these recent processor releases is the decision by their creators to port Linux to most of them, and the very public announcements of this. Virtually every computer technology company is acknowledging and accomodating Linux and open-source software. It appears that hardware will become even more of a commodity, with software serving as the differentiator. As open-source software matures, hardware manufacturers are enjoying freedom to experiment with new designs and can shorten product development cycles without dependence on a proprietary software vendor for operating systems and other software, while simultaneously reducing both R&D and unit production costs. The sleeper is finally awakening.
February 27, 2006
Another Direct HitEarlier in this weblog, I talked about the concept of a hand-held device that will serve as a combination cell phone, music player, video player and internet phone. Bill Gates recently announced Microsoft's plans to create a device sounding almost identical. You're welcome, Bill. And hey, while you're at it, you should consider acquiring Transmeta. Their low power consumption processors would make an excellent choice for your new device. Transmeta has fallen on hard times, so they should be ripe for the picking. Their technology is excellent.
February 10, 2006
Sun-Tzu and the computer industry
The dictums of Chinese general Sun-Tzu have achieved huge popularity among 20th/21st century
business leaders. Perhaps the most popular Sun-Tzu quote among such leaders has been:
"Whoever is first in the field and awaits the coming of the enemy, will be fresh for the fight; whoever is second in the field and has to hasten to battle will arrive exhausted."
We often see this philosophy in action. Apple Computer has achieved an appearance of leadership by constantly introducing new products the past two years, mostly new iPod models and recently a new line of Intel-based personal computers.
The above Sun-Tzu quote is often a bad analogy for the dynamics of the computer business,
where first-adopters are often eclipsed by their competitors. Sun-Tzu was referring to troops preparing to do battle, a situation where a person's physical condition is important. Extending the principle to corporate affairs is often inappropriate.
IBM and Sony will definitely be "second in the field", but will hardly arrive exhausted.
IBM and Sony's positions are perhaps better described by another Sun-Tzu adage:
"...first you win the war, then you fight it."
Bottom line: if the thermal management in IBM's new Cell Processor allows the chip to perform
without over-heating, then IBM most likely wins the war and all that remains is to fight it. If the Cell suffers from the same thermal problems of its Power5 predicessor, then IBM and Sony should not even bother taking the field until they regroup.
January 18, 2006
Why use Ruby on Rails?
I was corresponding with a friend recently, and asked his opinion of Ruby on Rails. The following two paragraphs are his response:
"Ruby's quite good, well worth learning, astronomically better than PHP or
Perl, and a lot more fun than Java. But for me it's not so much better than
Python that I'd switch (better in some ways that aren't very important to
me, worse in others that are important to me). I prefer the simplicity of
Python syntax. I already know Python well, and it has a lot of excellent
established well designed library modules and extensions, that have had all
the kinks worked out of them over the years. I'd use RoR if a project
required it, and I'd learn it if I had the time, but given the choice, I use
Python.
The TurboGears framework for Python is like Ruby on Rails in a lot of ways,
but there are things I like about it more than RoR. Especially the
templating system, called Kid. I'm not impressed by RoR's weird non-XML
templates. Kid is totally XML-centric, very Pythonic, and instead of trying
to reinvent a kludgy half assed template language syntax, it just lets you
use 100% full power Python via 100% standard XML. Never accept a template
language (like ASP) that tries to pervert XML and can't be processed with
off-the-shelf XML tools -- it's just not worth it. Also, some template
languages attempt to reinvent the if and loop statements and variables and
functions in half-assed ways, but I'm happy with Python and don't want to be
"protected" from it. That's the nice thing about Kid -- very Python and XML
friendly."
His point about generic XML compatibility is excellent. I also appreciate what he says about Python. I'm a Python user and think it compares well with Ruby.
So, why then would someone want to use Ruby on Rails? Well, I for one have yet to dive into XML, so I like the fact that Ruby on Rails doesn't require me to learn any XML. I have lots of experience with relational databases, so I can always work with my data from whatever back-end (MySQL or Postgres) my RoR app uses. I also think we will see the RoR framework extended to accommodate those who wish to work with XML. There is at least one Bay Area start-up (in stealth mode) that is working on extending RoR.
December 04, 2005
The Language Wars Have Begun
There has been much heated discussion about the virtures and pitfalls of different programming languages recently on various internet technical forums. O'Reilly Publishing's web site is one focal point for this. A recent O'Reilly title, Beyond Java, written by Bruce Tate, explores the topic. The author's thesis is that Java is about to fade away and be replaced by a more agile language; he feels that Ruby is the main contender.
The ensuing debate has been quite interesting and is very good for the computer programming community. A lot of people are seriously re-thinking their assumptions about programming languages. One interesting observation is the apparent lack of objectivity by many in defending their programming language of choice. Software developers are supposed to be logical, detached and efficient in their pursuit of programming perfection. Hey, looks like they are human beings after all!
One point that is being made continuously in these discussions is the importance of programming frameworks, especially in enterprise software development. It looks like framework development will assume even more importance in the near future. This computer language debate will most likely turn into a framework debate.
O'Reilly has published a brief interview with Bruce Tate and some other software evangelists, you can view it here. Enjoy.
October 12, 2005
Video iPod brings us one step closer
Apple's new video iPod is an excellent proof-of-concept for digital convergence. It offers full video viewing in a very small package, along with music playback. It also contains personal information management software. Apple has been very aggressive with the iPod, introducing new variants in rapid succession. This aggressiveness has kept their competitors off-balance and has whetted the public appetite for the product. Apple is definitely on the cutting edge of digital media technology. Their QuickTime file format is used in MPEG-4. FireWire is another technology that for all practical purposes was delivered by Apple. I don't care for the proprietary nature of Apple products, but they have done an excellent job with OS X and the iPod. The iPod has made Apple a full-fledged consumer digital device company, in the same arena as Sony.
The new iPod plays video that has first been downloaded onto a computer, then transferred into the iPod. The next logical step for this type of device is to add wireless and wireline internet connectivity. After that, integrate the product into a cellular phone (and enable VOIP over the wireline ethernet port.) That will be the product that enables mass market digital convergence to flourish. The market will then depend on infrastructure and content delivery. Apple's partnership with Disney is a first step in that direction. The other players are no doubt studying this with great interest.
October 05, 2005
Digital convergence is starting to look real
All the recent news on digital convergence brought back the memory of a breakfast meeting with an Alcatel executive in the Summer of 1999. At the time, Alcatel was enjoying recognition for their VOIP business telephone and had plans to expand VOIP into wireless phones and other devices. They were talking about developing a comprehensive product offering that would be used to build-out a complete IP-based digital network. They were negotiating partnerships with Cisco and some regional U.S. telcos. I remember Pierre showing me some very general diagrams of Alcatels "Intelligent Network." I appreciated the ambition, forward-thinking and vast resources Alcatel brought to bear on the problem domain. I had to decline his polite and generous offer to relocate to Paris, for various reasons, one of them being that it was obvious digital convergence was a way off from becoming reality. Alcatel was also stepping into the role of an early adopter, usually not a good position. They had to know that at least several competitors coveted the market and were waiting for the right market timing. In the world of digital convergence, market timing is critically important. The technology, product, and service dependancies in the integrated digital market are very difficult to accurately analyze, and any misstep can be very costly.
Here we are, with 2005 winding down, and now digital convergence is starting to look very real. The big news is the Internet portals, such as Yahoo and Google, starting to offer VOIP and announcing plans for widespread wireless access. IPv6 is rolling out, and new high speed wireless technology, such as WiMax, is on the horizon. Most cellular phones are now equipped for Internet access, many of them having their own operating systems that enable things like video games, appointment schedulers and other PDA functionality. Speaking of PDAs, Microsoft and Palm just announced a partnership where Microsoft will supply cell phone chipsets for Palm to integrate into hybrid PDA/phones. This was a brilliant move by Microsoft: offer a struggling PDA company a chance to avoid oblivion and sell lots of Windows CE licenses and cell phone chipsets in the process. Don't be surprised if Microsoft ends up acquiring Palm. It looks like we are now out of the opening and solidly into the middle-game of the digital convergence contest. The question now is how far are we from the end game?
No one player has all the pieces to dominate the market. The final moves will involve alliances. Technology standards toward the top of the software stack can mitigate the need for some alliances, but the current culture of greed and control make such standards difficult if not impossible to achieve. The other inflection point is the question of how some of the players will generate revenue. The question so many people are asking about Google's moves are "How are they going to make money offering free services?" I can think of two ways: advertising and transaction fees from credit card companies for sales made over the provider's infrastructure.
Make no mistake, the game gets more interesting and intensifies from this point. The stakes are too high. My favorites at this point, in the U.S. market: Google, Microsoft. Dark horses to watch: Bell Atantic, SBC. The wildcard: Yahoo. Other companies will have to line up behind these leaders. The telcos could end up partnering with the others. Other interesting possibilities exist, but this is the safe bet. The telcos have a dismal record developing and managing technology, I don't expect that to change. What keeps them in the game is infrastructure, that the others need to deliver content. Wireless technology such as WiMax may change that. Digital convergence will be driven by software; the companies that drive the market will need strong software offerings, capital and marketing muscle.
February 07, 2005
Hardware can still provide an edge
IBM, Sony and Toshiba jointly announced a new microprocessor today.
Called the "Cell" processor, it is basically a multi-core IBM Power chip.
It should perform up to 10 times faster than current microprocessors.
My last weblog entry talked about the importance of software, and how
it is the differentiator in today's world of commodity hardware. I still
stand by this statement. However, there are times when really fast hardware
can prove irresistible and cause users to ignore other considerations.
In all fairness to hardware developers, software is also becoming a commodity,
which makes hardware innovation just as important as ever. Also, even the
best software won't overcome problems with unreliable hardware, so
hardware quality is always important.
Other cpu makers have been preparing multi-core microprocessors recently,
so this new IBM/Sony/Toshiba offering will most likely become one of
several processors with similar performance (which will again make
software the differentiator.)
Interestingly, the introduction of new
microprocessor architectures underscores the importance of open-source
operating systems, such as Linux and BSD. Microprocessor manufacturers
can now release a new design without having to work with a proprietary operating
system maker to port to the new architecture. All that is needed to
port the OS is a C compiler and a small amount of machine or assembly
code for the new platform, then you compile the kernel for the new
system. The new kernel will most likely require some special tweaks
for the new platform, but this is still much easier than starting
from scratch or waiting for a proprietary OS developer to port their OS, or be
delayed by having to work closely with the proprietary OS developer.
Next, you compile all the open-source utilities and applications,
and you have a full-featured computing environment. This methodology
is not unique to open-source, it was the underlying principle in
the development of Unix. Of course, it is a bit difficult to port
Unix if you don't have the source code and license to use it,
so open-source is the next logical step in the evolution of a
ubiquitous computing environment.
So it appears that the open-source movement may be headed for a
productive symbiotic relationship with hardware platform developers,
allowing hardware developers to greatly shorten their product cycles.
Organizations such as the Open Source Development Labs (OSDL) promise
to play an important role in this process. Indeed, IBM announced they
plan to port Linux to the new Cell processor.
November 04, 2004
Software is the differentiator
We now live in an era of inexpensive yet powerful computer hardware. Anyone can walk into a store and purchase a $700 personal computer with more processing power than most mainframes of 20 years ago. We can even get a full-featured, multi-tasking operating system for free (Linux) to install on our inexpensive computers. There are several viable CPU architectures to chose from: AMD, Intel, IBM/Apple and Transmeta are available at the consumer end of the market. In the server arena some additional choices are available. Computer hardware is a competitive business, the trend being increased performance at lower prices.
Of course, the most powerful computer is worthless without an operating system and some type of application software. Not all operating systems and applications are created equal. The "survival of the fit" principle is very evident in the software world. This is also true with hardware, where some systems are more reliable, faster or offer a better price/performance ratio than others. Regardless, virtually all computers possess the same functional characteristics: they store and process data. Software controls
the type of data processed and the manner in which it is processed. Good software does this in ways that is useful and expedient for the person using it.
Given that people now have a lot of choice in hardware, the deciding factor in a
computer system purchase shifts almost entirely to the software available for a particular computer. The rule of thumb for buying a computer has been to first find the computers that will run the programs you need, then consider any other characteristics that may influence your choice. In the server world, the operating system and development tools available are usually the primary considerations.
So where does this leave us? To summarize, in a world of commodity hardware, software is the differentiator.