Internet Alchemy

The Shafe Report on P2P Computing.
A good survey of the state of P2P that tries to identify where the
gold is for new startups:

It is often said in computer circles that the people most likely to
make money during a gold rush are the ones selling pickaxes and
shovels (as an aside a recent report on small scale mining points out
that miners actually spend most of their money on gambling, alcohol
and prostitutes).

The point is made to justify selling tools to the companies searching
for the gold. However, when the technology is defined by a standard or
is provided by the industry giants then start-up companies need to
focus on either filling any holes in the standard technology or
providing vertical solutions based on the standard technology.

Start-up companies can offer products or service that solves a
critical problem for decentralized applications, like resource
discovery or reputation tracking. Novel solutions to these problems
could be licensed by larger software companies to address shortcomings
in their offerings.

Insatiable SVG.
A report by Eric van der Vlist from XML Europe on the state of SVG,
including its use on mobile devices.

SVG is invading your cell phone and this “graphical XML” might wipe
out “text XML” (such as XHTML and XSL-FO), just as graphical user
interfaces have wiped out text-based user interfaces.

Eiffel for .NET part 2.
Following up the first part of Bertrand Meyer’s Eiffel.NET series
detailing how the Eiffel.NET compiler shoehorns multiple inheritance
into the .NET single inheritance model. As I suspected it’s a complete
hack. The base classes become interfaces with separate implementation
and creation classes. The consumer of the classes has to rely on
documentation rather than language features to understand how to
instantiate one of the base classes. The method Meyer described
implements polymorphism but does not allow for mixin functionality
which I presume has to be done via aggregation.

What the common object model provides is not a stranglehold forcing
all languages to support a single view, but a kind of language bus,
enabling all languages to cooperate by agreeing on a basic set of
common mechanisms. With the multiple inheritance example, we were able
to present to the bus, and hence to other languages, a view that
doesn’t lose any essential property of the original model.

KaZaA Worm.
There’ll be a lot more of these before the year’s out.

XBox Economics.
An insightful analysis of the economics of the XBox

The economics of the XBox don’t add up now, and they get worse with
time. Sony and Nintendo can kill the XBox on cost alone. The “software
subsidies” that Microsoft expected are a myth. Game console prices
will continue to drop, from $199 to $149, then on down to $99. Will
Microsoft ever make it to the $99 level of this game? We’ll see.
According to XBox economics, it all depends on how much money they are
willing to lose.

F#:

F# is an implementation of the core of the Caml programming language
for the .NET Framework, along with cross-language extensions. The aim
is to have it work together seamlessly with C#, Visual Basic, SML.NET
and other .NET programming languages. In particular it is the first ML
language where all the types and values in an ML program can be
accessed from some significant languages (e.g. C#) in a predictable
and friendly way.

Updated RSS Microsite.
I’ve added some new aggregators and RSS tools to my RSS microsite.
Drop me a line at iand@InternetAlchemy.org if you’re missing and want
to be listed there.

W-CDMA Bugs Will Delay 3G Take-off.
More evidence that the 3G roll-outs are not going to happen in any
significant way for a couple more years.

Leading Japanese mobile vendor Matsushita has warned that ongoing
technical problems with W-CDMA technology will delay mass market 3G
take-up. According to the Financial Times, Matsushita president Yasuo
Katsura has been quoted as stating that Japanese W-CDMA glitches will
take until mid-2003 to solve. Katsura’s statement casts further doubt
on NTT DoCoMo’s ability to meet its target of signing up 1.29 million
3G subscribers by March 2003.

Java Performance and Scalability.
An in depth article about Java performance, especiially with regards
to garbage collection.

Developers must first get a solid background in garbage collection,
examine the different algorithms and implementations, come to
understand how they affect the application, and learn about the GC
problems Java applications have historically faced. Then they can
construct a mathematical model to predict and tune application
behavior.

By using the output from the -verbose:gc option in the Java Virtual
Machine (JVM) and the GC analyzer, developers can analyze object
creation, object lifetimes, and garbage collection of these objects.
Then they can use the knowledge they’ve gained to improve the
performance and scalability of their applications, using advanced GC
algorithms to reduce application pause times to acceptable levels.

Is a Cow an Animal?
An interesting discussion of covariance. An Animal class has an eat method
that takes a food as an argument. A Cow class inherits from animal and
overrides the eat method to take a vegetableFood argument. This is
covariance. The problem is when you have an Animal instance of type
Cow and you try to feed it some meat…