Doc Searls and
Brent Simmons have penned an article discussing Mac OS X from a Linux
perspective. Combining the Macintosh and Unix ways of doing things is
a complex endeavor, to be sure, and the results irritate both sides in small ways,
but overall, OS X is an amazing achievement. It’s harder to configure than a
Mac, and the Finder isn’t as polished, but it’s stabler and more responsive.
(Unix people have their own complaints, I’m sure, but they can just buy one of those
dirt-cheap Wintel boxes and put Linux on it. Unless the great styling and industrial
design of Macintoshes really is a draw.)
#
Scientology manipulates Google
I’ll admit that I don’t pay much attention to the battle between the Church
of Scientology and the people who claim that it’s a cult and a scam. I’ve
read enough of the material on the subject to be highly skeptical of
Scientology’s claims (it’s remarkably similar to Dogbert’s theory that all
of life’s problems are caused by invisible people named Juan and Cindy),
but I generally avoid participating in the conflict itself.
However, it seems that Google has become a battleground.
Recently, Google removed anti-Scientology
site Operation Clambake from its index. There’s been
much discussion about why that happened,
but according to Google, there was a claim made under the
DMCA. Apparently,
someone in the Scientology camp claimed that Operation Clambake was violating
their copyright and that Google’s cache of those pages (for searching purposes)
made Google also liable.
That’s a pretty weak argument, but it’s not too hard to understand Google’s
reluctance to face court time (even given Scientology’s apparently-poor record
with these things).
It’s also not censorship, as several people are shouting. Google is a private
service and is free to index or not index any site it chooses. It stays impartial
because it makes its service more useful, not because it’s required by law. Google
has enormous power on the web at present, because it’s such an effective search
engine, but it’s not the first search engine to dominate the field and it may
not be the last, either.
Operation Clambake doesn’t provide permanent addresses for its
news posts, but Operating Thetan, which pointed out Scientology’s attempts
to boost its Google rankings, is collecting news
and links. (via Wesley Felter)
#
How RESTful is the web
What is the essential architecture of the World Wide Web? Roy Fielding examines
the issue in his doctoral
dissertation, and declares it to be “Representational State Transfer”
(REST). Fortunately,
he then goes on to explain what that is (or, perhaps you already know what that
means at that point; I read the dissertation somewhat out of order).
Describing how the web is different from other distributed systems is difficult
without assuming some knowledge of networked software, but I’ll give it a shot.
There’s plenty of articles and the dissertation itself for those who want more detail.
REST sees the web as
a vast collection of “resources”, identified by
URIs. Each resource accepts
a limited number of commands, such as GET, PUT, and POST. The exact meaning of
each command depends on the resource, but GET always returns a representation
of the resource. For example, to read ZedneWeb your browser performs GET on
“http://www.eyrie.org/~zednenem/” and receives an
HTML file representing
ZedneWeb’s table of contents and most recent weblog entries. PUT is used to
set the state of the resource, but isn’t widely supported by browsers. POST
sends the data to the resource, which can use it for just about anything.
That probably doesn’t sound very impressive, so here’s a very simple example
of how the REST model
can be used. There are a number of proposals floating around for web-based applications
which are accessed through Remote Procedure Calls. Let’s say a company wants to use
such an application to store employee phone numbers. In the
RPC model, there might be a resource
called “http://example.com/employeeDatabase”. To get someone’s phone number, you
would send a POST to that resource with data meaning “get the phone number of
employee X”. The response would contain the requested phone number.
The REST model
looks at things differently. We’re interested in dealing with employee phone numbers,
so we would assign an address to each one, such as
“http://example.com/employee/X/phoneNumber”. Then, to get the phone number,
we simply do a GET. If we want to change it, we do a PUT. (We would want to take steps
to ensure that only the right people can access or change the data, but it turns
out that HTTP already has
that, too.)
The REST example isn’t
necessarily any simpler than the RPC one—you
still need to know where to get the information and how to interpret the result—but it
does have advantages. By working with HTTP’s
methods, it can take advantage of its security and caching architectures rather
than having to recreate them. By using the same protocols as the rest of the web,
software which has to use multiple services only needs to know one protocol instead
of several.
This is fairly abstract stuff, so don’t feel bad if the advantages or even the
differences aren’t clear. None of this is user-level, so unless you’re writing
software that deals with web services, it probably doesn’t matter how things
are implemented.
(Backtrace: the dissertation was cited by an article comparing
REST- and
RPC-style web services which was
mentioned in an examination of
SOAP security referenced in the
March Crypto-Gram.
This list of
REST
resources is a good place to start for further reading.)
#
The Jaguar of laptops
Good news and bad news relating to my PowerBook
problems. Last week, I called in again, spoke to another customer
support rep, and learned that the logic board was on back order and wouldn’t arrive
until March 20. Since that would be over a month in repair before the repair
even started, I complained and was put in contact with customer relations.
They felt that I had waited too long already, and were able to get the logic board
replaced and the machine sent back to me by that Friday, March 15.
If things have been quiet on ZedneWeb since then, it’s because I’ve been playing
with the new machine and Mac OS X.
But all is not well in TiBook-land. The AC
adaptor which comes with the laptop has a light which is supposed to glow green or
amber when it’s plugged in, depending on whether the battery is charging. Since
getting it back from repair, mine does not light up at all.
Merely annoying, really, but I called in and they sent me a replacement adaptor
which arrived on March 21. Guess what? Its light doesn’t work either.
Meanwhile, the laptop can’t seem to recharge its battery. That is, the battery
indicator in Mac OS X claims that the battery is completely charged,
while the battery’s built-in tester claims that it’s empty.
So I called up tech support again. The guy I spoke to suspects, as I
do, that the two problems are related, and they indicate either a problem with the
battery or… the logic board. The new logic board that they put in to replace
the old one, which had some sort of problem with the video control. So they’re sending
me a new battery, in case that’s the problem. But if that doesn’t solve things, it’s
time to send it off again.
Of course, this repair happens after I discovered how easy it is to set
up local file-sharing and start transferring stuff over. Not only is the TiBook my
primary-use computer, it also holds the only copies of certain files. And since
Apple explicitly warns that data may be lost during repair,
that means I need to start making backups. (Yeah, I know. I need to be doing that
anyway.)
On the positive side, the people I’ve spoken with at Apple have been intelligent
and helpful. I just wish I didn’t need to interact with them so often.
#