It's Like Déjà Vu All Over Again
"You could probably waste an entire day on the preceding links alone. But why take chances? We also give you Paul Snively..." — John Wiseman, lemonodor
David Kurtz: "When California was faced with an energy crisis last year, Vice President Dick Cheney asserted that 'conservation may be a sign of personal virtue, but it is not a sufficient basis for a sound, comprehensive energy policy.' And so what did Californians do to stave off rolling blackouts and rising energy costs? They conserved." True. [Scripting News]
Re-read the Cheney quote: all he said was that conservation wasn't "sufficient." I find it hard to believe that a reasonable person wouldn't agree with that. Put another way, conservation may have been a necessary, but not sufficient, tactic in addressing California's energy crunch. And while Californians did indeed conserve and can and should be justifiably proud of it, there's no question that there's more to California's recovery from the crisis than conservation alone. And this is all I think Cheney meant.
That doesn't mean that I think we should let the Bush administration off the hook over Enron and their ilk, however.
11:29:00 PM Google It!
Hmm. The court should not drill for oil either, but it should rule on anti-trust cases in the oil industry. It should not make telephone switches, but it should rule on anti-truct cases in the telecommunications industry.
The issue as I understand it is not about software design, it is about products, markets, and competition. I don't know nearly enough about the law to saw who is right. But the court has ruled against Microsoft. They lost more than once by the experts we've entrusted to make such decisions. There is nothing special about the software industry that is above the law, is there?
The issue, as I see it, is that software exists at a sufficiently high level of abstraction such that a software company can effectively be in multiple markets at once. Not only that, but software can be reflective and introspective. What market, for example, is software about markets in?
In the end, a court ruling that Microsoft can't write software that does X isn't very dissimilar from a court ruling that no one can write software that doesn't do Y, e.g. where Y = "include mandatory anti-piracy technology." Both are attempts to reduce the generality of the systems in question; both are being pursued at the behest of large competitors with deep pockets; both have extremely questionable long-term ramifications for customers.
11:22:18 PM Google It!
OOP Criticism. OOP criticism and OOP problems
The emperor has no clothes!
Reality Check 101
Snake OOil - Interesting anti OO rant. [Roland Tanglao's Weblog]
It's a shame that the author's style is "proof by assertion," because buried in all of the procedural hyperbole are some good points, e.g. that C and C++ are too low-level to serve as effective examples of much, and that some things that people have come to associate with OOP, such as garbage collection, have nothing to do with OOP in the first place.
But good grief—the rest of the piece is as pedantic and provocative as anything that it rails against. And in the final analysis, it's just plain true that, to most working programmers, OOP is a novel discipline compared to the procedural approach, and so of course some hard lessons, such as the hidden costs of implementation inheritance, are only now becoming widely appreciated.
But this is all academic: there are some amazing multiparadigm languages these days. Languages that I consider extremely important—by which I mean that I feel they advance the state of the art—include, at a minimum, Oz, Objective Caml, and E. Erlang might also belong in the list, but I haven't payed as much attention to it as the others.
If you only learn one new language in the rest of your life, learn Oz. If you learn two, learn Oz and E.
11:11:07 PM
Documents, reports, books, articles, vector graphics, multimedia interactive animations, 3d world descriptors, math formulas, chemical formulas are all examples of semi-structured data.
Genetic footprints, topic maps, invoices, stock quotes are all examples of fully structured data, even if very well described by XML markup.
Don't let the syntax fool you: it's the schema that counts. If the schema describes a fixed set of elements, not matter how complex, you should use relational mapping.
Otherwise, but this is much less common that it first appears, and only otherwise you should entering the native XML DB land. When you hit the wall on native XML DB and somebody shows you how faster, simpler and more elegant would be to use a relational mapping for it, don't tell me I didn't warn you.
Well, this is indeed interesting, but not necessarily very informative. What exactly is an "XML-native" database? What exactly is a "relational" database? The reason I ask is that popular "relational" databases such as Oracle, DB2, and the open-source PostgreSQL all have object-oriented features in addition to their relational features. Some of them even have some level of XML support baked in. Are they relational, object-oriented, or XML-native?
What about layers on top of such databases that implement whole new APIs for insert, update, and query, but run against one of the above-mentioned traditional databases? For example, both at work and at home I've been playing with RDFSuite, which rides on top of PostgreSQL. Does that mean I'm using an "RDF-native" database, given that all of my interaction from that point forward—from schema design to query—is in RDF?
Like most absolute statements, "if the schema describes a fixed set of elements, no matter how complex, you should use a relational mapping" is extremely questionable. I'd say more explicit things about this, but I don't know that my employer would care for me to. Suffice it to say that we currently use straight relational modeling with a really bad Java O/R mapping tool, it's been acknowledged that our data doesn't exhibit as much flexibility in analysis as we want, and we're investigating alternatives. So far, colleagues that I've shown RDFSuite examples to, using a very preliminary RDF schema and handful of sample data, have been pretty impressed—a couple of days' work was sufficient to allow me to express single-line RQL queries that couldn't be expressed at all in SQL against our existing schemata.
10:35:52 PM Google It!
Inflation calculatorThe Columbia Journalism Review .... Inflation calculator The Columbia Journalism Review has put up an inflation calculator: pick a year (1800-2002), enter a dollar amount, pick a target year (1995-2002) and it'll tell you what the buying power of your dollar figure is in target-year-dollars, adjusted for inflation. Pat points out the value of this for writers -- if I were going to attempt a work of historical fiction, I know where I'd turn...LinkDiscuss(Thanks, Pat!) posted by Cory Doctorow at 22:42 [bOing bOing]
I definitely want to keep track of this. Later, I'll link to something related that I also want to keep track of. Stay tuned!
7:57:42 AM Google It!
