And gases can't "want" to fill their containers, but when talking about the observed properties of gases, we anthropomorphize by describing the tendency as a "want." It's a rhetorical convention, not a conjecture that a gas is self-determining.

Google Docs has a lot of nice features. It's true that compared to OOo or MSOffice, it's under-featured, but then again, it has a large number of very cool collaborative and Web features you don't get with either of those. It's a trade-off.

But again, that's not the issue to which I replied, and I still contend that the original poster was wrong in his analysis of Google Docs as a piece of software.

You're engaging in a straw-man (these guys over here are wrong, so you are too), and yet you're using a terrible example because the people who understand the real estate market were years ahead of the drop. I can't count the number of very well written essays, articles and books I read about how the housing market works that included the prediction that the current bubble would burst within 5-10 years of 2005. Lo, they were right.

I bought a house at the height of the market, and what seemed like dozens of my friends asked why I would do so when it was so obvious that the market was at its peek. My answer and the answer of the real estate community as a whole was that three factors increased the real estate market from where it was in 1900 to where it was in 2005: 1) single-income families have very broadly become multi-income families 2) far fewer extended families live under the same roof and 3) the average family is far more mobile than it was in 1900.

These three factors increased both the demand and the available cash for purchases. At the same time, we became capable of managing long-term debt in ways that were not nearly as ad hoc.

So, has the bubble burst? Yep, just as it did in the late 80s.

Does that matter? Not really. The price will eventually return to where it was, though at a slower pace. What the doom sayers are suggesting is that the fains of the 20th century will also erode, and we'll return to $50,000 homes. That, simply can't happen unless the market changes pretty radically. Real market forces forced those increases, and unless we return to the model of single-wage, extended families, there's just no reason that people would want to sell their homes for so little, nor why people would stop paying the current rates.

Eh?

Heh. You don't now Big Oil very well, do you? This is an industry whose wealth is entirely based on expanding resources that they won't be able to take advantage of for 5-10 years. They spend more on exploration and R&D for long-term returns than any other industry.

You've contradicted your own point. The person you're responding to said that change in pricing would be gradual as supply waned and market pressures would facilitate a rational response. You're setting a timeline of 10-40 years, essentially backing up his thesis.

Ignoring the issues of pre-warmed applications, I'll re-state what I said above:

What Office apps are you using? I'm using Open Office and I just opened the spreadsheet app. it took exactly 11sec to open and present a blank spreadsheet.

On the other hand loading a 2-page long existing document in Google Docs just took 2 seconds (that's with a trans-national proxy through my company's gateway in the middle) in a browser that had not previously visited Google Docs (and thus had no cached JavaScript, etc.)

My experience with MS Office is that it's faster than OOo, but slower than Google Docs.

However, both MS Office and OOo speed up significantly once you've already loaded them once on most platforms. Why? Because they stay resident, taking up system resources. You can do the same thing in your browser with Google Docs. Just keep a tab open with Google Docs and all of your documents will come up faster.

The real bottom line isn't a matter of benchmarks, however, it's that the original poster's claim that Google Docs was "bloatware" ignores the fact that it's an implementation of a very large system which is at least as bloated in every fully-featured implementation.

Java was essentially dead on arrival. It had several bursts of potential success, but ultimately, it's just become the less-horrible COBOL of the 1990s and 2000s.

That's not to say I dislike Java. It's not that bad as languages go, but it failed to capture a market outside of the world of highly captive environments.

COBOL was the world's most popular programming language for a long time. Thing is, a language whose use constrained to black-box environments isn't really going to be considered "alive," in the larger sense of programming language vitality because nothing new builds on it.

Eclipse and Open Office are the only two exceptions to this that I can think of. One was written by the company that published Java and one was designed to be an IDE for Java. That's not my definition of broad adoption outside of captive environments.

So, JavaScript makes it bloatware? Last time I checked, Google Docs was faster loading by large factor than OpenOffice, MS Office or any of the other WP/Spreadsheet apps out there. How do you define "bloated," here?

You make it sound as if that comes with no advantages. There are dozens ... here are a few.

• Google's reliable storage which is backed up for you
• documents and be shared or collaborated on with other users
• the options for publishing to and interacting with the Web from docs is, frankly, a game-ender for locally hosted Office suites.
• The ability to continue to access your documents even if your new computer is a different OS or hardware vendor with no purchased software.

One demo of the idea of publishing data to the Web that blew me away was in Google's Official Blog about their public data sources, where they plotted a time-series of world fertility data. There's lots of decent examples on the Google Docs official blog as well.

There's also the fact that all Google applications allow you to export your data to local apps, if you wish. The Open Office format export is quite nice in Google Docs (import is OK, but at least for the spreadsheet it has a ways to go).

Only if you don't want to pay for it. Google Docs via a premium Google Apps domain does not have ads.

Ah... no. That's the reason that they're doing it, not the reason that it's popular. The reason that it's popular is that it's useful and free (again, if you don't want to pay for the ad-free version).

Most anti-spyware scanners don't give a rat's petard about applications that show ads or applications that store files remotely. Typically, the goal is to ferret out software that does either without the user's knowledge or ability to prevent. In both cases, Google Docs is 100% opt-in and entirely friendly to those who wish to opt out later on.

If your definition of spyware is any Web site that records your activity on the site or saves documents that you create for later use, then you need to include every ecommerce site on the planet. I don't think that's a definition the majority of the technical community would agree with.

Wind is free. If what you want is to have your hair ruffled, you pay nothing.

If, on the other hand, you want to build an energy grid based on wind power, it costs far more than you might imagine. The post you're responding to has some salient points in this respect.

The problem of replacing or upgrading the single most important piece of our national infrastructure has always loomed as the greatest problem with converting to energy alternatives. Wind and solar power have radically different properties with respect to the national grid, and you can't just plunk them in and go on. Doing that leads to unpleasant things like brown-outs that kill the elderly during the height of summer or depths of winter.

These aren't unsolvable problems, but they cost a LOT of money to solve, and no one is yet willing to step up and pay for it, as the advantages are not easily recognized.

I don't think the name matters. What matters is that, like C or Common LISP, the language is specified in an open (as in open to contribution) way. Sure, AT&T's Bell Labs created C (before the breakup), but it wasn't a proprietary language after its management was turned over to ANSI.

Go isn't at the stage where an international standards board makes sense, but I'm encouraged by the openness of the process.

It is a concurrent system programming language, and the fact that that terminology doesn't make any sense to you is probably evidence enough that it's doing something interesting (interesting does not equal successful... time will tell).

Because it's spurious optimization.

Modern HLLs (and even most of the languages that live in the spaces between machine language/C and HLLs) spend so much time doing bounds protection, reference counting, garbage collection, constructing, destructing, etc, that if you find yourself in need of the extra cycle indicated by a double-indirect lookup, then you need to use a lower level tool for that code section, not a 2D array. In many languages this is commonly found in image processing and large math libraries. PIL for Python, PDL for Perl, etc. are designed to perform these low-level operations in a way that the high level language can manage cleanly.

You don't need a huge tumor on the side of the language syntax just so that your reference counted, garbage collected, safe accessed arrays are stored without an extra pointer. That would be absurd.

You kids and your Pascal and COBOL.

Get off my lawn!

*returns to toggling in bootloader code for reading the OS off paper tape*

You're making a very old argument. One that was made by C purists when C++ was introduced: that simply adding syntax for something doesn't make your language special. C++ users disagreed and many defected from the C camp. Were they right? Possibly.

Of course, the smalltalk users were also right: these were not new features. C++ was just applying previously developed tools to the C language.

None of that mattered in the long term.

Is Go the new C++? I have no idea, but I don't think you're taking an objective, historical view, here.

As with many other Google efforts, Go isn't managed exclusively by Google (much like Android). So your statement really doesn't make much sense. There are many people who do feel that Go is the correct logical step from many existing languages. I'm on the fence, but as a disinterested observer, I think its reasonable to say that the community of folks who are anticipating that future are largely not Google employees.

They're far, far more portable than Java.

Java is essentially non-portable by design. It has only one environment where it runs: the JVM. Unfortunately, the JVM is only ported to new platforms in the loosest terms, though to be fair, they've done a large amount of work to make it more platform friendly under Windows, but it's still got a long way to go after more than a decade.

Python and Perl, on the other hand (along with other HLLs that don't use a platform-abstracting VM) are far more portable, interacting with system libraries and platform features in ways that make sense without having to re-specify the language to do so.