Software

Doomsday

This was the week Google surprised the world with Chrome, its own open source web browser. Just imagine the deadly effect that had on a dozen or more browser-specific start-ups in Silicon Valley. Lots of readers are wondering what I think of Chrome, like my opinion really matters. Chrome is okay -- faster, but not faster enough to make me change for that reason alone. It's better than IE and almost better than Firefox except there are no plug-ins to speak of. What I really wonder, though, is why Google bothered to do a browser at all? Now I know.

It's not like there aren't enough web browsers in the world. There are plenty. And though Internet Explorer still dominates the Windows market, Firefox (not to mention Opera, Safari, etc.) is there to keep Microsoft honest,. So why did Google even bother? There are two general opinions on this and they are not mutually exclusive. Naturally one opinion is widely held and the other is held mainly by me.

The first reason why Google had to do its own browser comes courtesy of my friend Dave:

"People are looking at Google Chrome and actually think Google is competing in the so-called Browser Wars," said Dave. "This is not the case at all. Google doesn't care what happens to Chrome. And, in fact would be absolutely thrilled if Firefox and Opera enhanced their browsers to the point where they trounce Chrome into extinction. Google doesn't make a dime off of Chrome. Its money comes from people using the web browser -- any browser.

"What Google does not want is Microsoft creating a browser that sucks. Actually, Google doesn't mind if Microsoft's browser sucks. What they really don't want is Microsoft to make a browser that sucks and everyone ends up using it. And, if the IE8 beta shows us anything, making a really sucky web browser is Microsoft's true ambition.

"Google's main concern is quite simple: Browsers should render pages accurately, and the JavaScript engine in the browser should be fast, efficient, and bug free. On both counts, IE8 is an abomination. JScript just doesn't behave very well and is buggy. And, IE's page-rendering engine simply does not follow the standard. Because of this, Google has to keep development on their Google Applications quite generic and simply cannot implement the features they want. You'll also notice that Microsoft recently has been putting on some very compelling web content that is only available if you use Windows and IE."

Now back to Bob. Everything Dave says makes sense and I agree with it, but it doesn't answer my real question, which is not "Why did Google have to do a browser?" but rather, "What made it impossible for Google NOT to do a browser?"

The answer to this latter question begins with Dave noticing Microsoft's recent IE- and Windows-specific web content, which cracks open the door on Google's greatest fear -- that Microsoft will turn off ads in IE.

Microsoft can't do that, can they?

Microsoft can do pretty much whatever it wants in this area. There is plenty of browser competition. They can hobble their own product if they like, though it would drive users away from IE -- from a product that brings Microsoft no direct revenue anyway -- so what's the risk?

Microsoft turns off the ads in IE and what happens? Google takes a huge revenue hit, is knocked down three pegs in the eyes of Wall Street, while pretty much nothing happens to Microsoft, which would have just shown the world who is still the sheriff.

I am not saying this is going to happen, but I AM saying that it COULD happen -- and that very remote possibility is, by itself, enough to make Google have to produce its own browser.

Let me be clear that there doesn't have to be any subterfuge here on Microsoft's part. They can simply turn off the ads in IE, declaring it a non-commercial product. If you don't like it, get another browser -- there are plenty to choose from. Microsoft's revenue would go almost unchanged while Google's would plummet, if only for a few weeks or months -- just long enough for Microsoft to come through with a second punch, that is if they have thought that far ahead.

If you are wondering whether people really sit around Google asking if Microsoft would actually do something like this, well they do.

So to avoid that eventuality (and to do all the other things that Dave said, above) here we have Chrome, Google's attempt to direct the future of browser development and take some momentum away from IE.

Chrome promotes WebKit rendering, which is also done in Safari. It would not surprise me if WebKit didn't make some inroads shortly with Firefox and Opera, helping somewhat to turn the tide away from IE. Yet WebKit will change, too, by adopting Google's V8 JavaScript engine, replacing JavaScriptCore in both WebKit and Safari. Thus all the open source browsers (and Safari) become better and more alike, which helps them against IE.

A rising tide floats all (open source) ships. Google needs open source browsers to become even more competitive with IE, hence Chrome is a reference design that Google knows will work brilliantly with all Google Apps.

So much for Chrome: Now for something REALLY scary. I've been hearing that peer-to-peer file sharing has declined a bit. Actually, it's the rate of growth that has declined, but in a market where volume is always rising and prices always falling, even a decline in growth can be significant. This is happening for lots of reasons (market saturation, summer vacation, etc.) but the effect appears to be real, much to the relief of the RIAA and MPAA, which hate people sharing music, TV shows, and movies that they see as violating the intellectual property rights of their members.

But I think something else is actually happening. People are just finding new ways to share files -- ways that are harder to detect and even more chilling for society to prohibit.

Look at where P2P came from in the first place. The idea behind BitTorrent and similar programs was that many people wanted the same content and few users could afford the bandwidth to run their own dedicated servers, so sharing files by caching and re-serving small pieces of files was very efficient, especially with flat-rate bandwidth. Depending on your point of view, P2P has been a huge success or a huge pain in the ass.

But all the while, the cost of Internet bandwidth has come down A LOT. Remember P2P was born in the 1990s when most users still had dial-up connections. With the cost of Internet backbone bandwidth dropping 50 percent per year for the last decade or more, the economics have changed dramatically and it has become reasonable to effectively have your own server. No, I'm not talking about YouTube, I'm talking about dedicated servers used in large part to distribute movies and music. I'm talking about any of a number of Internet backup services.

The poster child for this new kind of service is RapidShare, a German file-sharing service that will let you distribute files up to 200 megs each for free and up to two gigs for not much money -- 55 Euros per year -- with no limit on the total number of files, total storage, total downloads or even total simultaneous downloads. Rip your copy of The Dark Knight, store it on RapidShare, then send the download URL to anyone you like or simply post it somewhere on the web. It's not as efficient as P2P, but it sure is easier AND harder to detect since nothing but http is used.

Can you see where I am going with this? How are the MPAA and the RIAA likely to respond if this technique becomes really popular? They are going to want to spy on us more, even to the point of auditing (or attempting to audit) our network backups. More lawsuits, more grandmothers and little kids being sued, less privacy.

I'm sure the RIAA and MPAA will fail in the long run. Once custom protocols and ports are dropped and you can't tell the difference between a spreadsheet and I Am Curious (Yellow) the game is up. But we're still years -- and a lot of pain -- away from that.

links for September 4, 2008

• CheckInstall - Community Ubuntu Documentation
• macosxhints.com - Convert .dmg files to .iso files for burning elsewhere

links for September 3, 2008

• Ubuntu Unleashed: Howto Integrate Windows XP Desktop into Ubuntu Gutsy and Hardy Heron!
• Ubuntu Unleashed: Howto: Install VirtualBox in Ubuntu Hardy Heron with USB Support in 5 easy Steps!

How Much is Enough?

While to regular readers this may seem an odd time of the week to see a new column from me, get used to it, because I'm deliberately increasing the frequency of I, Cringely columns to something greater than one per week yet still possibly less than two. In part this is my response to having more than ever to say. It's also an attempt to create more opportunities for you to view the ads we don't run. But the one thing this IS NOT is a knee-jerk response to the fact that Hurricane Hanna is right now bearing down on my home in Charleston, South Carolina, determined to drown us sometime on Friday. Even without Hanna to inspire me, you'd still be reading this column today.

What has me riled up earlier in the week than usual is Comcast's decision to limit its customers, which might include me (more on that later), to no more than 250 gigabytes of total bandwidth per month. Other pundits have called it "the end of the open Internet" and a betrayal. But I'm not so sure.

When I started writing this column in 1997 my Internet connection was a business DSL line from Covad rated at 384 kilobits per second for both uploads and downloads. And in the fine print of that agreement 11 years ago the account had a maximum download limit of 3 gigabytes per month. I know this for sure because I violated the limit several times and was penalized for it by Covad, which charged something like $4 for every gigabyte in excess of the limit.

So was there ever a truly wide-open Internet? Not for me.

Of course in 1997 dial-up was still the norm and V.90 modems were the bomb at 50-60 kilobits per second if line conditions allowed. If you take that nominal 56 kbps and multiply it out, what we were buying was about 18 gigabytes of DOWNLOAD POTENTIAL per month. Nobody could use all of that, of course. Or rather nobody could use exactly that amount because you'd have to take a gulp of digital air from time to time, pumping white space into the pipe. In fact the maximal duty cycle expected of even the most strenuously exercised dial-up connection back then was about 15 percent, which worked out to around 2.7 gigabytes per month -- remarkably close to the 3-gigabyte limit I had at Covad.

So what I was buying from Covad back then was a faster pipe, sure, but in a sense not a larger one. We were, after all, mining diamonds of data, not coal.

My 384 kbps could have technically allowed me to pump up to 124 gigabytes per month if I could have figured how to do that at the time using my 386-90 PC.

So how does this relate to what Comcast -- America's largest broadband ISP -- is trying to do now by limiting its customers to a total upload and download of 250 gigabytes of data per month? The new number is more than 80 times my old limit but only twice my old pumping potential. The new limit would be comparable, then, if my 2008 Internet connection was 80 times as fast as my 1997 Internet connection.

Is it? That's hard to say.

I'm a Comcast customer and know how fast my Internet connection is, but if I were just a prospective customer, all I'd get from the Comcast web site is that their cable modem is "4X faster than DSL." Where Comcast used to give you a number like 3, 4, or even 5 megabytes per second, now all they claim to be is four times faster than DSL.

Yeah, but WHAT DSL? Which version?

I think they must mean the old 1.5M/128K DSL of days of yore, and which pretty much nobody has today.

To be 30 times faster than my old Covad DSL my current Comcast connection would have to run at 11.5 megabits-per-second, but it doesn't. My Comcast Small Business Internet link runs at 8 megabits down and 2 megabits up for which I pay what seems to me to be too much money, though I do get five static IP addresses in the mix. Most Comcast residential customers get about 5 megabits down and 1 megabit up, but they pay less than I do.

Now we probably have enough data to come to some conclusions. That 250 gigabytes per month is a lot of data (Comcast claims it is enough for 99 percent of its customers), but it isn't proportionally as much as I got from Covad back in 1997 (for just about the same price, I might add). That means this 250-gigabyte limit actually IS a limit and Comcast is not being as generous as it could be.

Still, there are many unanswered questions here, and my discussions with Comcast suggest they don't yet have all the answers, either. For one, as a Comcast Small Business customer, am I even held to the 250-gigabyte residential Internet limit? Comcast doesn't know. The bandwidth limit is for consolidated uploads and downloads, but does it include Comcast extra-cost services like VoIP phone service? Comcast doesn't know.

This feels to me like a trial balloon. Eventually I'm sure we'll learn that Comcast phone service and Comcast Video-On-Demand movie downloads are excluded from the limit. Go over your limit renting from NetFlix and you'll be punished, but not if you are renting from Comcast.

It's all about TV and movies, you know, and what if cable TV as we know it dies and is replaced entirely by digital downloads?

If that's the case we can make a similar calculation coming from a different direction. How much bandwidth capability would we need if cable did die? Say the average house has two TVs, one is running four hours per day and the other two hours. Say HD downloads require 2 megabits per second, which is actually quite an aggressive number assuming H.264 compression. How much bandwidth would those two TVs suck up in a month? It looks to me like 162 gigabytes, which is within the 250-gigabyte maximum, but just barely.

What Comcast has done here is draw a line in the sand that it thinks it can justify. They are trying to change the game with DSL from one of comparing download speeds. And while Comcast seems to allow its customers just barely enough bandwidth to survive in the radically changed environment of a no-cable world, that's only the case if peer-to-peer file transfer technologies such as BitTorrent aren't used.

Comcast doesn't say so, but these new rules effectively kill P2P in the long run, that is unless Comcast somehow sanctions that P2P, calling it Comcast P2P.

And THAT's what this is really all about.

iPhone Gmail behavior

I noticed an interesting behavior in the iPhone's Gmail implementation.  When I delete a message from the inbox, I expected this to behave the same as it does with IMAP, where deleting a message from the inbox just Archives the message.

But in the case of the iPhone, deleting a message actually deletes the message, (moves it to Trash on gmail).  So if you want to just archive a message you need to move it to the "All Mail" folder.

It looks like this has been documented.