LevelsOfDetail

So I've had the Curve for several days now, and it still impresses. The autotext typing feature is particularly handy, though I don't feel like I've figured out the optimal hand position for typing lots of text. I still mistype a lot of keys like n/b or s/a. Maybe it's just practice.

One thing I can't figure out though, is what affects the performance of the browser. All weekend, loading web pages was painfully slow. The browser would get stuck at the "Requesting" step for ages. It seemed to be a little faster in areas that had better signal, but it wasn't night and day. But now, I'm at home, and signal is hovering around one or two bars and surfing is much snappier. Is it the some other part of the AT&T network that is causing the slowdown?

To my horror, I discovered that the last three "shoots" on my photo were all inadvertently taken with a ISO 800 setting.

These camera makers really need to tweak the interface.

Back in the film days, you put one roll of film in. That roll had a constant ISO. You know what film you put in, so you remember what ISO you're using.

These days, it's just another setting.. but unlike any of the other settings, it's not displayed anywhere. Furthermore, LCD previews of your shots don't show that you've accidentally set the ISO too high, since the scaling down usually averages out the noise.

They really need to put the ISO setting in the view finder, or at least in the smaller LCD at the top of the camera.

Maybe the lesson is just to use the auto-ISO setting.

Sorry, lame heading, I know. I'm just too lazy to think of anything better.

A couple articles showed up on my radar that I wanted to mention.

First up: Divining from the Entrails of Ubuntu's Gutsy Gibbon by Bruce Byfield.

The thing I noticed about this article is that while he expresses many complaints, a vast majority of them are cosmetic. Improper words like "guided" instead of "automatic", or how synaptic is not as good as apt-get, or how the font selection and default setting sucks (see my hint to do something about that), or how icons on the desktop are misplaced, etc. etc. The extremely promising thing about this article is that he doesn't mention hardware compatibility issues!

The only area that Linux has been stuck in the mud is hardware support. It simply can't be improved without hardware vendor support. To see an distro review that doesn't seem to say 'Oh my mouse didn't work, or my video didn't work' is a sign of great progress on this front.

Bruce needs to realize that while Gutsy's default settings might not be best for him, he can actually fix them. And open source software doesn't quite work like commercial software where companies sometimes depend on reviewers to tell them what's wrong with their software. If he thinks things can be improved (for example, his suggestion to use the term "automatic partitioning" instead of "guided partitioning" is a good one), then there is a direct mechanism for him to report the problem: Ubuntu's bug tracking system.

All the problems this guy mentions can be fixed. And most of them aren't that hard to fix either. I haven't come away from a Linux review article thinking something like that in a long time.

Second up: 7 reasons why Linux won't succeed on the desktop by Alexander Wolfe on iTnews.

This guy is much more in tune with the problems of Linux today. I found myself silently agreeing with many of his assertions, the biggest one being that people haven't figured out a way to make money from the Linux desktop. As much as people deny it, it takes money to move these things forward. Most hackers who spend a lot of time on Linux code aren't exactly doing it for free. They've managed to find a company that knows how to make money from Linux. It's just that there are very few companies that actually try to make money only from the desktop, and there are even fewer that remotely succeed at doing this.

Part of the reason, at least from my view, is something I've also mentioned before. It's really hard for software vendors to ship software on Linux. As it currently stands, the only technically "sane" solution is for a vendor to open source their software in a way that a distro can include it in their release. The first annoying alternative is to ship your entire runtime with your app (but this is hard due to LGPL static linking situation). The second even more annoying alternative is to make a bunch of packages for different distros (which distros?)

Meanwhile M$ spends a whole bunch of time and effort and money to make this easier. They work really hard at not breaking backwards compatibility (well except for Vista), and they give you a lot of tools that make app deployment really easy. You can write one windows app and expect it to work on a huge percentage of the installed base. That makes it really easy and enticing for people to write software for the platform.

The problem isn't that Linux simply can't support third party software vendors, it's just that nobody has figured out how to do it right. So far the biggest successes look like QT-based solutions, or web apps (if you think about it, the browser has become a fairly widely deployed backwards compatible app environment). Seems like Adobe's runtime and Silverlight are also interesting prospects.

Chances are though, when the OSS community does figure it out, it's not going to take the same form of MS's solution (even if it did, it likely wouldn't be better enough to displace MS).

Every once in a while, I type something like "Linux vs Windows" into google, in a foolish attempt to assess the general landscape (and often find outdated articles)

This time, I came across this article. The author argues that the choice provided by open source software results in a usability nightmare for people who don't care for the choice.

The one paragraph that really rang true:

The tradeoffs made by the open source community between usability vs. choice will become increasingly important as various distributions and organizations (such as SuSe and Lindows) try to move Linux into the general desktop market space. It's extremely important for the open source community to be responsive not only to users' freedom to choose, but also to users' freedom not to have to choose.

With bazillions of distros, and bazillions of alternative programs that all try do the same thing, a Linux user has to spend a lot of time finding the right combination of tools to build exactly the experience he or she wants. And it's true that no average user is going to do this. It just doesn't matter.

Distro's like Ubuntu will come along and try to collect all the best-of-breed selections. But even users of Ubuntu know, not everything is quite up to snuff. Best-of-breed doesn't really mean anything if the the user still can't do what they need to (connect to an exchange server, for example).

Choice is good when it lets you choose a better solution over one that already works. Choice sucks when you have to choose between a bunch of halfway solutoins that all don't work in some way, and you wonder why all these people don't cooperate together to at least make a baseline that works.

Just came across Ubuntu unleashed which appears to be a Ubuntu tips site similar to my ubuntutips.net project. And damn, he has a 100,000 hits already..

Oh well. It's all for free, so there's no reason you can't have two right? Mine's gonna be better anyways ;-)

That thinks that linux font rendering (even with the latest David Turner patches) still is not up to par with windows... This post goes into much more detail, specifically pointing out that Linux text rendering isn't doing any gamma correct rendering at all. Apparently this is why the diagonals in Linux-rendered characters look so much heavier than the verticals or horizontals?

I saw that he hates how 'w' look ons Linux too.

This program continues to impress me. It seems like magic.

Apparently there's a new version. Looks mostly the same, except the UI has an Vista bluish-ness to it. Doesn't really fit in on XP.

I get really annoyed by random articles/blog posts written by Linux/Ubuntu evangelizers that all talk about how they had to use their friend's Windows laptop and it was slow and had all this crapware, etc.

What they forget, is that if it actually became easy for 3rd parties to ship software on Linux, then you'd see a whole bunch of crapware floating around on Linux as well. At some level, it is up to the user to be discerning about the software that gets loaded onto the machine. It's unfair to blame it on Windows. If you look at the industry as a whole, the benefits of Windows providing such a good software distribution platform have far outweighed the costs.

Of course, it's not really accurate to compare Windows and Linux as equals. For example, Linux distros come with more software in the default install. The distributors can usually test this software, and ensure that it isn't crap. So you can make the argument that if you install Ubuntu, then you're less likely to need to install some external piece of software, and therefore, you're less likely to install crap on your system.

Then there's also the fact that it's really hard to ship any software (let alone crap) on a Linux system. My friend Ramesh argues sometimes that it is actually this "high bar" to entry that keeps the awful stuff out. That may be true, but it weakens the argument that Linux is the platform of "choice".

A seasoned power user can easily keep a Windows system running well and free of crap. I do it. A bunch of my friends do it. It'll be funny to see the day (if it ever comes) where you'll have to go clean the crapware out of your relative's Ubuntu box.

And a final note. Microsoft could fight back by trying to integrate more useful software directly into the default install.. but we know what happened the last time they tried to do that. Apple gets away with it for now, since they're the underdog, but sooner or later, people are going to start complaining about them too.