July 2007 Archives

July 16, 2007

A Mac in decline

When I started working at my current job and was given a Mac laptop, I quickly started feeling its gravitational pull. I gradually spent less time on my PC and used it for fewer tasks. As a development environment the unix-based Mac was far superior. For lots of smaller tasks, things like FTP or media format conversion, there always seemed to be one or two really nice Mac applications, either freeware or trialware, where for Windows there would always be an endless array of crappy, undifferentiated stuff on download.com that would track its mud all over my disk and menus during installation and never completely go away. My email still lived on my PC, but then Gmail came along. Eventually the primary uses of my PC became burning DVDs (should've gotten that Superdrive!), email still on my older account, large-scale Photoshop work for which I needed my 2 monitors, and music production.

Recently, I had a dreadful fright when I realized the pendulum had swung back a bit. I had in mind a web research project that I knew was going to require a fair number of Firefox tabs and back-and-forth comparison, and I found myself wanting to do it on my PC. Why? Because my Powerbook has, sadly, been growing less and less pleasant to use. More precisely, applications seem to be getting more sluggish and less responsive, along with other minor symptoms. Of course it's a bit out of date now, at three years old, but circuits don't get slower over time, do they? And it's not a result of the kind of software or OS buildup usually associated with Windows over that kind of time period, because the hard drive crashed and was replaced about a year ago, the second major malfunction in the computer's life. So what is going on? I've been devoting an unfortunate amount of thought lately to that question, and here are my possibilities:

1. I have become more impatient, and the slowness is only in my perception. Or conversely, I've become more of a power user, running more resource-hungry applications at once. This may well be part of the problem, but there's not much I can do about it. I'm already displeased at the amount of time I spend policing open applications and tabs to try to keep things running reasonably smoothly.

2. Insufficient hard drive space. One of the things that frustrates me most about Mac laptops is the small hard drives. Aggravating this is the fact that every time Software Update runs, there's about 250MB of updates available for the OS and various iLife applications. I always wonder if the 50MB update for iTunes is in addition to the previous size, or replacing old content. If the former, some of these applications are getting pretty bloated! Anyway, I recently cleared out some stuff to get to approximately 16GB free space out of 60GB, and it doesn't seem to have helped.

3. Insufficient memory. This is my current favorite theory, largely because it's the one thing I can easily remedy. One of the symptoms is prolonged waits (accompanied by the beach ball/pinwheel of death) while switching applications, and this is usually caused by 'thrashing', or moving all the applications' data back and forth between memory and virtual memory on disk. I currently have 512MB of memory, up from the laughable 256 when I got the computer, and I'm hoping for a company upgrade to 2GB soon.

4. Increased demands of operating system. When I got the computer it had the 10.3.something version of OS X, and now we're up to 10.4.10. It seems beyond doubt that the OS has only increased its usage of system resources in that time, particularly with features like Spotlight and Dashboard, on the justifiable assumption that most people are using it on newer and more powerful computers. I would try to disable Dashboard, but I actually find the darn thing pretty useful. So there's not much I can do about this that I know of.

I'm really hoping the memory upgrade improves the situation, because everything else about my Mac is still great, but this problem is making it increasingly painful to use.

July 19, 2007

Two Software Gripes, Pt. 1: Firefox Memory Management

For all its life, that I know of, Firefox has been subject to criticism about memory usage. For a long time I took it for granted that it had a "memory leak," and only gradually did I start to wonder why the developers would fail for so long to fix a problem so basic in an application used so widely. Then a while ago I read a Slashdot discussion on the topic and realized that, to the developers, what many consider a memory leak is really a feature: namely, the aggressive caching of recently visited pages so that "Back" button usage can display the page instantly without having to fetch it again.

I had long thought that the most conclusive evidence of a memory problem was that, when I tried to close a bunch of tabs to free up system memory, there would be no decline at all in FF's memory usage. Others apparently share this experience. But I learned from this discussion that this is because a static number of recently visited pages are cached regardless of which tabs or windows they were in. This enables the sometimes nice feature of being able to get back recently closed tabs through the History menu. It just went against my perhaps outdated expectation that once a tab was closed it was gone, and should be gone from memory too.

To me, it would be a nice option for Firefox if it could only cache the n most recently closed tabs, where n could be 0 or 1 (or any other number of course). However, there are some vague indications that the software's architecture might make this rather difficult. I've considered trying to add it myself, as they say one should do when one has a complaint about open source software. But in my limited experience looking at the code of open source software, I've been extremely confused and unable to make any logical connection between the code and the application that I see when I use it.

Another thought: one Slashdot commenter said that with the Restore Session feature, his problem was alleviated slightly, because he could restart Firefox periodically without losing his tabs, and regain free memory. If we can do this by restarting, why can't we do it with a button on the GUI, or automatically? According to some commenters, it's because of differing memory management methodologies between C++ and garbage-collected environments like Java. The discussion shows that this is a controversial issue and I make no claims to having all the answers. It does also get quite abstruse and technical, but to me it's an important issue at a time when I know some users for whom their browser may as well be their computer.

 
Main
Previous:
June 2007
Next:
January 2009

Archives

Photos

www.flickr.com
mihalis' photos More of mihalis' photos

Colophon

Validation:
XHTML Validation
 
CSS Validation

Feeds:
RSS2
Atom

Powered by Movable Type 3.33
Hosted by Cornerhost