The new toolbar, which hyperlinks content such as ISBN numbers and addresses, has aroused ire in various quarters. The main concern from several businesses is that this could divert their (potential) customers to other web sites. Some have complained that this feature is very much like Microsoft's SmartTags, which was pulled due to trust and trademark concerns.
Are these people overreacting? Maybe. The feature may not be enabled by default. Also, just because a link is present does not mean that people will use it.
OTOH, Google has been getting an increasing amount of negative publicity from some folks. I've seen lots of complaints in places like alt.internet.search-engines and SearchEngineWatch of worsening Google search results, over and above the usual "My site has vanished from Google" after the latest new index push. I personally don't use Google nearly as much as I used to; Y!'s results are good enough for me.
Several years ago, it could be argued that Google's results were better than AV's, but hardly anyone knew that because AV had the mindshare. I wonder why Google is putting so much energy into new services such as this, at the risk of jeopardizing their core business?
Microsoft, Oracle, and some other companies have recently complained that the plethora of security software bugs is due to a lack of secure programming instruction. While it may be the case that there should be increased security instruction in CS curricula, I still have problems with this statement:
- Microsoft has more than its fair share of security bugs! Microsoft also has one of the most comprehensive screening processes for software hires in the industry. There are quite a few people out there who do have strong security backgrounds, but don't make it through Microsoft's screening process because its focus isn't on security. If Microsoft made security concerns a stronger criterion for hiring, instead of (for example) puzzles irrelevant to computing, perhaps they'd have less security problems.
- Security is expensive and time-consuming. Microsoft has a history of shoving products and services out the door as early as possible, with insufficient QA.
That's a bad recipe for promoting secure software.
- There are talented security folks who just won't work for Microsoft, period, because of their attitude towards open source, Unix/Linux, etc.
- A proper treatment of security would most likely require its own quarter or semester course in a CS curriculum. It's already difficult enough to cover the basics in four (or so) years. Tuition is rising, and some students are (rightfully) less enthusiastic about CS than they were a few years ago, due to the decreasing demand for s/w engineers, outsourcing, etc.
So, I'm not too sympathetic to Microsoft and some of the others. You all brought this upon yourselves when you decided that s/w engineers were expendable cheap labor.