kcw | journal | 2002 << Previous Page | Next Page >>

Last week Shannon pointed out an article where some startup game company needed a high-end database but couldn't afford Oracle so they went with MySQL and that's worked just fine for them. I have no response to that. If MySQL or any of the other limited-functionality high-end databases work for you then you obviously didn't need Oracle nor DB2. You can only make database software so fast -- how a database is structured and tuned makes a far greater impact on performance. So I'm not surprised that MySQL is quite capable.

But think of this: MySQL only recently added transaction support. We're talking about a very basic feature for businesses. The other databases are fast, have great uptime, are scalable, etc. Really, what differentiates Oracle and DB2 with respect to those other databases are what happens when the database dies. There's better support for continuous uptime and parallel processing and transaction security. As an analogy, a medium-sized PC running Linux is quite capable for most small businesses. But larger businesses run their enterprise on high-end machines like Sun and IBM mainframes. Doesn't mean that those mainframes are all that much faster or that their hardware components are more reliable. But those computers are designed for businesses that really need that kind of hardware. Same with Oracle and DB2...

Matt Johnston wrote a small opinion for OS Opinion. Now, I've read some of the opinion pieces at OS Opinion and they've all been rather bland and uninteresting. Sure, they have nice topics but the writers never seem to take those topics anywhere. Same for this article. The gist is that zealots are bad for your image and have hurt Apple sell to businesses. But now there are more mainstream people in business who use Apple products and those are the kind of quite advocates that Apple needs.

I don't disagree with anything he said. And that's the problem. It's an opinion piece and as far as I can tell he's espousing the opinion of the majority. Johnston is being obvious, which is why I thought that the article was boring. But I do want to add something to the topic: although you may not want zealots, they're always there. There will always be a percentage of the user base that are zealots and another percentage that are diametrically uninterested in the product they're using. It would have been nice to read what Johnston's opinion is on how to deal with zealots. Try to suppress them, use them, ignore them. But he didn't so the article comes out bland...

Last topic is about crime and punishment. I saw "Ed" and the episode was about a woman who stole money from the insurance company she works for, so that it could be used to help one of the insured people who had been denied his claim. Now, I have no real problem with that. If you want to do something that you think is the right thing to do, I'm all for that. If it breaks some laws, so be it. But then you have to accept the consequences.

You should obey laws. I believe that. Whether the laws are good or bad, society runs better when everybody plays by the rules. Sure, your company doesn't play by the rules. Does that give you the right to break other rules? I've never believed that you are allowed to do a wrong to correct another wrong. Let me rephrase that a bit. You should not expect to go unpunished for doing a wrong to correct another wrong. You can break a law to correct a great injustice. But I hate it when people then expect that the ends justified the means and they should not go to jail or pay a fine or whatever...

I have a little space left. It's time to set up a Mac OS X machine as my web server. Unfortunately, I don't have a space Mac to run my web site while I build the new machine. I was hoping that Michael would send me his old iMac so that I can fix it and use it but I think he forgot. So that leaves me with plan B: move the services to Thales and then build my server machine. That's not as easy as it sounds because it involves two moves -- move the web site to Thales then move them to the destination machine -- rather than just one move to the destination machine. Oh well, time to start planning.

Copyright (c) 2002 Kevin C. Wong
Page Created: August 21, 2004
Page Last Updated: August 21, 2004