Today's workplace survival phrase is:
"That appears to be a loadscreen bug"
I just read Neil Gunton's article on Open Source Myths. I've got to say that I agree with him, particularly on points 4 ("Open Source software is always better than closed, proprietary software") and 6 ("More choice is always better").
Every once in a while I need a really good hex editor. And let me start by saying that EMACS hexl-mode is a sad joke and the VI hex mode isn't any better.
I need a hex editor that does the following:
* Opens files in a paged manner (a gig of data doesn't fit so well in memory when you need to not shut everything down, and a half terabyte of data just doesn't fit).
* Adjust the alignment of the data so that it has nice even offsets (usually 0x8 0x10 or 0x20)
* Has a nice ASCII dump of the text on the side
* Runs under Win32 (sorry, not all of us can run under Linux at work).
And I tried a dozen different open source projects (because everyone can write a half-assed hex editor). Under Linux ghex did an ok job but really wasn't stellar. I finally gave up and downloaded a trialware HexEditor for Win32 and when it turned out to be a good product I shelled out the money for it.
Which brings me to my main criticism of free/open source software. Most of it is crap, and a lot of the stuff that isn't crap is badly in need of a polish stage.
This shouldn't surprise anyone, most of everything is crap. Most non-free software is crap and unpolished too, but you don't see most of that software. At my company we have dozens of tools to process artwork into something suitable to be used in game. If the average person picked up any of these tools they'd find the tool to be buggy, obnoxious, and generally difficult to use. However, in the very limited context of our art pipeline the tools do a very good job. But nobody here is pretending that these are stable, mature products that are suitable for the average software user.
Every time I look for a utility on freshmeat or sourceforge it seems like I have to try 5 or 6 different projects that turn out to be total crap before I can find that one project that turns out to be pretty good.
This isn't some kind of grand indictment of the Free/Open Source Software movement thing. I'm using Mozilla Firefox to post this, I edit code with EMACS, I edit config files with vim, pictures come off my camera courtesy of gphoto2, and my primary compiler is based off of gcc. There's a lot of really amazing software out there, there's just an awful lot of unpolished, unfinished, craptacular, student project-ish software that has consistently made it very difficult for me to find the good software.
Entries in the QA database:
07/20/04 2:22:05pm Matt - Fixed in v.55: The fix in the previous version made the bug incremental.
07/20/04 2:23:27pm IdiotTester - Not fixed in v.54
Wow, one minute and 22 seconds. It takes longer than that just to verify the bug. It takes longer than that to walk down the hall to the testing pit and give them a DVD. It also takes longer than that to burn the DVD, to copy all the files across the network in order to burn the DVD, and to open the craptacular bug database.
So, no, the version that was burnt at 8:30am isn't really likely to contain the bug fix because our happy friendly tachyon time drive is a little broken right now and we have to deal with boring events that happen one after the freaking other.
And this is why programmers don't really like a lot of testers.
I think we should bury dead sci fi authors with a very large magnet strapped to their body. Then, we can wrap the coffin with wire and wait for a few years for a movie studio to make a movie that bears some resemblance to the title of one of the author's books. At this point the author will not merely have turned over in his grave but should be spinning between 1k and 10k RPM (depending on the ratio of advertising dollars to scripting dollars). A Dead Spinning Author Array (DSAA) should scale evenly with time and be capable of delivering several megawatts of power to supplement a power grid.
Hmm...
If we follow the 90% rule (90% of everything is crap), then spam helps maintain a high quality of interpersonal communication as it's already producing 90% of the email (thus forcing the remaining 10% to be of high quality).
A friend of mine is running in the AIDS marathon. He says that I can kick his ass if he doesn't cross the finish line, so everyone please donate so that I can either kick his ass or buy him a very expensive drink.
Our publisher, in it's infinite wisdom, decided to schedule a press event right before our alpha date, and then moved back our alpha date. So I was stuck at work all weekend trying to optimize the asset streaming.
A few things I've noticed:
1. Clever loading schemes (e.g. radius based preloading with different distances for various asset types) don't seem to work as well as stupid schemes (e.g. drop everything into a bucket, when you're close to the bucket load the whole bucket).
2. The next game I write will have the following scheme: define a region in the editor, associate objects with the region. If you're close enough to the region, load the whole region. Let the artists and designers load balance the map.
3. No matter how optimal your loading scheme is, having the music system thrash the DVD pretty much screws you over.
Ah well... hopefully we won't be skewered too badly by the press.
According to a Reuters article a programmer is suing the games branch of Vivendi Universal for unpaid overtime.
It will be interesting to see how this turns out and weather or not it kicks off a whole bunch of similar lawsuits (there are a lot of disgruntled, burnt out programmers who have left the games industry).
There are, however, a few things that people should notice about this case.
1. VU is a publisher of games. A publisher is very different from a 3rd party independent development house (like the one I work for). Publishers foot the bill for the production of the game and they get most of the profits from the game (they also eat the losses for the many, many failed games).
2. According to the article Neil Aitken says that he was told to falsify his timesheet so that VU would be in compliance with California labor law.
3. There's a big difference between working 70-90 hours a week in order to establish a career or because you believe in a project, and working that much at the whim of another.
The subject of overtime in the games industry (and for programmers in general) is very touchy. It seems like a lot of people I meet feel like something needs to change, but nobody's willing to make the changes.
I love writing code. But I've been working overtime for a long, long time and it would be kind of nice to have a life again.