PDA

View Full Version : Scorched Earth Failure



XyZspineZyX
08-30-2003, 04:35 PM
S! All!

We attempted to run multiple missions with our campaign system last night. We failed. Following is a recap:

We got off to a late start. We were supposed to start at 10:00 EST. A lot of work went into this week, and I was actively keying-in waypoints until about 9:30. I then put my daughter to bed and right around 10:00 I got to my computer. This was the best part of the night...

The previous week I had not been able to fly. My TeamSound wasn't working correctly and my joystick wouldn't stay calibrated. While others flew and had a good time last week, all I could do was taxi out of the way and watch from the ground. My computer was actually having a number of problems. My wife wouldn't let me format the hard drive, so I decided to setup a new partition and really optimize it for Forgotten Battles. Well - to make a long story short I still had a couple of things to install at 10:00 before I was ready to go. We didn't actually go until 11:00. So we were an hour late right off the bat.

We had a lot of interest and launched the first mission with 32 people. The mission however did not load. After trying a couple more times, I discovered that it was my computer rather than the mission that was having problems and e-mailed the mission to Taipan. The mission finally got underway at about 12:00 - two full hours late.

In preperation for the test, I'd made some changes to the log parser. Previously I had to make sure that every vehicle type in the mission was in the database first. I could parse missions without the vehicles being in the database, but I couldn't make a new mission. Over the week I'd re-written part of the parser to fill the database with vehicle types whenever it hit one it didn't already have. BIG IMPROVEMENT. I'd tested this several times and it worked fine.

Well - there are a few differences between the log files for different versions of the game, and Australia has a different version than the United States. The differences are minor, but there are a few differences. Consequently I couldn't parse Taipan's log. The problem was related to the new code I'd written.

It's always good finding and fixing a bug, but that wasn't what we were trying to do last night. By the time I started to dig into what was causing the problem, it was 1:30 AM - getting a little late...

Had we gotten the first mission off on time it would have been easy to fix the problem and move on. But we didn't. Everything that could have gone wrong last night did go wrong. I've never had more unrelated problems when trying to test. Last night was as bad as it gets.

Another way to look at things is that we had committed to running that test on Friday night, and we really needed more time. We should have postponed last night's public test, and ran with just Taipan and myself.

I like to look at 'public' tests as being somewhat scripted. I say 'scripted' because you ideally have already tested everything and you know it works. Last night was not 'scripted.' The changes were such that I thought we'd be OK, but not only had I modified our program, but I'd also made major changes to my computer. The combination of the two created a disaster.

I want to appologize to everyone who showed up last night. It was my fault that my computer wasn't running properly and my fault that the test failed. We had a bunch of people show up to play, and while we did get one mission off, we did so after a two-hour delay and never did launch a second mission. I appologize to everyone for that delay and for the time they wasted waiting.

In the future Taipan and I will do our public tests on Saturdays. We will do PRIVATE tests on Fridays - and while the public is certainly invited to participate in the private tests I will not announce them and will make sure anyone who enters knows that there might be problems. Saturday tests should be far more reliable - in fact they won't really be 'tests' at all because we will know everything works.

In closing, I would like to put last night into perspective. Scorched Earth did not fail. Frankly, we never even got to the part we were trying to test... I failed and my computer failed. The RAF74 IL2/FB Dynamic Campaign System works. I will be re-grouping and using the log file from last night along with the waypoints i've keyed into the database to make sure we are where we need to be. I will even fly a second mission just to make sure it will load. If I get this done today, then we will test again tonight (this time successfully). If I am not ready tonight we will have to wait until next weekend - and if that happens we will have more functionality as well...

I wish I could promise that last night won't happen again, but the fact is that Taipan and I are putting a tremendous amount of work into this and that there are ALWAYS changes in the program. Usually any changes I make between Friday and Saturday nights won't be included in Saturday's testing, but there may be times when it makes sense to include those changes. While I can't promise that Saturdays testing will always be perfect though, I can say that what happened last night is not indicitive of our project as a whole and that we will take steps to minimize the chance of another public failure.

If you see us on HyperLobby tonight, it will mean we are ready. If I'm not sure it will work I'll be coding and testing privately tonight rather than playing on HyperLobby...


RAF74_Wall-dog

<a href="http://www.doghousecomputing.com/scorchedearth">http://www.doghousecomputing.com/scorchedearth/images/parsersmall.jpg

XyZspineZyX
08-30-2003, 04:35 PM
S! All!

We attempted to run multiple missions with our campaign system last night. We failed. Following is a recap:

We got off to a late start. We were supposed to start at 10:00 EST. A lot of work went into this week, and I was actively keying-in waypoints until about 9:30. I then put my daughter to bed and right around 10:00 I got to my computer. This was the best part of the night...

The previous week I had not been able to fly. My TeamSound wasn't working correctly and my joystick wouldn't stay calibrated. While others flew and had a good time last week, all I could do was taxi out of the way and watch from the ground. My computer was actually having a number of problems. My wife wouldn't let me format the hard drive, so I decided to setup a new partition and really optimize it for Forgotten Battles. Well - to make a long story short I still had a couple of things to install at 10:00 before I was ready to go. We didn't actually go until 11:00. So we were an hour late right off the bat.

We had a lot of interest and launched the first mission with 32 people. The mission however did not load. After trying a couple more times, I discovered that it was my computer rather than the mission that was having problems and e-mailed the mission to Taipan. The mission finally got underway at about 12:00 - two full hours late.

In preperation for the test, I'd made some changes to the log parser. Previously I had to make sure that every vehicle type in the mission was in the database first. I could parse missions without the vehicles being in the database, but I couldn't make a new mission. Over the week I'd re-written part of the parser to fill the database with vehicle types whenever it hit one it didn't already have. BIG IMPROVEMENT. I'd tested this several times and it worked fine.

Well - there are a few differences between the log files for different versions of the game, and Australia has a different version than the United States. The differences are minor, but there are a few differences. Consequently I couldn't parse Taipan's log. The problem was related to the new code I'd written.

It's always good finding and fixing a bug, but that wasn't what we were trying to do last night. By the time I started to dig into what was causing the problem, it was 1:30 AM - getting a little late...

Had we gotten the first mission off on time it would have been easy to fix the problem and move on. But we didn't. Everything that could have gone wrong last night did go wrong. I've never had more unrelated problems when trying to test. Last night was as bad as it gets.

Another way to look at things is that we had committed to running that test on Friday night, and we really needed more time. We should have postponed last night's public test, and ran with just Taipan and myself.

I like to look at 'public' tests as being somewhat scripted. I say 'scripted' because you ideally have already tested everything and you know it works. Last night was not 'scripted.' The changes were such that I thought we'd be OK, but not only had I modified our program, but I'd also made major changes to my computer. The combination of the two created a disaster.

I want to appologize to everyone who showed up last night. It was my fault that my computer wasn't running properly and my fault that the test failed. We had a bunch of people show up to play, and while we did get one mission off, we did so after a two-hour delay and never did launch a second mission. I appologize to everyone for that delay and for the time they wasted waiting.

In the future Taipan and I will do our public tests on Saturdays. We will do PRIVATE tests on Fridays - and while the public is certainly invited to participate in the private tests I will not announce them and will make sure anyone who enters knows that there might be problems. Saturday tests should be far more reliable - in fact they won't really be 'tests' at all because we will know everything works.

In closing, I would like to put last night into perspective. Scorched Earth did not fail. Frankly, we never even got to the part we were trying to test... I failed and my computer failed. The RAF74 IL2/FB Dynamic Campaign System works. I will be re-grouping and using the log file from last night along with the waypoints i've keyed into the database to make sure we are where we need to be. I will even fly a second mission just to make sure it will load. If I get this done today, then we will test again tonight (this time successfully). If I am not ready tonight we will have to wait until next weekend - and if that happens we will have more functionality as well...

I wish I could promise that last night won't happen again, but the fact is that Taipan and I are putting a tremendous amount of work into this and that there are ALWAYS changes in the program. Usually any changes I make between Friday and Saturday nights won't be included in Saturday's testing, but there may be times when it makes sense to include those changes. While I can't promise that Saturdays testing will always be perfect though, I can say that what happened last night is not indicitive of our project as a whole and that we will take steps to minimize the chance of another public failure.

If you see us on HyperLobby tonight, it will mean we are ready. If I'm not sure it will work I'll be coding and testing privately tonight rather than playing on HyperLobby...


RAF74_Wall-dog

<a href="http://www.doghousecomputing.com/scorchedearth">http://www.doghousecomputing.com/scorchedearth/images/parsersmall.jpg