Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[UmbrageOfSnow]

On the other hand, releasing the new versions in a forum post before putting them on the front page would have the same effect.  Any forum user would be able to get the new version and report bugs, but someone just trying out the game from a link would not initially encounter the game with something like the 31.07 crashing bug.

Toady could just forum post, and if nothing truly game-breaking were discovered before he goes to sleep, he could the get around to updating the front page.  He announces them all here anyway, so that is sort of just a rearranging his schedule thing.

[Logical2u]

There's already an alpha testing group of some sort. If you check the devlog, you'll see Zach did some testing of .31 before its release, and I believe a few other people did as well.

Releasing versions to alpha testers will cause the following scenarios to occur.
1. Forum animosity. Perceived favourites will be selected over many others clamouring for the same sweet, sweet DF, resulting in the afore-mentioned forum animosity.
2. Release delays. Not everyone can test all the time, so while Toady may finish with a release at noon on Tuesday, all but one alpha tester may give their approval before tester X returns from his month long vacation.
3. If not all testers are needed to give approval, then the alpha testers are essentially just people who get the release first, which, I'd hesitate to say, could result in leaks, and then the meaning of 'alpha testing' is completely lost, or alternatively there's no point in giving it to them first, because there's no guarantee that major bugs will be quashed.
4. There's no guarantee an alpha tester will actually do a good job (in any sort of decent time span). I know that I didn't find many gamebreaking bugs when I first started playing the new release... It took a lot of people a lot of time to get the data needed to quash a lot of these bugs. Thus distributing a new release to a broad audience (all of us) seems like a better idea. 

Just putting the versions on the forum before the main page seems like a slightly better idea, to be honest. I think Toady did that with the SDL merged .31 release.

[NW_Kohaku]

4. There's no guarantee an alpha tester will actually do a good job (in any sort of decent time span). I know that I didn't find many gamebreaking bugs when I first started playing the new release... It took a lot of people a lot of time to get the data needed to quash a lot of these bugs. Thus distributing a new release to a broad audience (all of us) seems like a better idea. 

There's also the sheer opacity of the interface to contend with.

Players had essentially no idea that the entire attribute system was broken, nor could they in any real way know it was broken until you had the DF Hack tools that could actually show you that you never gained any attributes, but at the same time, your attributes rusted away.

[Aquillion]

Like someone mentioned, there was an alpha test for DF2010.  But that was partially because there had been so long since a release, and Toady wanted something that was at least a bit stable.  (As bad as the bugs you got in the public DF2010 release were, the alpha test had a lot more -- many many random-seeming crashes, corruption to historical figures and stuff that could easily ruin forts permanently, tons of worldgen crashes, many types of embark areas that were glaringly broken...  I think at one point lakes would always pour directly into the magma, and streams over caverns would frequently collapse into them the moment you embarked.)

It's unlikely that these relatively short, smaller releases would have such a wide array of bugs...  and even if they do, I think that Toady doesn't feel the same pressing urge to try and make the release a bit (comparatively) polished and playable that he did after the long DF2010 wait.  With weekly releases, releasing another fix a week or so later to fix bugs in the last one isn't a big deal...  but if DF2010 had been totally unplayable, it would have been much worse.

So I suspect you'll only see significant pre-alpha tests when there's a big gap between releases.

[Hyndis]

I don't see how that would benefit things over the current model.

The current method finds major bugs extremely quickly. If there happens to be a showstopper bug it shows up within minutes of Toady posting the build.

If the bug is more subtle then someone will still notice it eventually.

More eyes means more bugs are noticed, and they are noticed much more quickly.

[Capntastic]

Anyone that plays DF is the alpha tester group, it's as simple as that.

[marcusbjol]

Downloading and installing new software is the user's choice.  The User has to learn that upgrading to the latest and greatest, especially if the software is alpha, things break.  Often. 

This is true with all Software.  When I was Outsorced IT, I recommended to all my clients to wait 6 months after software releases so the bugs can be worked out. 

The quickest way to find all the bugs is to have as many people as possible playing it.  Limiting that will only make the release cycle slower.  And if it gets out that there is a new version that only a select few are playing, Toady will get flack for that.

If you want something stable, play the d40.  If ya want the latest and greatest, get it.  But please, for the love of god, dont complain if you are using known alpha software and it doesnt work right.  Submit a bug report instead.

[Vattic]

There's already an alpha testing group of some sort. If you check the devlog, you'll see Zach did some testing of .31 before its release, and I believe a few other people did as well.

You beat me to it.

03/23/2010: ...Testers are on their seventeenth version overall, and I imagine it'll be that way for the rest of you as well once I get it sorted out, bandwidth permitting.

From the devlog.