Experimental vs Stable, What do you prefer?

Started by GigaLem, March 02, 2017, 12:42:40 AM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

What do you prefer?

Use the experimental right away and test out the new toys?
0 (0%)
Wait till the latest stable release so the bugs aren't as plentiful?
2 (100%)

Total Members Voted: 2

GigaLem

I want to know something

With the new things every experimental version has to offer, what do you do?

Are you willing to download new versions everytime a quick bugfix is made with an experimental that is in the works?
or
You're willing to wait no matter what til the newest stable release is made so you have more time to toy with it without as many bugs?

I know i'm the latter in this one but I want to know what you think, each side has their advantages and disadvantages

ccexplore

The point of the experimental release track is so that people can try out new things even though it may not be fully stabilized (ie. still has bugs yet to be discovered and fixed, potentially very bad ones).  Some bugs may not be easily discovered by internal or self testing, instead requires enough people have tried things out to show up.  Some things related to user experience simply need exposure to enough people to accurately assess.

We keep those risky things out of the stable release track, so people who are less interested in the latest-and-maybe-not-so-greatest can still play and edit levels without worry that something bad may happen causing, say, lost of replay or corrupted levels (for an extreme example, though both have definitely happened).

You can in fact potentially use both versions at the same time for different things.  For example, you may choose to do backroute fixes in an existing levelset you already released using the stable version, so the edited level will remain playable in stable versions and will not have risk of problems introduced by bugs in experimental version (even if you do backups, a crash that could wipe out hours of level editing is probably something you want to avoid).  In the meantime, you may play around with throwaway levels on the experimental version to try out something new like a new skill or something.  You may playtest an existing released level on the experimental version to verify it remains solvable and doesn't have new strange behaviors, and if you found any, you can report such bugs right away so they can get fixed, before the changes eventually get ported to the next stable-version release.

Also, some people may, from time to time, skip a version or two on the experimental release track if a particular version doesn't offer anything of interest (in the form of new features or critical bug fixes).

We've seen bugs getting discovered and fixed through people using experimental releases, so I'd say the system is working as intended.  The poll seems unlikely to tell us anything that would change how we are managing releases.

GigaLem

Quote from: ccexplore on March 02, 2017, 01:08:50 AM
The poll seems unlikely to tell us anything that would change how we are managing releases.
Well to be honest, I wasn't intending for releases but more for just what I really want to know.

If you ask me, I don't want to worry about any nasty bugs, like that draw leak versions ago, which is why I choose to wait

ccexplore

Yeah, that's fine.  Just keep in mind that the less people try out experimental versions, the more likely a bug or usability issue will slip through and end up in the next stable version, where it can be more damaging/annoying to more people.  I would suggest that rather than always completely avoiding experimental versions, be smarter about what you use it for.  Make the assumption that an experimental version, especially one with many new features or changes, will more likely have crashes or other major issues that can wipe out hours of work or play in worst case.  Don't use it on any levels where such risk is not acceptable to you.

GigaLem

Quote from: ccexplore on March 02, 2017, 01:22:13 AM
Yeah, that's fine.  Just keep in mind that the less people try out experimental versions, the more likely a bug or usability issue will slip through and end up in the next stable version, where it can be more damaging/annoying to more people.  I would suggest that rather than always completely avoiding experimental versions, be smarter about what you use it for.
Fair enough, I did use some experimentals for my Deadnought .nxp (one to test out the victory theme)