New Features vs Stability

Its my personal opinion that new features act as "smokescreens," if you will, to solving the real problems and bugs within the game. I'm curious to know how many other people feel the same way with Linden Labs and their updates. I guess my thinking on this topic goes something like this. The more new features you put out, the more possibilities the software will have bugs due to new featuresets. The longer this trend goes on, the crappier the software gets if the real issues aren't taken care of.

So, I'd really like to hear peoples' thoughts on this. Do you guys enjoy new features more, or would you be happier having stability in the software. Your Comments, thoughts, logic, etc are welcome.

What? No pie?

-Ghoti

I wouldn't call the rapid pace of new features "smokescreens", but i would say that they do tend to hurt the stability of SL. Linden Labs should slow down a bit with the new features, like once or twice a month, so there would be less unplanned shutdowns and more time for residents to enjoy SL, as it should be.

With the current vote distribution, which is 100 % for Stability, I wonder why there's not more people voting for proposal # 700, which is calling for exactly this.
It may not help to vote for it as Linden may ignore it, some claims, but NOT voting will definitely get us nowhere.
Now, go vote for #700 and let's try put some pressure on Linden Labs.

I just voted on 700. I had no idea that was even out there. Thank you for suggesting that! Good info to know :3

Pie.

Next poll:

When you go to work tomorrow, are you:

a) Walking?
or
b) Taking a lunch?

Same answer I gave last wednesday on the exact same poll.

My many years in the computer industry has taught me that:

1. Programmers don't like fixing bugs, especially other peoples'.

2. Programmers love writing code, but hate commenting, testing and debugging it.

3. Many bugs get fixed by total rewrites of major sections of code (see 2.).

4. Real time code has bugs that can be extremely hard to find unless very very strict standards are applied.

5. Programmers often have huge egos and will vehemently deny that their code is at fault.

6. When following the rules it can take a long time to write any program.

7. Programmer productivity can vary enormously between individuals.

8. It can take a year before a programmer is sufficiently proficient, particularly in an online environment like sl.

9. Bugs are a bit like LAG, the manifestation or symptom does not really identify the cause sufficiently to sic a programmer at it.

10 Shortcuts to improve speed may have difficult to correct side effects.

Taking into account all the points above, I believe that LL programmers have done a great job in taking the system as far as they have. Stricter controls over the programmers would probably have reduced their productivity significantly.

Many bugs have probably been fixed in the process of writing new code!

Just my $L 5 worth.

Ed

Fixed but replaced with new bugs in new code which often has features people would prefer not to have.

My feeling is that we are just beta testers for their real end project & goal. Which is fine by me. I love SecondLife, even with the bugs.

I did. Gave it all ten of my votes.

coco

I voted features.

(And not just because I'm tired of everyone complaining about stability.)

SecondLife is woefully incomplete as far as I'm concerned.

There are many things that I refuse to give up hope for, and I know better than to expect the transitions they require to be flawless or painless.

The items on MY wish list include: Mono, Havoc 3, re-mapable keys, better interface shortcuts, in-world email or something better for offline messaging than IM or notecards, improvements in object permissions, some kind of privacy controls, and a vast slew of other missing components.

In the interrum, I can put up with some unusual interface changes, a few crashes from time to time and extended downtime on Wednesdays, if it keeps us moving down the road... "to a better tomorrow".


(If I were seeing object corruption however, I'd be singing a different tune.)

--
A bad day from time to time helps us appreciate normal days a little more.

I get that impression at times.

The problem with features, Jopsy, is they are tacking them onto an unstable base. You and I both want new physics and a real virtual machine (Mono for sure, but I'd prefer Havok 4 and the rest of the Havok Engine), and unless these two backend systems are implemented sooner than later they will get even more difficult to integrate.

Even the new Group tools are essentially user interface changes despite backend database updates needed for their implementation. The UI appears to be the only place the current development team is focused on. While flexiprims and local lighting were the only true new features I've seen in my short time here, they too should have taken a backseat to physics and the virtual machine.

When you can point to the debacle that is the new snapshot interface I think it's fair for people to yell for stability over features, as the few features to leak out have hardly been for our benefit.

My vote is for features every time ;0

I'm one of the freaks that actually rather likes the new snapshot interface...

I do agree... much of the development seems to be cosmetic/UI specific and seems to steer clear of the deeper internals, but I would expect that.

Flexi-prims, however, are a great add (now let us use flexi on phantom+physical items dangit!). And the new lighting system has certainly helped silence one of my biggest SL complaints ever.

Changing *anything* to do with physics or the LSL engine had better occur rarely and with a tremendous amount of recruiting people to test out the changes on the preview grid so they can be prepared for the changes on the main grid.

Actually I don't hate the new snapshot system, I just have no need for its features and it gets in the way. If it had been implemented as an alternative I don't think any of this moaning would be continuing to this day over it. It would have been easy enough for them to implement either a button toggle or preference choice to alternate between the two interfaces.

I'm not discounting flexi-prims or local lighting, just saying that we're still in a very Renderware quality environment, without custom bump maps, shaders or anything developers of 3D environments have taken for granted for three years. Even simple things like terrain painting (something you can do on the Xbox Map Maker for Far Cry Instincts), things that impact the look of the world, seem to be back-burnered for UI tweaks.



This is absolutely true, but as I said, the sooner the better. Everything from prim limits to performance would be affected dramatically (especially in terms of Havok, as the initial Havok 1 linux release wasn't even feature equal to the Windows version, and Havok 2 physics on consoles like Xbox and PS2 shows a world of difference on far less powerful hardware than we use for SL).

I played There today while SL was down. SL's graphics and user interface positively beat the tar out of There. SL "looks good enough" I think. Custom shaders and all that jazz would be nice, but there are so many fundamental problems that need attention, that I don't mind waiting for SL to be "like UT2006" in appearance.

What we need is a Havok upgrade to address some of the deep think/sim crash issues, Mono so that our scripts don't run like chilled molasses in busy sims, and some serious attention to the region handoff code. Andrew Linden made some mention of "journaling" in sims that should effectively fix the Twilight Zone issues whenever you go over a sim border. (But, he said Havok 2+ would be necessary first.) The reason I'm saying this is that I think it's hugely important that the underpinnings of the world operate correctly. Right now, it's not so hot.

There's also the problem of bugs that are reported in preview and not fixed before release. One asks oneself, "Why do they ask us to test this for them, and then release it even though it's full of bugs we reported?" I've said this a number of times and I'll say it again: bugs go unfixed for years. It doesn't give me a good feeling to see feature after feature added, while so many old bugs apparently remain untouched. Surely there must be some that are waiting on fundamental architecture changes, but there's crap I reported around this time in 2004 which is still broken to this day.

I don't mean to imply by this that I don't want to see ANY new features. 1.10 was a fabulous release. However, I think a lot of SL users would really appreciate more attention being focused on long-standing bugs, to the point of giving them a lot of priority over adding new features.

Hm. Maybe they should move the voting feature to the "known issues" page instead... let us vote for prioritization on which bugs get fixed first too.