We need reliability first

Reliability is the first thing we need for any activity. We need reliabilty for education, we need reliability for in-world as well as cross-world business. We need scripts that work, information networks that work, displays that rez. We can't play anymore the "guess game". Will it rez or not? Should I rez another one? Is this object lost or not rezzed? Is this object inventory empty or just not loaded? Is the notecard i've put in really part of the inventory? Should I put it again and end with a duplicate? Should I wait? How much time? Will my mailers work today or my network be broken again? Will rpc work or timeout? Will llSetScriptState work today or not?

Landbots, camping, griefing are second problems. Anything else is second problem. Without reliability, nothing serious can be done. Just a 3D laggy chat with questions appearing below answers. Any work will be a waste of time, crashing, reloging, bug reporting, and never having the job done. Just lost an important notecard today. 5 hours of work gone poof. Soft clearing cache did not help. Hard clearing cache did not help. Switching client did not help. And no, it was not mis-filed or trashed. It's lost. I'm now trying to trigger an inventory reload until my 11k assets are loaded but it never reaches the end. Counter stuck at 10,645 after 30 minutes, but I know I have more. Playing the guess game again. That's enough.

We need to know WHEN an operation is finished or if it failed. Not play the guess game nor wait forever. Think about a system that MAYBE will file data. MAYBE will execute a program. This system has a name: Second Life.

You're right, but LL's interests are to grow the population as fast as possible. What we'll see is a dynamic equilibrium between these two factors, because unreliability causes people to give up. LL will understandably focus on the issues that affect the most residents as the highest priority, while working on infrastructure scalability issues.

So, you're right, but that's beside the point. Unfortunately for us. Frankly, SL is a game, and many of us are willing to put up with a certain amount of problems and stay in the grid.

Sorry to hit you with the hard reality, but that's how it is. Reliability will become more important as SL becomes used more for RL purposes, but at present that's a serious minority. On the other hand, there's speculation that some recent changes (e.g., voice chat) were pushed by commercial interests (that is, RL companies that have private islands for commercial purposes). To the extent that these companies represent big bucks to LL, we'll see a bigger push for reliability.

For example, I work in data communications. Enterprise customers (companies running their internal networks) have put up with a lot of unreliability in order to get new features and low costs. In the last decade, though, the emerging market is telecommunications, and those customers (e.g., phone service providers) have a much greater demand for reliability, and they're willing to pay for it. So companies like mine shift their focus from adding features as fast and cheaply as possible to improving reliability.

The bottom line is that LL will focus on whatever affects their bottom line the most. If paying customers demand reliability, we'll get it. But my suspicion is that most of their revenue comes from participants who are essentially playing a game, and that will continue to be true for quite some time.

splitting hairs...

I agree with the OP. I'm scarce in world at the moment, mostly because I get too quickly frustrated with uncertainty when I'm trying to get something done. I recognize that many of the problems we face are likely due to packet loss which results in incomplete data being transmitted, and then not resent.

When I have to deal with a 'production system' the priorities I invariably set for it are:

Integrity - is the data correct? Viewable only by those that should see it, updatable by only those that should... and free of corruption. Without high data integrity (there is no 'perfect') nothing else matters. Is the data recoverable in the inevitable case of system failure?

Reliability - can the data/service there when it needs to be? Is it responding quickly enough to be useful? Are there unplanned outages? Are some transactions just failing?

Availability - How much of the time is the data or service available? Is it easy to bring back online quickly after an unplanned outage? Are planned outages reasonable?

etc... etc.. etc..

Yes, SecondLife is a form of 'recreation' (game is too narrow a term I think), sure as such it doesn't need to hit the same levels that a financial app on wall street does, and it certainly doesn't pull the same kind of revenue as wall street apps.

The problem is... Most data services provide specific functions. SecondLife is so open-ended that as soon as things are running well someone comes along and finds a new way to use SecondLife that bogs it down again. It's a blessing and a curse.

All systems expand until they can expand no further. They either implode when that limit is reached or they grind away expending great effort and to push that limit.

So far, SL seems to be in that second category. While it may be frustrating... I certainly prefer it to implosion.

inarticulate stream of concsciousness writing disclaimer: Haven't had my caffeine yet... sorry.

You are both right. I'm here to play. My recreation is building & programming. And building & programming (especially programming) on an unreliable game is no fun. It is frustration. So the game is no more a game. LL's interests are to grow the population as fast as possible. But what is the population's intererests in SL? CONTENT. Without content, SL is just a laggy chat system. Jopsy, you gave us one of the most delightful places in SL. I often go relax in the Cave or the Cloud Castel. Daryth gave us the dragons. But I can barely rez as a dragon, now.

To be more constructive... UI. The UI fails to give feedback of IF and WHEN an operation has succedeed. Example : i'm "boxing" items. I drag a folder from my inventory to the Content window. When is the copy finished? Can I close the Edit window before I see the whole list? Can I quit the game now? Not sure. So I try to check by opening the object with the Open menu. Argh... only half of the list is here. Is the copy running in the background? Did it failed? Or did the copy succeeded but the reload failed? Don't know. So, I restart from the beginning. Open in Edit, select all Content and Delete. But the list doesn't empty! Is it cleared or not? Don't know. Yes, it is asynchronous. But I never know WHEN the UI state matches the server state. How can I distinguish a pending copy from a failure to copy? A failure to copy from a failure to load back?

So... In your opinion, a level editor for [insert game title here] is a game in and of itself? In the end, a level editor is no different than the content creation tools we have in SL, SL's being even more open and less defined (and under the definitions of a game, farther away from being a game).

SL has games, but it in and of itself is not one.

Your pedantic semantic comment is valid, but the point remains that the primary market for SL is for entertainment, and overall quality demand accords with that.

I agree completely with the OP concerning the priorities I'd like to see LL use. However, their priorities are different for the reasons I mentioned.

Jopsy summarized it pretty well. However, SL hasn't come near hitting the limit; and the on-line population will continue to grow even as reliability goes down. LL's ideal strategy is to optimize the cost/benefit ratio of development, where the benefit is measured in terms of either profit (near term benefit) or population (presumably producing a longer term benefit in some way). Adding features provides benefits; increasing reliability does too. However, they will certainly test the limits of our tolerance. (This is news? haha)

As a builder or scripter, the most frequent problem I have is that I often get disconnected without any indication. I can be editing away for hours and as long as I don't rez or shift-drag-copy a prim I won't notice it, and all my careful adjustments get lost. This is terribly frustrating. So I save frequently and test connectivity frequently (by pulling up my profile, for example).

Of course, that kind of reliability is less problematic than llGiveMoney() calls that fail! Ah well, such is (second) life.

I deny none of this and have never stated as otherwise.

In my opinion, programming is a game. A game I'm paied for when I do it at my office, and not paied for when I do it at home.

Back to UI. The UI was designed with the asumption that the server is fast. All operations runs asynchronously and no feedback is provided (progress bar for example) because they run in an "enough short time" for the visual result to be the feedback.

If rez takes one second, the rezed object is the feedback. If it takes more than one minute, it is aborted with a blue message. In no system the user can wait one minute with stopwatch in hand to proceeed with his task, especially if he runs multiple in parallel. If speed don't come back to SL, and even if core reliability is achieved, the UI will fail. At least for content creators. Although it may be fine for chatters, who I suppose are a majority of users.

I agree with Jopsy completely in this post. But as a lay-person with no real knowledge of what it takes to keep the SL servers humming, my question is this:

At what point does LL stop, take a look around, and see the mess of things not working correctly? The open-ended nature of SL will of course make it difficult to keep on top of every way the platform can be compromised, but I guess I just don't grasp why, when things break, does LL blindly keep pushing forward instead of taking time to correct thier mistakes and fix the once working elements of thier platform that have become corrupt or broken?

They've already proven you can achieve growth with no real reliablity, and I agree with Jopsy that it has been a thrill to witness LL push the limits of what was thought possible. But at some point, somewhere along the line, somone really needs to sweep up the mess pushing forward with such great effort has left in its wake. Where's the clean-up crew? Why do I still need to hit World/Region/Estate to get the water back on after tp-ing? Stuff like that.

Because I'm still not so sure we aren't all going to bear witness to a really big implosion.