why linden labs continues to be broken

Firstly, I admit I was wrong. I thought ousting the old CTO was a bad idea, but it turns out that he really seems to have been the sticking point in both client and server development, since both have come a very long way in just the past couple of months. Though still to do is to entirely replace the rendering engine with something that doesn't blow goats, and to finally admit that you're using entirely the wrong grid topology for your application. (which is why things lag so quickly and become unusable so often)...its not the speed of the hardware, its the network latency between nodes that does that.

But the biggest problem, it seems, is Linden Labs having put all of their eggs into one upstream basket. For those of you who don't know, LL's upstream provider is Level 3, and Level3 is notoriously awful at dealing with their own network issues. Not that there's any one provider that's perfect, but the general idea behind an application with a user base so large and decentralized is to diversify upstreams. You can't just use Level3 and expect everything to work all the time, Lindens. If you don't want to pay for multiple upstreams for extra bandwidth and resiliency, then you can just move all your crap to a colo-center that does it for you.

See, then this sort of thing just wouldn't happen.

Actually.... LL are rolling our European Servers and other countries/continents will have their own localised data centers. This will imporove the topography of SL and make it truly universal and worldwide.

The biggest shame of all. you got to hear this stuff from the backdoor, they announce nothing, then BAM, release and update.. What they MUST understand is, SL is now big business and when they screw up, Business's suffer in a big way, not just in sales, but with aborted meetings and schedules, events and exposure.
This goes against what LL have been doing the first 3 years I was inworld, they repeatedly said over and over.. lookout for Havok2.... lol...

Ohh i knew it, level3.net. I have a VDSL Connection 50Mbits/10Mbits and suffer from severe packet loss happening at the gateway to level3.net and on their routes.
I talked to my provider- Deutsche Telekom - the inofficial statement they did on their gateway packet loss was something like: Our peering partner level3.net puts more traffic in the lines than they have capacity. I guess that says all about that.

Linden Labs is broken?
How does a company "break"?

Ask Enron

Thats exactly it. Level3 oversells their capacity, as do most American backbone providers. The way out of this is for Linden Labs to diversify their connectivity, not only between data centers, but inside each individual data center as well. Doing so would require the hiring of a very brainy network engineer who understands BGP *really* well. Those people are very expensive, which might be part of the problem.

Another part of the problem is that LL may have, from within the distorted reality field generated by their ex-CTO, signed an exclusive agreement with Level3 in order to keep costs down.

There's that "cost" thing again.

The fact is, Linden Labs is making enough money to keep chugging along, and they're forming close relationships with the big boys in entertainment and computing. They don't actually NEED to diversify their connectivity between providers (though if they did, I would suggest coupling Level3 with at least Cogent in every data center that is accessable to both, and dropping the data centers that aren't entirely) in order to be profitable and to please their financial and commercial interests. The only thing they'd gain is a more stable, loyal user base.

That's the "right" thing to do of course, but there's usually a difference between "right" and "profitable".

They update things when updates are not needed
example, new interface.

For anyone in SL, consider this...

If SL worked 100%, 100% of the time... who could you blame for, lateness, incompedence, shoddy workmanship, failure to reply to NC's, IM's, bad scripts... the list goes on...

In RL, ppls blame the post, emails being 'lost', letters never arriving, didn't get that all important text message, trains late, traffic..... the list goes on...

In SL, with the way things are, its our 'Get out of jail free card'...... 'You know how SL is at these things'.... 'Yeah.. np, understood'.....

Let me paint you an image...

Blizzard is generally adored (with a few far between hiccups) by WoW users, who generally (as the polls constantly show) feel that they're almost always getting their money's worth and having a good time.

And then there's Linden Labs.

The difference?

Management and finance. Blizzard is willing to pay brilliant programmers what it takes to keep them, and their project managers generally only interfere when things go offcourse, nudging them back on. LL is precisely the opposite.

One has all their content sitting on your hard disk and one doesn't?

edit: no! that's not it.. One lets every user in the environment create stuff and one doesn't!

edit edit: no! that's not it either! One provides a single world view and ones doesn't!

That whole dynamically user modifiable environment argument is old and very, very worn out. They've had six years to make it stable, and it still hasn't happened. Do you understand how long six years is in the software world?

No more excuses, just fix it.

And yet nobody else has come even close to what LL has done.

Yes, of course. I don't see why they're not pressing the "fix" button.

LL and SL is currently serving as an example to any company wanting to do the same thing:

We'd better get our act together before we go live with this project. We need to be able to compete with what everything SL offers, yet be more stable and have better costomer service.

If I were a virtual world contender I wouldn't go live until I had a scripting language, sim crossings, content creation, load times, physics, and load handling down so smoothly that when the system goes public SL users go "Holy crap, it feels like SL 4 years ago in stability, but with a population of last week!"

That could well be true - I don't doubt it at all.

But until it happens, the "they've had enough time" arguements used above don't really work very well.

lol a lot of us would love a SL of 4 years ago period. Features and Population included.

Wouldn't you miss Excite?

If SL had stagnated from 4 years ago, I doubt it would have 50K + ppls logging in at any one time, maybe 5000 or less. Even the OpenSim projects would look good compared to then. The only real difference would be perhaps the lack of lag on 2 basic points, lack of ppls ingame and the upgrading of our connections.

As a Social platform, it probably would not have suffered too much, but SL is sooooo much more now, the diversity being that equal to wandering the globe in RL. However, there are many more platforms out there that are attractive to the Socialites.

Not all 'features' are as welcomed as others, some simply defy logic, but diversity is key to progress.

And sim boundaries wouldn't rubber band so much!
And TPs would work, the search would work, there wouldn't be so many outcries of content (childporn I don't think existed on the grid yet, because no one was that stupid)...

Sure, there were fewer people, but almost everyone you meet you could talk to and be friendly with.

Don't forget this one: LL wondering if they're going to be able to keep the lights on!

Do you go to the industry conventions? I do. The reason why no one else is getting into the same exact scene as Linden Labs (an end-less game with no rules, plot, or point, but with user-modifiable worlds and tunability) is because, as is continually stated over and over, everyone is banking on Linden Labs failing.

And they are, very, very slowly. The rest of the industry---and by that I mean the sorts of people and organizations who would otherwise come in and give LL some competition, see it as far, FAR too much trouble with very little reward. The model of giving away the software for free and then not requiring a periodical fee of some sort, but instead relying on land and business owners to infuse SL with money has *yet* to be proven as a stable, workable method. Even with people like IBM and Best Buy *cough* filling in some of the holes, it's still not really viable, and everyone knows it.

However, when a bigger fish (*COUGH*GOOGLE*COUGH*) ends up making an offer that LL can't refuse, look for some interesting and finally, most welcome changes made by an organization who can actually afford good programmers.

No, Second Life really, really is still just very, very fancy IRC.