Gavin Hird

Ebbe Linden wrote: Gavin Hird wrote: Ebbe Linden wrote: Gavin Hird wrote: Bobbie Faulds wrote: There won't be land, per se, not like the SL model anyway. You'll pay a yearly fee for a domain and host your land yourself. That was exactly what I did for Highfidelity. I have said it before, the new platform is codeveloped with HF, and there will be some changes that sets it apart, but technically very much the same. No, it's not codeveloped with HiFi. Two different companies. It is only that every time I read or hear something you have said, or others have said (who think they are in the know) I get deja vús about what I see in HIghfidelity. – Which I think is not very good (neither from a technological or marketing perspective.) There are similarities as we're in a similar category (VW, VR). There will be many...just like any other interesting category. We have different approaches though. That's not the spirit Ebbe. It needs to be 'There will be many, but they will fight over the scraps as we take the lead!' You really need to stand out, and you need to enter the mobile space and leave the PC behind (for the mass market)

Ebbe Linden wrote: Gavin Hird wrote: Bobbie Faulds wrote: There won't be land, per se, not like the SL model anyway. You'll pay a yearly fee for a domain and host your land yourself. That was exactly what I did for Highfidelity. I have said it before, the new platform is codeveloped with HF, and there will be some changes that sets it apart, but technically very much the same. No, it's not codeveloped with HiFi. Two different companies. It is only that every time I read or hear something you have said, or others have said (who think they are in the know) I get deja vús about what I see in HIghfidelity. – Which I think is not very good (neither from a technological or marketing perspective.)

Bobbie Faulds wrote: There won't be land, per se, not like the SL model anyway. You'll pay a yearly fee for a domain and host your land yourself. That was exactly what I did for Highfidelity. I have said it before, the new platform is codeveloped with HF, and there will be some changes that sets it apart, but technically very much the same.

Cinnamon Mistwood wrote: I hope people like individual, isolated experiences, because they have not stated how people will meet others. They are catering to the selfie generation. Maybe the new viewer will come with a selfie rod. – Kindoff like an Occulus mounted on a rod theater binoculars style.

melt Mills wrote: I am considering buying a new computer and have looked at the new iMac. I have never had an apple computer before and was wondering if anyone could tell me how they go with second life. Or does anyone have any suggestions for something better? When you say the new iMac, do you mean the 5k machine? Although this has the most "hefty" GPU before going to a Mac Pro, keep in mind that it has a 5k screen to haul around and runngin the SL viewer in this will probalby be sub-optimal. If I recall right arstechnica.com have tested the 5k Mac for gaming purposes and concluded it performs quite good, but if you halve the screen resolution to one of the 27" iMac standard resolutions performance suffers. I am running the SL and Kokua viever on and iMac 27" with the 660M processor and 512 Mb GPU memory. As long as you don't try and run in Ultra mode, it runs pretty decent even with shadows and AO turned on.

Horrible with potential.

Continue to develop my opensim grid XMIR

LlewLlwyd wrote: Gavin Hird wrote: I worked in Apple product management. I know my audience. ;-) And for the record, Photpshop was delveloped for the Mac as was Excel originally. FOR the record, a study comparing Apple and PC "audiences" found that the average reading age of documents produced on the former was mid-teens, and for PCs twenty-something, and that the former displayed five times as many orthographic errors of one kind or another. ***As you have conveniently demonstrated*** For the record a study of forum posters who gripe about typos usually have English as their only language, and are out of valid arguments. ;-)

Dresden Ceriano wrote: Gavin Hird wrote: I worked in Apple product management. I know my audience. ;-) OMG!!! How dare we question you? Now we know for certain that no one who isn't a complete Apple whore has any sort of creativity whatsoever. ...Dres Yeh, how dare you?! ;-))

Mac users don't choose OS X (which is what they choose) because of SL. It is only that when they try to, it's a total letdown because ot the technical train-wreck the Mac viewer is.

I worked in Apple product management. I know my audience. ;-) And for the record, Photpshop was delveloped for the Mac as was Excel originally.

Sassy Romano wrote: Hmm, can you cite the evidence that says that Mac users are amongst the most creative or why they're more likely to spend more money than other customer groups? Or for that matter why they are most likely to accept subscription based models? iTunes and "creative types" (music mixers and most likely to use Photoshop etc. isn't enough) You only need to look at number of applications in the iOS App store, turnover in the store, and revenue paid to developers for those apps. It is 6x higher than for the same Adnroid apps. Also entire creative industries were created on software originating on the Mac. Photoshop included.

The reasons for people choosing Macs are mutliple and price is really not much of an issue at all. Making snide comments about it is futile. ;-) What Linden Lab should ask themselves from a business standpoint. Why on earth are they chastisizing the customer group (Mac and iOS users) who are MOST likely to accept subscription based models (Premium), are amongst the most creative and are likely to spend more money than any other customer group they can recruit?

Tomarax Davidov wrote: The problem is systemic with this particular software and really needs to be addressed, but the chances of that are not likely unless the Lindens are doing some really major changes and aren't telling us about it. The chances of doing anything fundamental to the Mac viewer are slim to none. If you look in the About SecondLife menu item in the viewer it says Compiled with GCC version 40201. That is a compiler Apple deprecated for system 10.6.8 and stopped supporting in 10.8 completely. It is also built with the Carbon framework and an old version of QT. It is like they would design the Windows version with a Windows 2000 target. THAT is how outdated the thing is. The BIG thing (gasp) for the 3.7.27 version I use now is that they are able to compile it with Xcode 6.1 after hacking it to run GCC. For iOS development we are on Xcode 6.3. To get up to speed they'd need to rewrite a good portion of the Mac viewer with the Cocoa frameworks, make it 100% 64-bit, ditch QT (which is terrible on the Mac.) use modern frameworks like CoreData, SpriteKit and SceneKit, and of course modern memory management, the built in WebKit, QuickTime, NO Flash and sandboxing to name a few. Unless they get the epiphany to write a viewer for iOS, it is only going to deteriorate. Only then will they be able to start making headway. I'd say we need a SWIFT viewer!

It might have to to with the Retina display which of course the SL client is not optimized for at all. I started the 3.7.27 version of the SL client on my 2012 MPB connected to a 2560 x 1440 display with the viewer running in a 1680 x 1050 window. Initial CPU/GPU temps were 64/46 and after 4 minutes it is at 100/92 (deg C). Fans starts going and brings it down to 80/72 where it keeps it. Meaning the system is happy with that temp with no alarms for any of the sensors. Corrected the SL version that was missing a 7.

There might be. The design point for MacBook Air and MacBook is full day battery operation with "normal" use such as office, mail, web, movie playback and music. They are not really built for running the GPU full tilt for prolonged periods. The MacBook Pro should handle this fine.

Some of the Mac portables will handle the SL client quite good (those with descrete GPUs) while those with Intel integrated graphics will struggle more. There has also been systemic overheating issues with some Mac portable models througout. The first one I used for SL with Intel processor (not PowerPC) back in 2009 melted the GPU 3 times and each time the logic board was replaced under warranty, but at the end of the day I ended up getting an iMac for SL use. There are also some issues with a 2012 Macbook model, so it might be worthwile to check Apple's support pages if the serial number of the machine has an issue and is eligible for replacement. The CPU temperature on Apple portables often ends up between 85-100 deg C, the GPU somewhat lower when using the SL client

Good! As a technology for SecondLife it was DoA. Not scaleble, not sustainable, a sidetrack and distraction. So this is all good.

Russia has about 152 million or which the bulk lives in Europe, Ukraine has 46 million not part of the EU, shall we go on...?

Since when did Europe only consist of the EU? You left out closer to 200 million Europeans from your chart..

Actually the Blue Mars mobile app kindoff did that. It would be more fun though if others could observe you while doing that.

I think a mobile app that would satisfy the instant need for presence and attention that many need inside a virtual world would be a good starting target. That would of course be for the current SL/OpenSim type virtual worlds and not a version 2. To create a (version 2) virtual reality that is relevant throughout the day of everyone is close to impossible simply because people have enough with filling their first reality. Virtual realities are first and foremost something you can engage in if you have a surplus, or lack of ability to realize first realities for whatever reason. Were we may see a wider market is for agumented realities and also for produced realities that takes a movie or TV type production to the next level with virtual presence in the scenes. The dynamic nature of SecondLife / OpenSim may lend itself to such use with some enhancments such as flexible rigging, better geometry and enhanced UV mapping for the avatar. I am of the opinion this can be done without breaking current content significantly. Content also needs to be scaled to realistic dimensions, but that are minor adjustments in the big picture.

LlazarusLlong wrote: From what you say, Gavin, it sounds like LL have to choose between a platform where a satisfactory product might be developed, but for which the market is unlikely to pay a price acceptable to LL, or a platform where the market might pay the price, but on which the product could not be satisfactorily developed. The risk of entering the iOS market is low(er) both because there is significant unified volume (100 mill devices per quarter) and the willingness to pay is quite high. The offset is development cost and time to market. Siginficant portions of the development cost would have to be absorbed by that market alone since the reuse of all interface code + a renderer written in metal could not be reused for Android development. It could, however be used in the Mac market, but the volume is only around 25 million units/year at the moment. On the other hand the halo effect btween iOS and OS X is significant. The risk of entering the Android market is too high as it stands in my opinion, and going to consoles would be pointless. It is a fact that PC volumes have been declining year over year, and the gamer portion of the market has probalby largely been absorbed by SecondLife and OpenSim in combination, but OpenSim is lacking in content. Horizontal scaling of SecondLife and opening up to Hypergrid transfers would transform the system substantially to both make it more attractive for commerce, creation and cultural scaling while also lowering cost (see my early posts for explanation.) It would also preserve the current user and content base.

I agree with you that tech development will bridge the envelope, but so much so that this is also the case for mobile devices where the maker of the GPU for Apple's A-series processors a few days ago announced a moblie GPU with 512 execution units, again significantly increasing the graphcis processing pwoer over the A7 and A8 processors. What is currently possible on the A8 can be seen in this video In my opinion, it is therefore a better business plan to move the bulk of the processing to the mobile client, even though the envelope is a little bit narrow at the moment. The risk of moving into the mobile space is of course the competition for attention is fierce + when entering the Andriod space, the platform is highly fragemented and is full of underpowered systems. The willingness (and ability) of Android customers for payments for services are significantly lower than for iOS customers. The financial risk of moving into iOS is lower, but the investment in client development is substantial in that all client side code essentially have to be rewritten in Swift + Metal, and the abysmal state of this can be witnessed in the state of the Mac viewer which is till being built with a compiler Apple deprecated 4 major operating system version back. One of the latest "features" of the 3.26.x version of the LL viewer was the ability to build it with Xcode 6.1 while we are on 6.3 for latest mobile development. It is still built with a hack to get GCC running in that environment. Early on in this thread (maybe page 10 or so) I have outlined an alternative to build on the current SL and widen both the use case, base and still make it scale reassonably well. I share your view that they may be on a path of throwing the baby out with the bathwater by defocusing on SL proper wile offering little to no migtation to SLv2, essentially asking the customers to hit the reset button.

It is not so much where the signal goes, as it is to the very substantial cost you add to the equation. 1. A normal SL session has a bandwidth requirement of 50-200 Mbyte/hour depending on how much the avatar moves around and how dynamic the environment around it is. 2. Streaming the same session from SLGo would add 2Gbyte+ per hour for reasonable quality (more than 800x600) Scaling this up to even a concurrency of a few thousand would put a massive strain on any datacenter network connection and is something you only would see from CDNs like Akamai delivering canned content like movies or music to end point consumers. Data center server don't have more graphics processing capability than what is required to run a low rez admin or terminal session. These servers are packed as densely as possible in the datacenter to reduce (bulding) cost. To be able to stream the SLGo session, not only must the data center run and render the scene but it must also compose, comress and stream it to the network. The scene cannot be rendered on the blade server CPU, and there is not room in the chassis to place graphcis cards, so 1. They will have to install server hardware that takes 3-4 times the space as the normal server, meaning significantly higher space cost 2. The power consumption for the server goes up perhaps 6-8 times per served endpoint 3. The data center needs to cool off 6-8 times as much heat as before which adds to the power requirements and is problematic in multiple ways in terms of space efficiency in the data center Who will pick up all this additional cost? - That's SLGo's problem? To propose a business plan like that to be board of any sensible company would get you fired in a second. It may work for very small scale deployments. For a scalable solution, not a chance.