On October 18th, Jessica Lyon poked me about an upcoming blog post she was preparing for Firestorm which would make mention of a 64-bit Windows build, offering me the opportunity to talk to her about it ahead of the announcement going public.
At the time, my schedule was such that I couldn’t get back to Jessica immediately, so by the time we did get things worked out, the official blog post announcing both the team’s immediate plans for their next release and the arrival of 64-bit flavour of the viewer had been published. However, this didn’t stop me from taking the opportunity to sit down with Jessica and members of the team at the Cheeky Tiramisu Café late one afternoon in order to find out more about the promised “Firestorm 64”.
64-bit versions of SL viewers have been in demand for a considerable period of time. There has been some degree of resistance to them in the past, although there are a number of developers and self-compilers who have produced their own 64-bit versions of one viewer or another. The resistance has been for many reasons; Windows viewers are already Large Address Aware, for example, allowing them to use the additional memory common to computers using the 64-bit version of the operating system, thus helping to negate one of the biggest reasons for developing a 64-bit build.
Given that 64-bit builds have been seen as potentially problematic in the past, I started by asking what had prompted the Firestorm team to decide to go ahead with one.
“Our Windows 64-bit code was developed by Nicky Dasmijn as a sort of side project she wanted to do to scratch an itch she had,” Jessica informed me. Nicky, who started-out contributing code to the project, is now the project’s Lead Developer. Sadly, she’s also a little camera-shy as well, and managed to successfully escape my conversation with the team, hence why her profile picture appears here.
Jessica went on, “None of us were expecting her to drop the code into the repo when she did; but since she did, and since we had already decided to do a public beta, I figured, ‘Why not? Let’s get it out there in alpha form to see public reaction, and to see what the cost versus benefit might be’, neither of which we know for sure yet.”
I noted that when discussing 64-bit viewer builds at a recent Firestorm Q&A, there were concerns from the team about potential issues with maintenance, such as bugs and additional regressions, and for how it might negatively impact support were they ever to try for 64-bit viewer versions. I wondered what else had changed, other than Nicky working on the code herself, to persuade the team to push ahead on the 64-bit front.
“The expectation is that the 64-bit version won’t have different bugs than the 32-bit,” Jessica replied. “In fact the hope is that it may have better performance and fewer crashes, which if true, should actually take some load off our support team. But we don’t know for sure as we’ve only tested it on a dozen or so computers.”
I wondered if trying to offer a 64-bit version of a viewer might be the proverbial catch-22 / can of worms situation: the viewer needs to be put to public use in order to see what the response to it is like, but if it is put into public use, it’s going to be awfully hard to prevent it becoming an accepted and expected version of the viewer.
“Well, the feedback will determine whether we move forward with it, but I think chances are good,” Jessica said, before giving me a wry smile, “As for the can of worms; yes, we’ve opened it, and we’re not going to be able to get those worms back in it now. Folks are going to want it, many will want it even if there is no noticeable benefit.
“But other TPV’s are also working on 64-bit windows too, I spoke to Latif [Khalifa] from Singularity and found out he also has coded up Windows 64-bit for them. So to add another metaphor to the mix: the cat is out of the bag, and x64 for Windows is going to happen with or without us, and due to user demand it will likely become a standard presence going forward.”