I posted this earlier today on the Frontier-user list.
Choices aplenty, for now
- Here are the choices that exist today for people using Frontier or the OPML Editor, as I see it.
- 1. Don’t update your Mac to Mavericks. That’s what I’m doing. I depend on the OPML Editor for my daily work. I’m concerned that one of Apple’s automatic updates will update me to Mavericks since it’s a free release. So I’m avoiding those as well.
- 2. Run the Windows version of the OPML Editor in VMWare or Parallels.
- 3. Run the Windows version of the OPML Editor on Linux using Wine.
- 4. The Windows version seems to be in decent shape for the foreseeable future.
Funding development work
- About doing the work to get the codebase to run under Mavericks, one of the original guys tried to do this, but gave up. Too much had changed. The GDI on the Mac has changed, as has the networking model. We had been using deprecated APIs for a long time. They finally pulled the plug on them.
- However — if you want to pursue that approach, it’s best to use the tools available like Kickstarter or Indiegogo and let’s pool the money and then try to attract the developers.
- I would say a bounty of $25K would probably get some interest. More would probably be even better. 😉
- But we’d have to be careful to employ someone who has the programming ability to deliver. There are a lot of people floating around who say they’re gutsy programmers, who would probably not be able to do this port.
- Frontier is a pretty complex piece of software internally. But it is organized in layers, so the work probably is fairly compartmentalized.
- I didn’t want to say that the guy from the original team that worked on Frontier code recently, trying to get it ready for Mavericks was Brent Simmons, but he said it on his blog, so I can.
- And he wrote a great post detailing the issues. Anyone considering this project should read this post.
- I think we should be able to find a machine someone could use for the “short path.”
- Re the database being converted to 64-bit, I think that would be pretty easy actually. The places that depend on the address size being 32-bit are pretty well factored. And it would a huge deal to be liberated from the 2GB limit on the database size.