Modern App development and PowerPC

What would be good is a low resource (or rather, variable resource depending on settings) maintained web browser built with the same skill and efficiency as Coreplayer - that would be worth paying for!

 

Windows XP still lives in the form of Windows Embedded which will be supported at least until April 2019. There have been successful effort / hacks to retrofit the Embedded security updates back into XP SP3. I wish Leopard received the same amount of love.

iOS developers have been enjoying the fact that the userbase has been adopting new versions rapidly: it only took 5 days for the iOS 7 market share to cross the 50% mark back in the day. So there hasn't been much incentive to support the older versions.

The other factor here is that the developers are known to love the fastest hardware: shorter compilation times means more work gets done. A few thousand dollars every few years for a new Mac is a pragmatic investment. And fastest / latest hardware again means no love for the no longer compatible OS versions. Yes you can run a VM but realistically it won't pay the bills.

 

Well, is that even possible? Yes, G5 (at least dual and quad) are powerfull enough (today), but otherwise, web is an overkill for these machines. And it will be worse each and every day. For example, simpler web is still OK in browsers like Dillo even on a G3 laptop, but load anything with javascript or multimedia of any kind.. and you (your PPC PC) are dead.

 

That's kind of my point. In the same way Coreplayer was specifically written to squeeze out every last drop of power from the G4, a browser in the same fashion would be more capable than what we currently have. But yes, even if such a browser existed, it would only be a matter of time before that folded under the weight of the WWW.

 

Well, I can imagine a browser using a beefy cloud service on the backend to run, for instance, Firefox, then streaming the contents back to the desktop. In the same spirit as https://github.com/sidorares/vnc-over-gif

 

A real avenue of longevity for PPC machines would be helper hardware - either PCMCIA, USB or PCI based processing units. You might argue, why waste money on such add ons when it's less expensive to buy a better machine? All I'll say is look at the Amiga community, it's now possible to emulate an Amiga (for free) that far surpasses any original Amiga and yet there's still an increasing market for Amiga accelerators/add ons, including the new Vampire boards that emulate the Amiga chip set in hardware with FGPA.

 

Well, we'd have to default to MOBILE WEB, for starters. That alone strips so much junk from websites that it makes web browsing on PPC quite tolerable. That's partly why Classilla, of all browsers, can run a lot better/faster on some tasks than TenFourFox.

We're at that point where regular Firefox will no longer compile on PPC, much less run with any kind of workability like "real" Firefox. (Sorry, Cameron "don't call me Camster" Kaiser, but I have no better name atm for "real" FF). Roccat, while it doesn't have the Firefox issues, has Safari issues and can never progress beyond the stock Safari unless you relink it against Leopard WebKit. And LWK may very well run into an OS-specific hurdle that impedes or destroys development completely.

Perhaps we can get CorePlayer efficiency in TFF now that we're on Feature Parity releases instead of The Camster working (and hoping) on getting the newest to PPC. Or, for a temporary solution, use Erik's TFF suggestions in the sticky here.

--- Post Merged, Aug 6, 2017 ---

I had to keep that last paragraph. It's just that good.

I looked at it the opposite way UNTIL 10.6 hit the market. It wasn't until August 28, 2009 that PowerPC truly died. The writing was on the wall at that point, and it was PowerPC that ultimately suffered. 10.6, as I've said before, should have been the Last Gasp of the G5's, much like 10.5 was the Last Gasp of the G4's and 10.4 was that for the G3.

They just had to get that App Store in there somehow, right? And making it Intel-only made sense (because supporting two architectures like that is brutal and leads to consumers running away because something didn't work).

Could App Store have been saved for 10.7? Yes. Should it have been? No. G5 users could just be told the same thing that G3 users got told during 10.4--it's not for you. It would have been a slow and painful death, but said death would have been 20 times easier on the PPC world.

 

Very true, people are still writing for the Amiga.

Exactly. The shift to Intel was inevitable as the G5 roadmap faded out of sight but it was the audacity of treating those who'd invested heavily in state of the art PPC hardware like suckers and realising it was all about marketing and money afterall that was hard to swallow. Lesson learned.

All TFF tweaks only have a limited effect - nothing short of turning off javascript gives you the speed bump - and that breaks 99% sites.
As you say, mobile sites win if you're on low end hardware, I can get by even on G3 by mixing browsers but it'd be nice if there was one scalable browser for all PPC sizes.

 

There were rumors flying around about a PPC 10.6 over on some of the older Hackintosh boards. If it does exist, it's locked up pretty securely along with Intel builds of 10.1, 10.2, 10.3, and a vanilla 10.4.

 

It would be novel but would it bring much more to the table now - still being an outdated OS?
What would be good is either a dedicated, optimized attempt at a Linux distro or a flavour of Morphos that wasn't so Amiga-centric (or licensed to one machine at a time). Problem is, even when someone attempted a brand new OS, they came here seeking support and were largely rubbished for even trying.

 

Doubt that there is a usable version of 10.6PPC otherwise Apple would have released it at least to calm some of the fury from its pro users and businesses that had invested heavily in G5s. The decision to ditch PPC was probably taken even before Leopard was out the door.

As for the Marklar builds, they wouldn't be much use without the prototype hardware Apple tested them on. Doesn't really matter whether they leak or not.

 

Yep, and then it's a question of IF that old code was ever kept...

 

I don't even think there was any to be perfectly honest. What SL became was a further optimisation of Leopard without the overhead of replicating functionality for PPC. It was mostly under the hood improvements and stuff like Grand Central Dispatch might just have been possible on G5s but I doubt the rest of the polish was. I would guess the very earliest builds of SL were just stripping PPC code from Leopard without breaking anything.

It went to $55 IIRR if you wanted it on a fancy small white flash drive, such as was bundled with the MBA 2010. Back then, a lot of people had issues downloading GBs of data over the internet with data caps or just slow connections.

 

Great post and I agree 100%. I think things will change though, as with the 8/16bit scene, as time moves on the hardware will become more scarce and it's use will be in the domain of an even small niche of enthusiasts with (presumably) more disosable income, which will provide a workable target market.

Regarding a new OS/new apps, I'm not thinking in terms of a modern parallel to other systems as clearly PPC hardware can't cope, I just mean a tidying up and repackage of what we already know works eg ad blocking, script blocking, mobile site/user agent spoofing, Youtube-dl, Fox Boxes etc etc. It's not a great stretch to get all this working as is but you see so many newbies arrive here, ask "can this be done?" and then disappear when a long list of instructions follow. I'm not a supporter of the Linux ethic of 'you have to get your hands dirty or you don't deserve to be here' - not everyone wants to delve that far in.

 

No, I've no clear image nor do I know how difficult it is - not being a programmer, but for example, @Lastic did a great job with SMtube but it requires separate installation/config of Xcode and Macports which put people off, an installer there would be great.
Or a variant of TFF that doesn't have a lock down on plugins and can call Quicktime to play video in browser or have a pref pane to choose user agent/javascript settings on a site by site basis like Omniweb used to.

 

I'm a professional Mac developer as well as a collector of old Macs (PowerPC and 68k).

I dropped PowerPC support in the last app I work on about 3 years ago. I started development of that app on a PPC Mac running Tiger in 2006, so the codebase was relatively easy to keep PPC compatible for a long time. That wouldn't be true with a codebase started a few years later using blocks, ARC, the modern ObjC runtime, etc.

It's not that supporting the PPC architecture itself is hard, rather, supporting Leopard means giving up usage of (very nearly) all of the many new and enhanced APIs in newer OS releases. For my own apps, the number of active PPC users is near zero, so it really doesn't make sense to spend extra time supporting PPC at the expense of new development that will benefit everyone else. I wish the situation were different, and I went out of my way to support the minuscule portion of my user base on PPC for a lot longer than most Mac developer, but it just isn't worth it anymore.

I do occasionally do development on my PPC machines (as well as on NeXTStep), but nothing serious, just for fun.