Microsoft approached IBM to work with them on the XBOX 360 CPUs and in a choice between Apple and Microsoft, it wasn't difficult for "Big Blue" to decide who their priority should be.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
The ******tification of the internet killed PPC.
- Thread starter DearthnVader
- Start date
- Sort by reaction score
Microsoft approached IBM to work with them on the XBOX 360 CPUs and in a choice between Apple and Microsoft, it wasn't difficult for "Big Blue" to decide who their priority should be.
You are right about 2008, I did not notice that in models table. The point still stands, IBM was way ahead of what Apple had with Intel by 2007 even in terms of bare frequency.
Though G5 at 2.5 is still faster than Xeon at 3.0, so that 3 GHz thing has a purely marketing value.
Is that right though? I always felt my MP 1,1 ran rings around my PM G5. Even my janky Banias based hackintosh Dell outperformed my G4 PB 1.5, though both had similar clock speeds.
Oh is it time to break out the Precisions? I'll need to reimage my Mac OS back onto mine though; it's got Haiku on it right now!
The internet is not a commodity. The internet is infrastructure.
Users, on the other hand, are a commodity. In rendering users as a product for advertisers to consume, post-web 2.0 platforms devised — invented — a commodity.
This, of course, raises the “security through obscurity” variable into the discussion, as these are PowerPC Macs we’re discussing.
I doubt folks still use their PowerPC Mac to, say, log on to their bank site (well, at least one might…).
And yes, other social (networking) services, in which users are the product, are especially vulnerable to security shortfalls. But unless one is being, say, spearfished (because they’re of some known, high importance), then the exposure risk of using, say, a 2004 Power Mac G5 for online services — from ssh to playing YouTube videos — is not going to be terribly high. The incentive for a hacker/cracker, if out of curiosity, may be there, but it certainly isn’t for extorting capital.
All due respect, but this reads like a condescending talking-to by a ne’er-do-well relative who think they’re being constructive and helpful. None of the regulars on this forum need that lecture.
And if one’s seriously holding bitcorns still in 2024, a virtual currency backed by nothing, then the joke’s really on them at this point.
I will make no claims re G4, it does feel slow. PM G5 – well, we can test stuff, if anyone is interested (and time is there). I am not suffering at all with the Quad even today. (Of course, something without proper hardware and/or software support may be slow – but this holds for every platform.)
I had MacPro 2.66 back in the days, did not feel any “wow”, but cannot compare directly, since I got PM G5 years later MacPro was gone.
It's got to be a Latitude, like what I had or it's not a fair fight comparing desktop CPUs with mobile ones. This is bringing back bad memories of hacked installer discs like Jaz and iAtkos and the fact that you never got full acceleration with the Radeon 9200 under OS X as the only drivers were PPC.Oh is it time to break out the Precisions? I'll need to reimage my Mac OS back onto mine though; it's got Haiku on it right now!
I was intrigued - not enough to drag out my G5 and 1,1 (now a 2,1 after a makeover) but enough to see if there were Geekbench comparisons just to give a very rough ballpark figure but I'm not sure the versions both ran were the same, so comparisons would be invidious.
The point is that the entire web has moved to TLS/HTTPS, and largely disabled HTTP, for good reason. The fact that obsolete hardware and software become less useful is an unfortunate consequence, not malice.
No, but if the options to turn off encryption, MFA and identity verification were available, lots of people would do so out of laziness or ignorance, and there is no incentive for any browser manufacturer or web service provider to provide support for a minuscule number of user that do not even use their antique hardware and software as their only/primary computing device.
OP wants to "just use the internet and an email client". I am pointing out that the reason this is difficult is not only new protocols, new browser features, ads, tracking, unnecessarily resource-heavy web pages, sloppy coding and high-resolution video, but also long-overdue security measures.
DMARC/DKIM/SPF has largely made spam and quite a few scams go away. Digital ID has significantly decreased the number of people that have their identities stolen and their bank accounts emptied by largely no fault of their own. EMV/NFC has made skimming a non-issue. These measures need to be forced on all users, it cannot be a choice.
You can still trade a BTC for 56000 actual dollars. 🤷♂️
Sure, the concept is silly. I would never consider crypto for an investment, for many reasons. However, it is not all that much more silly than the concept of state-backed money. Neither has any value if nobody trusts the system.
Regardless, I don't think it is a joke, and I do not believe anybody deserves to have their property stolen through a spoofed e-mail, app or website.
Last edited:
I was certainly not thinking of malice as the intent of that drift.
But this discussion is happening during an interesting moment, an intersection wherein:
- there are multiple ways to use internet-based services on obsolete hardware with the current transport layer security revision (TLS 1.3) and openssl;
- a renewed interest by both retrocomputing enthusiasts and folks younger than when the last PowerPC Mac left the factory on the non-securitized web, whether out of novelty, or as a critical look at the Rube Goldberg-complexity of “doing commerce securely” on a physical infrastructure and atop logical protocols (like http and the WWW) which were never designed or intended for it;
- a desire (although from folks who, again, never had the experience or privilege to experience life without access to an internet appliance or even the internet itself) to disconnect and find ways to spend time in person with others locally;
- the Moore’s Law cycle is all but over with, with cramming new cores onto a wafer being the current workaround to this physical limit (and a renewed consideration for better optimizing code to mitigate code bloat and code written when Moore’s Law had not yet reached its sunset; and
- near-countless examples everywhere of obsoleted hardware running current OSes (whether OCLP/dosdude1 for macOS or any number of currently maintained Linux flavours for 64-bit and even a few 32-bit).
I am mindful, even factoring all these considerations (and probably others I’m not yet aware of), these aren’t the things which can bend the inertia of prioritizing internet infrastructure, principally, as anything other than a wealth extraction and exchange medium — tailored, principally, for those who use only the newest (and, increasingly, pre-disposable) hardware available.
But it does usher forth a necessary conversation, now only being taken more sincerely and seriously, on de-growth strategies and of the inevitable move from a linear economy to a circular one, if we hope to still have a habitable place to live by the time the “the COVID kids” (born after 2020) find themselves ageing out of life in a hotter, less hospitable, and more severe world in which food growth and supplies are expected to dwindle.
So the point of re-use and re-purposed use for what’s already built and functioning, hardware-wise, is growing more relevant as we all move forward.
Again, whilst not discounting the utility of remedial kludges like MFA and ID verification in order to keep a growth economy humming along (with consumers continuing to use their hardware to buy all of the things and to park their monies in places which will avoid public revenue generation), one would also be foolhardy to ignore how these workarounds, cryptographic measures for the benefit of commerce, and so on carry real carbon budget costs, even when using the most efficient of new hardware.
We have long premised the notion of perpetual growth on a belief that all that inefficiently expended carbon, to facilitate economic growth, wouldn’t have long-tail costs which wouldn’t come due. When we as a society design and build hardware whose utility is dependent on a cryptography (making transfer of ownership more arduous, even impractical, and eliminating modular parts replacement as components in a piece of equipment wear out), we are not designing for the future. We are designing for maximum wealth extraction (and subsidizing losses by kicking it down the street).
What I’m getting at here is it’s easy (and intuitive) to concentrate on what’s before us right now. What’s less so is having to look, as a starting point, to several years or generations ahead, to consider how that intuitiveness for how we do things right now (like measures to constantly release hardware designed with heavy planned obsolescence, to facilitate buying and selling things on the internet, and/or relying on “The Cloud” solely) may not be in the best interest of that future.
It’s not a fun picture to look at, especially for anyone preoccupied by the impact of quarterly earnings statements and dividends paid to their investment/retirement account/pension. It is, nevertheless, a picture we all need to consider for long-term perspective on those who will survive us. It’s a responsibility we all need to undertake.
The features, tracking, ads, code bloat, and so on are, again, concentrated on A) short-sighted, “right now” economy; B) the myth that growth, whether carbon-positive or not, must be and can only be perpetual, because capitalism; and C) coders and, often enough, their employers aren’t taking the urgency for de-growth and circular economies very seriously in their lax attitude around code optimization, stripping bloat, and prioritizing backward- and forward-compatibility in elegant ways or, at minimum, to gracefully degrade (an idea with some traction a dozen years ago, but along the way was ditched for reason A).
I wish the industry lots of luck there.
Indeed. And in April 1929, or March 2000, or February 2007, impatient investors still believed in an economy, one built upon the act of speculation, could never fail.
Considering the above-mentioned hard limits on growth called Earth (the only one, not the additional six needed to continue the resource extraction, growth, and per capita consumption of G7 members, extended to the whole world), we are still under the illusion that an extraction-based, wealth-aggregation economy — under an aegis of capitalism — is anything other than the longest, most complex Ponzi scheme the planet may ever know.
But that’s another conversation for another time.
State-backed money, depending on the state, persists because it often is backed, at least in portion, by physical reserves of exchange. That’s why state-backed currencies persist as they have for, well, millennia.
Cryptocurrency is backed by what, again? And in the event of a catastrophic, whole-systems failure of an interconnected digital economy (like, Carrington Event-level, or worse), how might crypto step in?
I concur!
But how we ask and approach these questions as a society, and how we answer them — whether we want to take a “right now” view of answering them or a long-term, circular economic, and whole-system approach — should shape how we’re developing secure protocols with widest compatibility in mind.
Looks like macports is back up. https://ports.macports.org/search/?selected_facets=categories_exact:gemini. I'll see if any of these compile as is.
cgmnlm and offpunk build and presumably work (these are ports which I added to MacPorts, but I only tried to open some page to confirm they are not broken, which worked fine).
lagrange won’t build as is; I have it building locally, but something with the app not finding resources folder was there, and I never got back to fixing that. If you are interested in this app, I will dig out portfile I used. It should build, but won’t work as is.
There's actually a set of problems that makes it hard for macOS to keep up when there are forks of Chromium for Windows XP out there that work as long as you have a Pentium 4 or better. For example, macOS is maybe the most moving target there is, which is impressive considering how short the lifespan of precompiled executables can be on Linux. But between every transition, dropping 32-bit Intel, Carbon, Rosetta and Classic, that's all within the support life of Windows XP, that just shows you how long XP lasted in one form or another. Windows XP SP3 was getting its last updates just a few months before macOS Catalina came out. And SP3 still worked on ancient hardware from 2000.
But it goes deeper than that, because Windows 11, deep down, isn't that different from XP or even NT 4. That makes backporting newer software a lot easier than it is on, say, Linux. It's not like macOS is any better on that front, most Linux ports have all the same issues they would on Linux.
The funny thing about being able to run Chromium in Windows XP on a Pentium 4 is that the P4 is a terrible chip. The Pentium 3 was better in almost all ways, and the P4's replacement was entirely based on the P3. So, it's not like the AIM Alliance was the only one struggling with the death of Moore's Law. It's just that the P4 introduced the SSE2 instruction set, and both Chromium and Firefox require it in their current x86 versions.
As for how well any of that runs? I don't actually know, I've not got a P4 sitting around to try to browse the web with.
But this should give you an idea what a near minimum spec XP system is capable of.
But it goes deeper than that, because Windows 11, deep down, isn't that different from XP or even NT 4. That makes backporting newer software a lot easier than it is on, say, Linux. It's not like macOS is any better on that front, most Linux ports have all the same issues they would on Linux.
The funny thing about being able to run Chromium in Windows XP on a Pentium 4 is that the P4 is a terrible chip. The Pentium 3 was better in almost all ways, and the P4's replacement was entirely based on the P3. So, it's not like the AIM Alliance was the only one struggling with the death of Moore's Law. It's just that the P4 introduced the SSE2 instruction set, and both Chromium and Firefox require it in their current x86 versions.
As for how well any of that runs? I don't actually know, I've not got a P4 sitting around to try to browse the web with.
I remember Steve Jobs saying at the keynote moving to Intel was the performance per watt and the future roadmap for Intel was far better than PPC.
Ok, so lagrange builds as `-gui +tui` (yeah, it is set up in a silly way, should be fixed so that just `+tui` works correctly), but does not work, since it cannot find its resources. I have no idea why, and upstream did not suggest anything either.
`+gui` is broken, perhaps will need either restoring a working Cocoa code or otherwise tweaking it to build Unix-style.