Probably not worth the effort and hassle for such old and obsolete hardware when a more recent Mac model (even if 2nd hand) would be a far more sensible approach... I mean, one can purchase a MBP7,1 for under $200/200€ these days. Same for a MacBook5,1/MacBookPro5,x which will support Sierra and High Sierra with a little patch for direct installation. All GMA 950/X3100 systems have had it, that's it; they're 10 to 12yr old...
Yup. I really appreciate all the work that the OS X hackers have done to get later versions of MacOS to run on unsupported earlier Macs, and I've benefited from it. But often, depending on your Mac model, and peculiarities with your own particular Mac, the amount of effort one can put into getting a Mac containing a GMA 950 or X3100 GPU (and I assume some of the other GPUs that aren't fully supported by versions of MacOS not intended for Macs that contain those GPUs, even after patching the OS to run on those older Macs) to run Mavericks or Yosemite properly (and stably long-term) may not be worth it, as is shown by so many comments posted to these threads--the amount of time and frustration spent is usually better spent either making or saving a little money to buy a later Mac model that for many people is just the next year's model, and which won't have unsolvable issues like no sleep, no brightness adjust, and no graphics acceleration (which many people should want even if they don't use apps which benefit from it the most, because some parts of MacOS rely on it to work properly). I understand the desire to save a little money, or just keep an old system running, or just to see if it can be done on a particular Mac, but people need to understand the issues they might encounter.
As far as better Macbooks go, both the Early 2009 and Mid 2009 Macbooks are 5,2 models that contain an NVIDIA GeForce 9400M GPU, natively support El Capitan (10.11), and will also run patched versions of Sierra and High Sierra that work perfectly when installed using dosdude1's MacOS Sierra Patcher or MacOS High Sierra Patcher (at dosdude1.com/). If you can find an Early 2009 Macbook, it should sell for no more than about $70 if it's in stock condition, though it may cost more if it contains 4 gig of RAM and a large hard drive or an SSD drive (which makes a huge difference in speed). A Mid 2009 Macbook still for some reason sells for about $150 or sometimes more, but the Early 2009 model is nearly identical except for being a tiny (imperceptible) bit slower (2.0 GHz vs 2.13 GHz).
I've installed Sierra on my daily-use Mid 2009 Macbook using dosdude1's patcher, and it worked very well--no bugs that I could find, or at least none that affected me, and then later I installed High Sierra using dosdude1's other patcher, with the same successful results.
On the other hand, let me tell you (part of) my saga, in greatly edited detail, attempting to install Mavericks and Yosemite using MacPostFactor, so you can decide if possibly facing these issues is worth it to you.
I've been wrestling for the past few days to get MacPostFactor 2.0.1 (I also tried 1.0.1) to try Yosemite, and then Mavericks, on a Macbook4,1 (Early 2008 model, essentially the same as the Late 2008 model) that relatives in my household use for web browsing, using Chrome. This Macbook had been running Mountain Lion 10.8.5 just fine, which I'd installed a few years ago using MLPostFactor v0.3--at that time, MacPostFactor hadn't been released yet. Recently I saw that MacPostFactor had been released in June 2015 (I hadn't bothered to check for some time), so I decided to use it to replace my MLPostFactor version of Mountain Lion with a MacPostFactor version, thinking that MacPostFactor must be superior to MLPostFactor, even though I hadn't been seeing any issues. I guess I thought it might fix issues that weren't obvious. Within a day of using MacPostFactor to install Mountain Lion (10.8.5) on this Macbook, the Chrome web browser began to show terrible window redraw problems, especially when trying to scroll a web page. I couldn't update Chrome to see if that would fix this, since the Macbook was already running the last version (49) of Chrome that runs on Mountain Lion, and using Safari wasn't an option since Mountain Lion's version of Safari is too old to work properly with a lot of web pages, so I decided to try using MacPostFactor to upgrade the OS to Yosemite.
It turned out to be a huge task, and the patched Yosemite never worked right. I wanted to leave the Macbook's internal drive's files untouched at first, and install a clean, new MCPF-patched Yosemite onto a Firewire drive so I could boot the Macbook from it and try it out before deciding if it was right for me. Getting a bootable Firewire drive took a while--though some people have no real issues trying to do these patched installations, I had almost nothing but trouble. The first several installer boot volumes I created using MCPF, onto a couple name-brand USB flash drives that I'd formatted using Disk Utility as GUID, wouldn't boot (I say several because I had to try several times, onto different volumes on these flash drives, before I got one to boot), and then once I got one or two MCPF-patched MacOS installer volumes that booted, the MCPF-patched MacOS installer couldn't see any volumes available to install onto (this might be a bug only with MCPF 1.0.1, but I haven't pinned this down yet). I found the fix for this later, while reading through MacRumors forums on MCPF--someone found that the MCPF-patched MacOS installer gets confused by punctuation characters in volume names, and when it sees one on any volume connected to the Mac, it fails to create its list of target volumes. One of my USB flash drives had volume names that contained commas, periods, and parentheses, so first I removed the commas, but that didn't help, so then I removed the periods and parentheses too, and the MCPF-patched MacOS installer was then able to see all the attached volumes (most of which were on the USB flash drive).
There were many other weird issues preventing me from getting to where I could install Yosemite onto the Firewire drive, but I finally got there by trying every combination of steps I could think of. When I booted the patched Yosemite from the Firewire drive, I saw that it was version 10.10.2, which dates from January 2015, instead of the last version of 10.10.5, which dates from August 2015, so I must have downloaded it some time ago. I saw so many video glitches, both in redraw of regular OS elements (window title/grab/tab bars, system animations, scrolling, etc.) and in the jerky performance of video players online (YouTube, etc.), that it was unusable. Some people don't see these video glitches (or not as many) even on some other Macbook4,1 models, so I don't know why I did, except that many Macs and OS installs have unique issues. I don't know if these problems were because I installed 10.10.2 instead of 10.10.5, but maybe I'll download 10.10.5 and try that. Because what I was seeing wasn't usable, I used Disk Utility to erase the Firewire drive's patched Yosemite volume.
Next, I tried to create a patched Mavericks installer USB flash drive to repeat the experiment with Mavericks, but MCPF again failed over and over to create a bootable install volume onto either of my two USB flash drives, until it finally created one that would boot, but its patched MacOS installer couldn't see any volumes available to install onto, because this was before I learned about the MCPF-patched MacOS installer's problem with punctuation characters in the names of attached volumes.
So I finally decided to try MCPF's option to install the patched Mavericks directly onto the Macbook's internal, booted drive, from which I was running the MCPF app, which you're supposed to be able to do, according to the instructions. When MCPF asked me to locate the copy of the Mavericks installer app, I showed it, in the Macbook's internal drive's Applications folder, which again the instructions say you're supposed to do. MCPF then created a Recovery volume on the drive and booted from it to perform the installation, but the installation process immediately failed. I rebooted the Macbook from a Firewire drive backup of the Macbook's internal drive, and found that MCPF had erased the drive's regular boot volume without warning, which is one possible reason why MCPF failed to install--it had erased the volume where the copy of the Mavericks installer app was located, before beginning to use it to do the installation. This is incredible for obvious reasons, not least of which is the lack of a warning that it will erase the target volume, and that you may lose data unless you have it backed up--general warnings about potential data loss posted on the website aren't good enough.
I almost gave up, and used Disk Utility to wipe the Macbook's internal drive to remove the funky MacPostFactor recovery/installer partition, in preparation for copying the backup back to the drive. But I decided to try, one last time, one of the patched Mavericks installer volumes on one of the USB flash drives that hadn't worked before, and for some reason it finally decided to boot on the Macbook, and it saw the Macbook's internal drive, so I let it install. Oddly, the progress windows displayed by this installer were all standard Apple installer windows--there was no sign of MacPostFactor. The resulting installation of Mavericks is OK so far--speed is good (especially with the SSD in this Macbook), Chrome 65 is working fine with no redraw problems, video players work (though I've only checked a few within Chrome, and some seem to be dropping every other frame, which is tolerable to me), I have normal sound, etc. But it's not perfect, due to not having a proper sleep mode, no brightness control, etc. And that sleep mode problem is worse than most people report--if you close the display, the Macbook will sleep in a normal amount of time, but when you open the display, not only will the backlight be off, but the display itself won't even show a faint unlit image, so it's worse than just not having the backlight running--if that were the problem, you could use a flashlight to see the screen image, and use the trackpad and keyboard to close your work and do a proper restart or shutdown. As it is, with no image onscreen, your only solution is to press the power button to force the Macbook to power off. Good luck with unsaved work. A partial workaround for this is to install the NoSleep utility to prevent the Macbook from ever sleeping--instead, just a dark gray screen will appear at sleep time, from which you can "wake" the Macbook by pressing a key on the keyboard. This works even if you close the display. Unfortunately, the backlight remains fully lit even while the display is showing dark gray, and even if the display is closed, since NoSleep can't overcome the fact that the patched Mavericks has no control over the backlight, and so the battery is drained at the usual rate by the power-hungry backlight. MacPostFactor 1.0.1 seems to install NoSleep, since it contains the package for NoSleep, but MacPostFactor 2.0.1 doesn't contain this package, so you have to download and install it manually (it's available at
www.macupdate.com and elsewhere). But if someone is fine with the backlight being lit all the time, and that you always should shut down the Macbook before closing the display so excess heat doesn't build up that could theoretically fry the logic board (since the patched Macbook won't go to sleep even with the display closed), then this kind of installation might be OK--it's been OK for my family members who understand these limitations. I knew from reading through the forums about most of these problems, and that they weren't really fixable, so they don't surprise me.
But if I'd known the process was going to be this difficult, I would have skipped it, reinstalled Mountain Lion onto the Macbook's drive using MLPostFactor v0.3, and bought an Early 2009 Macbook on which I could natively install El Capitan, or High Sierra using dosdude1's patcher.
I didn't try MacOS Extractor, because there's nothing on the MacPostFactor website (and I didn't see anything in the MacRumors forums) saying whether the resulting patched MacOS is any better than what you get from MacPostFactor, and the instructions for MacOS Extractor are much more elaborate than for MacPostFactor.
And all of the above isn't even the half of it.
Your mileage may vary.
