Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
The Discoveryd fiasco is a good example of purging good code in an attempt to make a self promoting reputation and failing.

The original Mac OS X discovery code was written by a very experienced developer whom knew the TCP/IP stack down to the bit order of packets and how the weight of 1's and 0s of a specific stream can affect antenna impedance while still managing the stack performance.

They knew all the inferred performance rules of the TCP/IP standards and the bizarre WLAN rules dealing with wireless impedance issues. It was also written in straight C with almost all the code size and logic path optimization turned off so the compiler would not remove the intentional round about logic needed to handle these wireless TCP/IP timing issues. He was that damn good.

Here comes a new college graduate (with very little hardware knowledge) following academic coding fads and trends into their dream job. Concepts such a object-oriented design, code reuse, automated logic path optimization and so forth is preached to make the code more "efficient."

New systems development manager took the pitch. Discoveryd was green lighted to replace some of the most rock solid Mac OS X code going back to the NeXT days. This new code was not written in C but a higher level language that abstracted the final binary where, unless reviewed intently, the developer was not aware of all their permutations of logic automatically applied for such a heavily called piece of system code. Thus "inefficient" code was removed without knowing it's real intent.

Typically code like this is stopped from public release by a good internal QA check. For whatever scheduling or management issue, this bad code got out. The new college graduate took a reputation hit and the original code is incompatible with an new internal scheme of OS X networking.

This teaches a hard lesson. If an old technology is kept in use, it has survived years if not generations of change where the "new" stuff did not deliver or was misapplied.

At least they are still avoiding writing any system code in Swift letting that academic post doc project stay in the third party realm. Been told most of the OS Swift calls are just wrappers around Objective-C or straight C functions doing the heavy lifting.

I had a feeling it was this kind of scenario that pushed it out. At my age I've met enough of these types to just shake my head. To cocky to questions their actions and intent. A Gift for Gab to sway the powers that be. Oh well, at least Apple realized it and jettisoned that junk.
 
I had a feeling it was this kind of scenario that pushed it out. At my age I've met enough of these types to just shake my head. To cocky to questions their actions and intent. A Gift for Gab to sway the powers that be. Oh well, at least Apple realized it and jettisoned that junk.

From what I was told is the old discovery code was rolled back in with an API rapper with the new internal networking scheme. Then a long hard resource review was adjusted.

Keep in mind, the current C compiler source code that's licensed to this day still has lines written by Dennis Richie himself back in the 70's where it first ran on a PDP-8.
 
  • Like
Reactions: kagharaht
Agreed. I don't understand why they have to push yearly OS updates, especially as the OS is free, so it's not like the shareholders are pushing for more bucks...
Apple sells hardware. They market new software features or optimizations (e.g. battery life) as if it's part of the hardware.

Thus the necessity to have major iOS and OS X updates every fall - at the same time they release most of their new hardware.
 
I think a part of the yearly release cycle is planned obsolescence -- "let's make sure iPad 2 doesn't work anymore so people upgrade, ha ha". El Cap seems to be a rare exception, judging by positive reactions to the betas from owners of older machines.
 
I think a part of the yearly release cycle is planned obsolescence -- "let's make sure iPad 2 doesn't work anymore so people upgrade, ha ha". El Cap seems to be a rare exception, judging by positive reactions to the betas from owners of older machines.
More like hardware cannot keep up with software
 
  • Like
Reactions: Tucom
Try this, I've had this work work well.
In Wi-Fi settings remove all the preferred networks, turn Wi-Fi off and back on, then reconnect your network/s.
This will hopefully clear all the previous settings and it should work properly
Thanks, I'll give it a try.
 
It's ironic that now that Yosemite is finally smooth and relatively bug free, here comes El Cap with a good deal of to be expected bugs. Not sure that is really going to be progress. Seems like they have to break things in order to move forward all the time. Don't quite get why...
My Yosemite is far from bug-free. I have the weirdest problems all the time, and I've reinstalled it twice now. Even the installer itself was buggy; it took about 6 hours to install just because I had stuff installed by Homebrew and MacTeX in /usr/local. I haven't had a decently stable machine since Mountain Lion, and my last really stable setup was with Snow Leopard.
 
  • Like
Reactions: jezbd1997
More like hardware cannot keep up with software
You're think that allowing users to install a system update that turns their tablet from a perfectly functional machine into a flat, shiny snail is fair and it's not Apple's fault the machine can't keep up with the software?

Interesting.
 
You're think that allowing users to install a system update that turns their tablet from a perfectly functional machine into a flat, shiny snail is fair and it's not Apple's fault the machine can't keep up with the software?

Interesting.

Economics of scale. Enjoy your Linux box.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.