Apple Seeds Second Beta of macOS 10.14 Mojave to Developers

[albebaubles]

But that had nothing to do with a rewrite (or lack thereof), just with poor planning and lack of automated testing.

there was no automatic tests. You didnt waste the 2 digits for century because storage was at a premium then.

They dont teach you computer history either I see.

Have they taught you about page faults?

 

[bugfaceuk]

But that had nothing to do with a rewrite (or lack thereof), just with poor planning and lack of automated testing.

Oh god no. Dude. No.

They were saving the memory space. No-one thought that those systems would still be in use. They assumed they would be re-written by then, but as we've learned over the decades that almost never happens. Google "Technical Debt", it wouldn't exist if it wasn't for the fact that everything is an iteration on the previous version rather than a re-write.

If you are looking for evidence, rather than just arguing, then just look to the DNS issues in Mac OS a few years ago. They rewrote the DNS (just one component of the whole of macOS) a few years back and it caused all colours of problems. That is proof that rewriting even significant components is rare, let alone the whole OS.
[doublepost=1529455487][/doublepost]

there was no automatic tests. You didnt waste the 2 digits for century because storage was at a premium then.

They dont teach you computer history either I see.

Have they taught you about page faults?

So now I'm the old man saying that all of this was fields? Is Computer Science in the process of forgetting what it had learned?

Argh. I'm getting too old for this s**t.

 

[albebaubles]

So now I'm the old man saying that all of this was fields? Is Computer Science in the process of forgetting what it had learned?

Argh. I'm getting too old for this s**t.

It IS cyclical...when the iPhone SDK came out those of us with experience in limited resource had a huge lead on those that did not. ARC, SWIFT -- created so the masses can write ****** code.

 

[anthogag]

Rotation is still only 90 degree steps. Why isn't there an unconstrained rotation?

Also, when I have several files picked (highlighted) in Finder, often I would like to see the contents and information of a file before I add it to the picked group. If I click the file I lose the highlighted group. Maybe if I mouse-over the file it could show it in a pop-up preview window.

 

[macsba]

In the Developer notes, was there any mention why Mac Pro 5,1 with Metal GPUs can't run Mojave for beta 2?

 

[imageWIS]

First one was buggy as hell had to downgrade to HS, hopefully this one will be better

Which is painful as HS is no picnic.

 

[syclonefx]

My installation has been stuck at the same spot for the last 2 hours.

 

[GtrDude]

Copying path has been CMD+option+C for some time now. In any case, I've been using it for what feels like years. Opening a new finder window with the same location is a bit more tricky. Anyway, in 10.14 you can set up an automator step for that and have it on our touchbar or something

Thanks for the info.
Very good.

 

[Macbookprodude]

Hopefully, someone out there can get this to run on unsupported macs, such as the 2012 MacBook Pro non-retina, though i THINK I may jump ship from Apple.. I don't like their direction and without Jobs, the company is not going to have a future as someone in there will mess up.

 

[Superhai]

First one was buggy as hell had to downgrade to HS, hopefully this one will be better

Yeah, that is how it is supposed to be, but remember they may release other features as well in subsequent betas. Those will inevitably introduce other bugs.

Oddly the current High Sierra release is fine. So what exactly are they breaking .. I mean fixing that is worth breaking everything in the process. I don't get it. Why not start from stable and get better as an idea? It defies logic.

New features will introduce bugs, optimizes and fixes as well. Code is typically written by people and people make mistakes. Some are really stupid...in hindsight, some are really hard to discover, and some are archaic ones which wasn't really triggered before or masked by other stuff.

This conversation is always coming up every year when Apple (or anyone else for that matter) release major revisions, especially betas.
[doublepost=1529467763][/doublepost]

OSs are usually rebuilt from the ground up to include new features, changes, frameworks, etc.

Mostly there are only minor changes, although a few parts are rewritten. As an example you can find the source code for the xnu kernel macOS uses, at least the Intel part, it is available from the first to the very recent one, and you can see for yourself how some parts have remained more or less the same, but some parts could be entirely different.

Writing an OS from the ground up takes an immensely amount of time and resources, which is why most parts of the OS remain. It is the reason why one typically say that one OS is based on another one. Like macOS have parts from BSD, Nextstep etc. it simply means that source code from those are included and refined in a way Apple prefer. New large frameworks like Cocoa, Metal, etc are written from "the ground up" but as you might have discovered these are multi-year undertakings, and span several OS revisions.

When you actually start coding on various projects you will understand how the benefits and pitfalls of full code rewrites affect your projects.

 

[chucker23n1]

there was no automatic tests. You didnt waste the 2 digits for century because storage was at a premium then.

Yes, and automated testing would have increased confidence that upgrading to four digits didn’t lead to side effects. That’s the point.

They dont teach you computer history either I see.

Have they taught you about page faults?

What’s your issue?
[doublepost=1529468220][/doublepost]

Oh god no. Dude. No.

They were saving the memory space.

I know.

No-one thought that those systems would still be in use. They assumed they would be re-written by then, but as we've learned over the decades that almost never happens. Google "Technical Debt", it wouldn't exist if it wasn't for the fact that everything is an iteration on the previous version rather than a re-write.

Oh believe me, I know what technical debt is.

If you are looking for evidence, rather than just arguing, then just look to the DNS issues in Mac OS a few years ago. They rewrote the DNS (just one component of the whole of macOS) a few years back and it caused all colours of problems. That is proof that rewriting even significant components is rare, let alone the whole OS.

That is precisely the point I made a few posts back. What are you even arguing?
[doublepost=1529468300][/doublepost]

ARC, SWIFT -- created so the masses can write ****** code.

Nothing to do with the masses. Humans write code, and humans make mistakes. ARC solves plenty of common mistakes.

 

[Candlelight]

Looking forward to the public beta next week.

 

[73CortinaV8]

but there's no mac hardware new enough to run it

 

[john123]

I’m sure you’ve heard the old maxim, “Those who can't do, teach”. I think the fact they’re former devs is telling.

As someone that has over 25 years experience developing enterprise grade software, I can tell you they’re full of hooey. If that’s truly what they’re trying to communicate they’re either being deceptive (unlikely), or they really don’t have proper understanding of what it means to work on large projects over years (quite likely), or their propensity for rewriting everything is what got them pushed out of the tech industry and they don’t realize that was a contributing factor (also a possibility).

I also think there’s a perspective shift that needs to happen around the idea of what a ‘major’ project is. It is unlikely anyone in university will be involved with, or produce, a product that would be considered major in the real world (barring incidental exposure during internships). I’m not trivializing the work university students do, but nothing done at school is going to be on the scope of a project that requires dozens, scores, or even hundreds or thousands of developers, working in concert, for years, to develop a single product. And iterating releases over that product almost never involves total rewrites. Too many lessons would have to be re-learned, too many ‘gotchas’ would have to be rediscovered and solved for again, far too much code would be rewritten almost identical to how it had been before, and far, far too much development effort vs massaging what you already have and smoothing out the freshly introduced destabilizing bits.

It would be helpful for people to let go of this idea that a company is being silly for not embarking on a ground-up rewrite of something like a major OS. It’s virtually never done and for many good reasons. If Apple were to do it, it would take several years, at least, and would still require a large number of iterations to ‘get right’. That, and the fact that they certainly don’t need to, I don’t see it happening for a long, long time, if ever.

I suspect it’s not the professors who are full of hooey but rather the poster’s understanding of what those professors said. Anyone who has even the most basic and rudimentary experience with software development knows better.

 

[eoblaed]

I suspect it’s not the professors who are full of hooey but rather the poster’s understanding of what those professors said. Anyone who has even the most basic and rudimentary experience with software development knows better.

Indeed, which is why I prefaced my list of possible explanations with “if that’s truly what they’re trying to communicate”... a misapplication of an idea mentioned in class is quite possible, I agree.

 

[albebaubles]

Yes, and automated testing would have increased confidence that upgrading to four digits didn’t lead to side effects. That’s the point.



What’s your issue?
[doublepost=1529468220][/doublepost]

I know.



Oh believe me, I know what technical debt is.



That is precisely the point I made a few posts back. What are you even arguing?
[doublepost=1529468300][/doublepost]

Nothing to do with the masses. Humans write code, and humans make mistakes. ARC solves plenty of common mistakes.

Im saying automated testing wasn't even in existence at the time.

 

[Guy Clark]

There is still a very long way to go and those considering upgrading from High Sierra at this early stage would be taking a huge risk. That being said I updated from DP1 on an external partition and it went without hitch. One notable bug is with VLC 3.0.3 is you have to open a video file manually as it cannot be opened by selecting from the VLC interface. Also Steam although working is buggy.
Full Dark Mode on macOS Mojave is a huge draw for many and temptation to upgrade from High Sierra even at this early stage is considerable but treat Mojave as with any other new macOS release. I shall be waiting until at least two point updates following final release to ensure many of the bugs have been ironed out before using Mojave on my primary partition.

Update:

Video files can be opened from VLC interface if they are shown in list or column. This was not possible in DP1.

 

[gleepskip]

Im saying automated testing wasn't even in existence at the time.

Such children. Not understanding the scope of the problem in Y2K, much less comprehending the complexity of writing a new system from wholecloth.

Unless you own your own company with wads of capital to spare, good luck getting your bosses to front the cash to rewrite a working system. I don't remember how many times I've looked at a system, wishing I could rewrite the whole damn thing, either to replatform with modern toolsets, undo a tangle of poorly designed, cobbled-together code base, or had better design concepts that would make the system more stable and supportable.

In general, the most you can hope for is to chip away at a problematic system by piecemeal. Perhaps, eventually, you will have rewritten a system one component at a time. Even then, working on someone else's dime... good luck with that.

 

[kissmo]

There is still a very long way to go and those considering upgrading from High Sierra at this early stage would be taking a huge risk. That being said I updated from DP1 on an external partition and it went without hitch. One notable bug is with VLC 3.0.3 is you have to open a video file manually as it cannot be opened by selecting from the VLC interface. Also Steam although working is buggy.
Full Dark Mode on macOS Mojave is a huge draw for many and temptation to upgrade from High Sierra even at this early stage is considerable but treat Mojave as with any other new macOS release. I shall be waiting until at least two point updates following final release to ensure many of the bugs have been ironed out before using Mojave on my primary partition.

Totally agree with you.

I jumped on the upgrade wagon at this stage and I must say I loved the dark mode. It's pretty darn good.
Although - the problem is that I have had issues to the point that for some reason I could not boot anymore in Mojave so I rolled everything back.

You should definitely wait for the final release before installing it on any machine - Unless it's a VM or on a separate partition and you do have a purpose/reason for doing that.
Again, it's in pretty early stage and I think waiting for the classic AppStore update is the best thing to do.

 

[kemo]

Did they fix the option to boot from external USB 3 SSD on Macbook Pro 15'' 2017 with Touchbar? I was not able to (showed black screen with while booting). The very same disk booted without any issues on Macbook Pro 15'' 2015.

 

[adrianlondon]

My Calendar app no longer remembers the previous window size, and always opens too small for me. It was fine in Beta 1. I remember having a similar issue previously - maybe one of the Sierra betas. Not a big deal as it just takes a second to drag it to the size I want, but it's not right

Feedback logged.

 

[MarcoPohlo]

Were the maps dark in previous beta as well or is this a new feature?

Also is Allow your Apple Watch to unlock your Mac working for anyone?

 

[SwissGuy93]

Oddly the current High Sierra release is fine. So what exactly are they breaking .. I mean fixing that is worth breaking everything in the process. I don't get it. Why not start from stable and get better as an idea? It defies logic.

The problem is: OSs are really complex things. It wouldn't be wise to do a whole lot of optimizing while implementing new features. They mostly implement the features (where they maybe even have to break something else), test them, tweak them a little, and then they start optimizing / adding the finishing touches on them. But first they have to make sure that the features actually work as intended. That is why we see a lot of bugs in the first three to four betas. And then it gets better as things get optimized.

It's a tiny bit like open heart surgery. First you have to put the patient on bypass, then you operate and then you check if the heart starts pumping again. After that you can close him up. It wouldn't be wise to close the patient up before you know if his heart is working properly again.

 

[justperry]

There is still a very long way to go and those considering upgrading from High Sierra at this early stage would be taking a huge risk. That being said I updated from DP1 on an external partition and it went without hitch. One notable bug is with VLC 3.0.3 is you have to open a video file manually as it cannot be opened by selecting from the VLC interface. Also Steam although working is buggy.
Full Dark Mode on macOS Mojave is a huge draw for many and temptation to upgrade from High Sierra even at this early stage is considerable but treat Mojave as with any other new macOS release. I shall be waiting until at least two point updates following final release to ensure many of the bugs have been ironed out before using Mojave on my primary partition.

For people using their Macs daily professionally yes, as for the rest it's the most stable release I have been on since beta 1 and I have been on macOS/OS X since it's inception.

The problem is: OSs are really complex things. It wouldn't be wise to do a whole lot of optimizing while implementing new features. They mostly implement the features (where they maybe even have to break something else), test them, tweak them a little, and then they start optimizing / adding the finishing touches on them. But first they have to make sure that the features actually work as intended. That is why we see a lot of bugs in the first three to four betas. And then it gets better as things get optimized.

It's a tiny bit like open heart surgery. First you have to put the patient on bypass, then you operate and then you check if the heart starts pumping again. After that you can close him up. It wouldn't be wise to close the patient up before you know if his heart is working properly again.

Except you won't die when the OS is borked.

 

[chucker23n1]

Unless you own your own company with wads of capital to spare, good luck getting your bosses to front the cash to rewrite a working system.

Even if money weren't an issue, even if the engineers had nothing better to do, even if it were easy to estimate the scope of such an effort, it would still be a terrible idea, as hidden in the existing software to be rewritten are plenty of assumptions and workarounds that haven't been documented. Perhaps they (inadvertently or not) avoid bugs. Perhaps they were expected behavior from users. Perhaps they weren't originally intended, but have since effectively become public APIs. In all those cases, a rewrite would break some of them.

I don't remember how many times I've looked at a system, wishing I could rewrite the whole damn thing, either to replatform with modern toolsets, undo a tangle of poorly designed, cobbled-together code base, or had better design concepts that would make the system more stable and supportable.

In general, the most you can hope for is to chip away at a problematic system by piecemeal. Perhaps, eventually, you will have rewritten a system one component at a time. Even then, working on someone else's dime... good luck with that.

Nothing wrong with rewriting portions of the code, especially when there's strong refactoring help from tooling (especially feasible with static typing) and automated tests that at least try to document existing expected behavior.

But yes, definitely piecemeal.

And you will introduce bugs. And accidentally "fix" some.