I do get it...

You can make money by painfully having to make better products and fix bugs so folks have an incentive to purchase the updated software, or you can license the software for a limited time and get people to pay to continue to use the software, thus not having to deal with making the product better or even fix anything, so you can get rid of that expensive dev team and just hope nobody else writes anything better than what you've paid to create.

Look at Office... it does pretty much everything most offices need. It doesn't need much by way of improvement at this point. If anything, people want Microsoft to just leave it the fuck alone so they don't have to learn how to do anything new or deal with them fucking up some feature they really loved because it just works for them as is. How the hell is Microsoft supposed to make any more money off that dead horse?
They aren't going to get that many new sales... so Office 365 comes out as a subscription model, requiring you pay for it again and again.

(This said, the code does need to get updated to deal with security problems... and it costs money to do this. You can't very well rely upon sales of new software to pay for keeping software secure, and because software security is a huge shell-game anyway, you'll always have to keep goofing around with the code, dealing with the next suprising security breach that someone creates).

Ragnar - From a consumer standpoint would you rather pay X fee all at once, or 1/9 X every year for 10 years? Yes, the second instance means you pay more money over time, but you don't have to pay it all out at once.

From the seller standpoint, they just need to keep you for at least 9 years.. SO they need to find a way to make that work.

And yes. I get that the real problem is when it's not 9 years to come out even but 2 or 3 that it really fucks the consumer. But many consumers can work a way to break the payment out over longer time as opposed to up front.

You can make money by painfully having to make better products and fix

bugs so folks have an incentive to purchase the updated software, or
you can license the software for a limited time and get people to pay

That's not a benefit of cloud; it's a benefit of continuous delivery (rolling releases). Cloud-hosted software does make that easier, but it isn't the driving force behind the model. If it were, the cloud people would be focused on that instead of on figuring out how they can best lock-in customers (but to make it more palatable, we call it "sticky" instead of "locked in").

Look ... I *like* the idea of putting as much as possible behind the glass.
I'm in the glass business myself. I have a problem with software providers who make their glass out of six foot thick iron bars.

Time to move to Gitlab.....

So it's true. GitHub goes to Microsoft for $7.5 billion. Nat Friedman will be GitHub's CEO.

I've wondered about that, too.

Although, there's a lot of code written around the idea of using GitHub for your repository, I've noticed. Several IDEs and whatnot that seem to cater to it in some unholy fashion that irritates me.

Now that Microsoft owns it, I wonder how much of that kind of thinking will go out the door, and folks will wake up to the idea that you can create git servers of your own, pretty much anywhere... and that decentralization is a good thing for everyone.

Hell, git is *designed* with decentralization in mind.

Since when does Microsoft take historical design decisions into account when upending everything that something is doing?

Apparently it gets even weirder than that. Last year M$ moved all of its internal repositories (including Windows itself) to git. Apparently their code base is so yuuuuge that checkouts and commits were taking hours, so they developed a filesystem overlay that does sparse checkouts and simulates the presence of unmodified files on the local filesystem by reading them directly from the repository when accessed.

Embrace and extend, I guess. Things have just gotten *weird* lately. I watched a video of a Micro$oft program manager demonstrating the virtual filesystem and when he wanted to show the changes in a file, he fired up vi as his editor.

It seems that Henry Spencer was correct when he pointed out in 1987 that "Those who do not understand Unix are condemned to reinvent it ... poorly."
DEC, IBM, Apple, and now Microsoft ... the inevitable lumbering of a clunky proprietary system towards the open unix world. With their shift towards being a cloud services company, M$ seems less interested in deliberately building barriers these days. Bill Gates (may he die slowly, cut into a thousand pieces) would never have allowed Micro$oft code to come into contact with something like git.

And yet, SourceSafe...?

Well, then, perhaps there is a god. That was one god-awful piece of shit.

God does not need a source code repository because he completed the entire universe on the very first build.

However ... I wouldn't mind downloading the source and making some adjustments.

2018-06-07 12:34 from IGnatius T Foobar
Apparently it gets even weirder than that. Last year M$ moved all of

its internal repositories (including Windows itself) to git.
Apparently their code base is so yuuuuge that checkouts and commits
were taking hours, so they developed a filesystem overlay that does
sparse checkouts and simulates the presence of unmodified files on the

local filesystem by reading them directly from the repository when
accessed.

Literally every large company that has an established multi-GB repo has done something like this. My first exposure was with Google's code-base, where they created a set of wrappers around Perforce that does, in effect, exactly this same kind of thing. Facebook does something similar; the difference is that FB contributed back to the Mercurial user community. You can find their contributions online.

Summary: this is not abnormal.

Also, God uses Lisp. He just hot-patches live on the server.

11 years ago I posted a short article in my blog (Skeptic Tank) about how Xenix would have been the sensible way for Microsoft to bring out a non-toy operating system, and in fact it was their plan until they went down the OS/2 route and then later brought in Dave "VMS r00lz, unix dr00lz" Cutler to design the albatross that is today's Windows operating system.

I'm thinking about this again today, but now I'm thinking about it in the context of our now largely post-Microsoft (except within the confines of the traditional desktop) world. What would have happened if, for example, they had built their next generation operating system on top of Xenix, instead of spinning off SCO?

The answer is that the Real Computer World might have been satisfied enough with Windows/UX (or whatever) that Linux might never have been created, and Microsoft might have ended up controlling the mainstream operating system *everywhere* (instead of just on desktops). And most people might actually have been ok with that. Sure, there would still be some BSD's kicking about, and they'd get used for things like embedded network stacks and niche Apple products, but most people wouldn't care. Instead of today's reality of one clear winner of the Unix Wars (Linux) plus one mainstream non-Unix OS, there would just be one mainstream Unix OS.

Funny how Microsoft was the architect of its own failure to monopolize outside of the desktop.

It also would imply that we'd have a much safer computing ecosystem as well; since Xenix would have been a Unix-derivate (and thus a POSIX-derivate eventually), it would have significantly fewer places for viruses and trojans to hide.
I'm not suggesting that they wouldn't exist -- existence depends on popularity, that's a given. I *am* saying, however, that anti-malware services on Windows seems to actually *beget* more malware in a kind of arms race, because MS keeps adding more and more places for malware to *hide*. With Xenix, there would be substantially fewer places for malware to hide.