ProcMon for Linux
github.com
github.com
Specifically, I hope they release their own branded Linux distro, built on Ubuntu/Debian, with an optional KVM-hosted NT kernel for Windows syscall support. Maybe call it Windows NC (for Next Century).
I'll probably leave MacOS behind at that point on my work machine, especially if they create a display-PS compositing window manager that's anywhere as good as MacOS's.
Microsoft had a lot of support for SUSE-specific products around that time, like SLES Support in System Center (https://social.technet.microsoft.com/wiki/contents/articles/...)
It does remind me of the old Microsoft Linux joke (http://www.mslinux.org/) of the early 2000's, which has aged more poorly then a younger me would have thought.
[0] https://azure.microsoft.com/en-au/blog/microsoft-showcases-t...
"I don't remember agreeing to that."
-- Janet Reno(I found this immensely amusing 20 years ago, and mirrored/restored this a decade ago, finding images from multiple wayback snapshots... )
MS Linux to have Start Button
Microsoft is working hard to incorporate the well known "Start" button from the Windows Platform into X Windows' Gnome interface. "We just can't figure out how the hell to get that darn foot out of there! The damn thing is like stuck." The team will have this feature ready by product launch.
SQL Server (already converted) MS Office suite (including Project and Visio) and Sharepoint Exchange (for Outlook's best features) ActiveDirectory (scalable auth + DNS + client management is still hard to beat, but MDMs are catching up)
Beyond that, all they really have is an open source platform (.Net), and a few open source projects (VS Code, PowerShell), and Visual Studio. Oh, and GitHub, I suppose.
Microsoft will not release their version of Linux for sale. They're going to try to make it a high-availability Azure-only service. They're still chasing Apple's and Google's services-only architecture and Amazon's cloud-dominance strategy. Windows will become their legacy server and user client product.
Had it not been for FOSS UNIX clones, they were probably right.
As a multi-decade linux user, the thought of a Microsoft distro still makes my stomach churn. The grand-daddy of software copyright enforcement branding the paragon of f/oss... it is just wrong.
But I bet it wouldn't last. MS's institutional inertia would soon see the thing fork away from linux, becoming something else. Something like android. Something more horrible.
I just don't like Windows anymore. So much of the modern development tools are Linux only, and depends on shell scripts. Running the WSL is not ideal either, even though it helps a lot.
Only thing keeping me from switching directly to Linux is the bad memories with the various Linux Desktops I tried in the early 2000s. Clipboard didn't work, font smoothing was very inconsistent, the fight with the (GPU especially) drivers, simple windowing animations was jerky, browser was always slower (e.g. scrolling on web sites, and videos).
Font smoothing is occasionally inconsistent between modern and ancient apps, but the same happens on Win. As for the clipboard, I've only seen clipboard issues on my ancient Arch install, which I'm surprised even boots up at this point.
Second, MS would need to port some sort of compatibility layer, which would probably be 10000 person-years to achieve anything like the backward compatibility they currently support.
Third … well, that would be it. The vast majority of non-techies would be ecstatic to find they can upgrade for free for ever, their software would just keep running, and hardware would work basically from day one (because the companies developing cutting edge hardware Windows would pivot their driver development) until the end of time (because FOSS drivers, while slower to appear for most hardware, are supported essentially forever).
It depends on what desktop environment / window manager you use.
I don't notice too many weird issues when I run i3 and bsp on my Linux laptop. I think the nice thing about a tiled window manager is if you go down this route, you often hand select a bunch of small components that do 1 thing very well. Even going as far as evaluating and picking a specific compositor (or maybe not even using one at all).
You are right tho about Windows. Its UI is very stable and buttery smooth for actions like moving and resizing windows. It has a level of polish that out of the box DEs on Linux usually don't have.
https://sourceforge.net/projects/linuxfxdevil/
https://b00merang.weebly.com/windows-10.html
Leveraging the FOSS work and cutting down on maintenance costs, and focus that only for LSW, to guarantee the functioning of that abstraction layer, and leave the harder parts (hardware support, security etc) onto public to pick up - at the cost of a few coupons, t-shirts, coffee-mugs etc thrown at them via "hackathons". (And even perpetuate the "passion for software" hence "public github contributions". May be give a slightly higher tier github subscription for free, they've bought that org anyway. And people are anyway suckers for Internet points)
ReactOS has been a thing for 20 years now and it just isn't that popular.
I see more and more people - and not just hardcore devs - using Linux.
For me, Linux has been a better desktop experience since around 2005 sometime: faster, easier to configure to my exact liking, easier to install, less hassle to maintain. (Except the time when KDE 4 was new and Ubuntu changed to Unity.)
For a long time I even though it was just a matter of time before everyone would realize but now I have realized that people are just different. Some people actually like Mac and Windows. Some people are driven mad by inconsistencies in my favorite DE (KDE) that I cannot even spot after someone tried to point them out. I however can't stand CMD-tab on Mac but Mac users - like me with the alignment issues on KDE - cannot see how it can be a problem.
Summary: it is not like any of them are way ahead in general. I see more and more people using Linux and expect numbers to rise a few percentage points more.
https://github.com/microsoft/ProcMon-for-Linux/search?q=tele...
https://github.com/microsoft/ProcMon-for-Linux/blob/70d5196e...
Is there anything this gives us that sysdig doesn't?
The entire point is move people away from Linux and back onto the gimped version of the platform provided by MS.
Case in point: https://news.ycombinator.com/item?id=23867754
Ya'll stuck in the Microsoft of 20 years ago. I hold no great love for Microsoft or its products, and large corporate power can always be abused, but screeching "EEE!" at the slightest hint with no evidence based on assumptions on what they might do in 5 years is just silly.
MS (like all corporations) only cares about one thing, making money. It's not silly to distrust a corporation that has repeatedly in the past stolen, subverted, and act in bad faith as soon as a 'threat' to their complete dominance has shown up.
Yes, most of the old-guard are gone. (Gates, Ballmer, etc) But I'm not entirely ready to trust them yet. (Windows10 telemetry, not honouring user-defined configuration [Peer-to-Peer updates active even though it's disabled in user-facing options], Candy Crush and all other game 'apps' being pulled from the MS store.)
So they aren't even acting wholly trustworthy now, even under this current management. Why would I trust them to be nice with Linux?
Extend by offering stuff exclusive to WSL, e.g. their directX and cuda stuff.
Embrace by releasing procmon, powershell, ... . <-- You are here.
Then offer nice additional features that happen to only be available on some blessed config/integration with AD/...
Extend by offering stuff exclusive to Android, e.g. their Java and NDK Native APIs stuff.
Embrace by releasing Play Store only APIs, ... . <-- You are here.
Then close down NDK to anything that isn't Android relevant.
... and that makes it better/worse/look a three-headed monkey is behind you?
How are you enjoying the BSD upstream updates done by Sony?
The time of rainbow colours and cheerful music is long gone for me.
The EEE debates are very boring and I understand it can feel like MS is being singled out but they deserve all of it. I often wish MS rises to the top again just so newer generations can experience first hand what that feels like. But now that Google and other companies perfected the boiling techniques I'm not sure it would be noticed.
I wish we could all stop fighting each other over these issues and just collectively bashed and scrutinized ALL the companies together, everyone would benefit. But that's easy to say, I can't even read about MS or Google without getting irrationally mad.
Then close down NDK to anything that isn't Android relevant. <-- You are here.
https://github.com/termux/termux-packages/wiki/Termux-and-An...
You managed to write 44 words without addressing that. But hey, at least you tried. Most EEE conspiracy folks usually stay silent when asked about the extinguish strategy.
WSL is only the realization that Linux killed POSIX and noways the Linux ABI is more relevant than POSIX.
Something that other UNIX clones like Solaris, illumos, SmartOS and FreeBSD also support.
This tool seems to be the first foray into the second ‘E’.
Extend would be forking the linux kernel and adding a bunch of interfaces for it that somehow have patents owned by MS so that they could not be reintegrated into mainline.
Extinguish would be to get all the distros to start using MS'es kernel and interfaces instead of mainline.
If we think back to the web, MS releasing software for linux is like them having a webpage. MS forking and adding proprietary interfaces to the kernel would be like all the DX filters and activeX apis added to IE. MS getting dominant players to use those interfaces would be like all the webpages having "Works best in IE6" logos.
This is still firmly in the first E. It could be argued that back when you could only develop Azure Sphere (their IoT linux distro) on windows it started to creep into the second E, but now you can develop for it on linux too, so perhaps that puts us back on the first E. The directX stuff in WSL is definitely a second E move though.
- This is EEE!
- No it's not, there is no third E
- Well, back in the 90s such and such happened
- Microsoft has changed! Look at VS Code!
- Electron is slow.
- Buy a better machine.
- what about the millions who can't afford a better machine?
- Electron can be good, just code better.
- but muh privacy
- exactly, VS Code is going to take over the world with it's knowledge about your coding habits
- It's all about Azure nowadays, anyway.
- Azure has terrible customer support, we should use bare metal.I miss having access to Windows games.
And I could build a PC as I wish. Weird times.
WSL2 has a linux fs which is fast and a bridge to windows fs which is still slow but I just live on the linux fs. WSL2 also runs native docker (and windows docker client and leverage the docker daemon in wsl).
It is not completely seamless but pretty good.
About to abandon that plan because it’s just too much of a pain. (WSL extension Locking files needed by npm specifically)
stopping now [-ferror-limit=] 20 errors generated.
https://github.com/microsoft/ProcMon-for-Linux/issues
There would have to be some limitations when not running as root, but it would still be useful for looking at your own processes. I always like it when an app can be installed "per-user" as an alternative to a systemwide install.
std::size_t calculateBackpressure()
{
return currentReadQueue->size() + currentWriteQueue->size();
}
Recommendation to MS folks: find someone at MS who knows C++. Barring that, at least run ASAN and TSAN on your stuff before you release it.
The “third e” is extinguish.
This seems more like a product of love from one developer. Doesn't seem like a cunning move in some grand scheme. Companies don't have personality and even company culture is very hard to maintain across many offices spread out over the globe.
Then again it is different people with a seemingly different playbook. I’m still slightly uncomfortable with it.
I am also slightly uncomfortable with that, but they are not the player they once were back in the 90ies, they definitly don't have an OS monopoly these days, and that is a good thing. Everyone who wants to use other distributions is free to do so.
They simply moved from selling windows and office licenses to selling office365 and Azure, and using open source helps them do this.
Ex. PyMSSQL is now officially supported by MS. This implies, their costs of maintaining an ODBC driver for Python-MSSQL will be largely reduced thanks to the OSS model.
Same deal with Linux-Azure tooling (there have been many tools that came out), or be it .Net-Core.
They are looking at "extinguishing" Linux not by doing better than competition, but rather by absorbing the competition and getting them to work for MS in the long run. Great business move, not sure for the larger ecosystem.
But then, OSS will always be there. If not Linux, something else.
In the meantime the world has changed, and Microsoft has changed. And still people keep referring to eee as if this is still the same Microsoft.
All the big tech companies are now making money with services.
Microsoft made a lot open to Linux. Why? Because they like you to be on Azure.
I really think they are not going to extinct Linux. They are going to hook you on their services and make a lot of money out of it.
I think it is different now though. Microsoft was trying to leverage their near-monopoly on the desktop into a monopoly in new markets. Now, although Microsoft retains the bulk of desktop/laptop market share, desktops and laptops are not as significant as they once were. Microsoft's attempts at the tablet and phone markets were largely unsuccessful. A lot of apps that used to be desktop-based are now web-based. For desktop apps, cross-platform development solutions (e.g. Electron) are a lot more popular than they used to be. The alternatives to Windows on desktop/laptop (macOS, Linux, ChromeOS) are arguably more compelling and more practical than they used to be. Linux, not Windows, dominates in the server space. In cloud, Microsoft is stuck in second place position. The generation coming up through the school system right now is more familiar with Google Docs than Microsoft Office. Microsoft's language platform (.NET, C#, etc) has decent marketshare but competitors such as Java and Python are ahead.
So even if some people at Microsoft want to repeat Microsoft's EEE strategy from the 1990s, Microsoft doesn't really have the same foundation that it had in the 90s to make such a strategy successful. And that 90s strategy, even if it worked in the short to medium term, doesn't appear to have been a long-term success – Internet Explorer may have killed Netscape, but arguably Chrome has killed Edge. Microsoft's attempts to EEE Java likewise failed – Visual J++ and J# are history, and Java is still more popular than .Net/C#.
JVM yea, but Java itself as a lang + frameworks/libs like Spring? I don't think so.
I may be biased, but also I've seen a lot of arguing about Java vs C# (as a langs) feature-wise
Which one is ahead in features is debatable. Certainly C# has some goodies which Java lacks, but more recently Java has been catching up.
Which one is ahead in marketshare/mindshare/ecosystem – I think Java is the clear winner here. In part because, despite Oracle's overall ownership, Java gets a lot of attention from other big vendors such as IBM and SAP and Google (putting to one side all the regrettable Oracle-Google legal issues), and a lot of focus from the open source community. By contrast, .Net is basically just Microsoft plus a boatload of small-to-medium vendors, and the .Net open source community has always been smaller than the Java open source community.
(Obligatory disclaimer: I used to work for Oracle, although not directly on Java.)
Microsoft is in the IBM position of 1985. IBM made a bunch of crucial missteps because they saw the PC as part of a larger mainframe and minicomputer strategy and was unwilling to cater to the microcomputer customer as just a microcomputer customer.
So they were blinded and dropped the ball a bunch.
Microsoft, like IBM in the 80s doesn't have full stack monopoly anymore. But unlike IBM, they have admitted this and most importantly that they can't claw it back. IBM ran off to fairytale land trying to claw things back with OS/2, the MCA bus, and their line of PS/2 workstations.
Microsoft knows this well, they built the first versions of OS/2 for IBM.
Like a good general who studies the battles of the past, they're trying to do the early Apple/early Google strategy: make good products and get developers to be enthusiastic about them.
Their will be new marketing opportunities opening in the future and their job is to position themselves to capture it.
They won't get there by pretending like it's still 1998.
The question is to what extent and in which areas Microsoft's interests and strategies align with ours. The answer to that question has changed in recent years, especially if we consider the alternatives.
Proprietary drivers are like getting an a third-party accessory like a new set of rearview mirrors for your car. However, you get your cars for free from the manufacturer, and complain when the mirror you bought doesn't fit the new car frame.
The ideal answer here is working with the manufacturer to get those issues fixed upstream. Where the comparison with car accesories stop, is that software drivers isn't the core product you are being sold, so there is no real reason for not upstreaming it.
Now, onto the real point the comment you are replying to made: the TUI interface has been very stable and good at backwards compatibility for multiple decades. New extensions exist for displaying images, getting mouse coordinates, more colors, etc. But the basic mechanisms have stayed the same, hence stability. Drop a greybeard UNIX tech strakight from the 80s inside a WSL2 environment, and they will be able to do their business as usual.
I'm getting better at working with MAC GUIs. Last time I used one (6 years ago), I was able to open a terminal myself to fix the issue the person had with their storage device. Unsure if the GUI changed much since, but I can likely still manage to do so from the CLI.
Also, can you give an example of a non-ugly MS cli app for windows?
And why would you want an example of a CLI app? That's the whole problem with Linux - you take some of the biggest software distributions for it and even they'll have you change some secret ascii file somewhere on the disk for it to work properly, while on windows you just click a few 'Next' buttons and you're done installing (before people start talking about package managers and how easy it is in comparison to Windows, I know about that, but they're honestly not foolproof). If Linux wants to get popular among masses (rather than a bunch of proud geeks who feel that the ones who use a GUI are all dumb people), it would have to be better at that front. Just updating a graphics card driver is so much pain on it - it would randomly not work, or it wouldn't update all symbolic links that you have to waste a day in finding, or always something like that).
Instead, you should give examples of the best looking GUI applications for it. Our company has a flagship product on both Linux and Windows and we use Qt for the UI. Exact same cross-platform code produces relatively ugly looking UI on Linux. In almost all discussions people start blaming both users and developers alike as if it's their problem.
That is either because of a Qt problem or a problem in your implementation, not with Linux in particular.
> And why would you want an example of a CLI app?
Because we are discussing an article about a cli app by MS and I thought your post was at least tangentially related. So please explain to me how this MS is somehow more ugly/pretty than other linux cli apps or what GUI apps have to do with anything in the article.
I agree Windows applications don't have a lot of consistency as well, but it's at least better than Linux where half are on terminal only, and half have some kind of casually created GUI.
Agreed about the casually created GUIs on Linux, but I'm happy someone has put the effort into them.
That's a lot of work for a simple tool when a text-based terminal interface solves the problem very neatly. Personally, I'd do it with strace, sysdig or something else that would spit a long text file I could parse rather than updating a terminal screen.
Good luck using a command line spreadsheet software and producing good results in time ;). I am a little surprised by your belief that only two classes of GUI software in the entire world are useful (because you use those two?), everything else is garbage. :)
That's your big assumption. It doesn't happen very often and, more often than not, those 5 steps you could do with one command without your hands leaving the keyboard will require a couple clicks of the mouse at not-always repeating spots on the screen.
Terminal users aren't masochists. They use terminals because it makes the work easier. I use an IDE to write Java code. I use Emacs to write Python code because it's easier to me. I use a terminal to investigate because it's much simpler than attaching a GUI-based IDE to a running remote machine.
And nobody seriously uses text based spreadsheets. We I need to crunch numbers or reads lots of data, I use Jupyter and pandas.
I would have chosen different colors though, but I also imagine they wanted to restrain themselves to the base ANSI color palette for compatibility reasons.
WSL1 yes I agree, it is just a translation layer between Linux and NT syscalls, and features such as eBPF would not be implemented in such a layer.
[1] there may at times be various enhancements to HyperV paravirt drivers that haven't been upstreamed yet