ReactOS 0.3.14 Released | Hacker News Reader

ReactOS 0.3.14 Released

reactos.org

110 pointsjhack34 comments

Having followed this project for around a decade, it amazes how slow it is being developed compared to Haiku.

On the surface it looks like windows 2000, but as soon as you try and do anything useful, you bump into problems. Firefox hangs the system when loading. Fonts don't always draw correctly. LOTS of things are missing. Quite often i press a button on a dialog, and it does nothing - why was is it even enabled if it doesn't work?

Still a long way to go!

I have been watching this project for 7 years already; why doesn't ReactOS simply target running IE7, 8, 9? if they could do that millions of people would use it in VMs.

Unfortunately WINE and WINE derivatives are not in a state where they can be trusted to render pages as IE would render pages. WINE changes very quickly. The drawing code is no exception; they've spent the last year or so integrating a DIB engine, which will be among the flagship features of 1.4.

WINE is not likely to reflect a pixel-perfect representation of what your IE users will see and that's the simple fact. I've been bitten before by trusting the output of an IE in WINE. For testing IE, there is still no substitute for the official Windows platform. I don't think there will ever really be one, but it may be a decent distant-future goal for WINE: consistent pixel-perfect rendering in every IE version.

I think that's because it's an order of magnitude more complicated than Haiku. BeOS was fairly new and well designed when they started. Windows however has nearly 25 years of evolution to reproduce.

ReactOS is remarkably functional considering though!

I still don't understand the motivation for this project. If you ought to run a windows application, you use windows anyway. If you like to move to an opensource platform, you have to invest in migrating your applications anyway.

I don't get where ReactOS fits in this picture.

also how people can spend their precious time in re-implementing an existing architecture? There's lots of interesting and creative projects to actually build something new.

> There's lots of interesting and creative projects to actually build something new.

Can you recommend one that involves building a kernel, from scratch, that at least _some_ people will use?

excuse me for answering with a question to a question, but

why would you need a new kernel? There's BSD, there's this list here: http://en.wikipedia.org/wiki/List_of_real-time_operating_sys... , there's GNU Hurd... why do you need another one?

Besides, if you really want a new kernel, why should it be win32 binary compatible?

Excuse me as well for answering your question to a question with a bunch of questions.

> why would you need a new kernel?

1. Why wouldn't you? Requirements change. Would you expect humans 100+ years from now to be hacking on BSD?

2. What if it's (kernel hacking, that is) the only domain you enjoy working in? Different programmers have different interests. Should you just give up to work on something `popular'?

> Besides, if you really want a new kernel, why should it be win32 binary compatible?

3. How else would you expect to find users and contributors when many people already have their own $OS_OF_CHOICE?

Also, consider the year of the initial ReactOS release: 1998. A FOSS WinNT clone would've made sense at that time.

:-) actually 1998 voids your argument for "building a kernel from scratch". It's not any more a new kernel in fact :)

see my other comment in this thread about the business case. Without a business case, any exciting project would eventually disappear.

> see my other comment in this thread about the business case.

OK, I read it and I think you've a point.

> Without a business case, any exciting project would eventually disappear.

I don't understand what you're arguing here. What is the `business case' for GNU Emacs?

ok, name it "motivation" instead of Business case.

for GNU Emacs, the motivation is completely clear: it's a tool for developers, and some of those developers contribute and make it better for everyone.

For ReactOS... ok, someone seems to have a motivation, but that's completely beyond my understanding :)

You either build something new or you build something people will use. Building a Win32-compatible kernel is not something new. It's only a reimplementation of something very old.

It's not saying it's not an interesting project in itself - it's just not something I would attempt

simply because there are less boring things :)

> If you ought to run a windows application, you use windows anyway. If you like to move to an opensource platform, you have to invest in migrating your applications anyway.

This is a false premise. Even before ReactOS to get traction, there was WINE. From your point of view WINE "doesn't fit" either?

And people are spending their precious time in something actually useful for the entire world. They canonize a proven and currently used (in massive proportion) architecture and take it out of the Microsoft's whims. They put their time and effort in a public solution which is an alternative to Windows, yet not a sadomasochistic option like Linux & related (trying to shove down to mainstream user's throat something that was not designed for him). Finally, they do something that is obvious and had to be done a long ago, and I'm shocked how there could be people questioning the even the reason for ReactOS's existence.

(thanks for not pulling down my karma :))

I am in fact an open-source developer, but there's some business case behind each of my project. I just don't see a business case behind ReactOS.

I know about WINE (never used it though), and it makes at least a bit of sense to me, unlike ReactOS.

Not everyone is driven by a business case. Some people want to help feed the poor, some just want to tend their rose garden all year long, and some want to provide a compatible but open source alternative to the OS provided by Microsoft.

a project of this scale needs proper funding, otherwise it would take forever (which actually seems to be the case).

the guys are building a competition to Microsoft. Some may do it for the pleasure of doing that, but I see it as a dead-end, as there is no business driver.

Think of Ubuntu and the money they get from support, then think about having the legacy of Windows.

if I try to think as a typical Windows sysadmin, I would just use Windows. For those guys, ReactOS is not bringing any benefit, and they will for sure work with a product that they know already. They would only see ReactOS as a source of unpredictable problems and costs.

Nice. I won't pretend I've been following the project closely, but it really is an interesting and powerful thing they're doing.

does it work on a bootable flash drive?

I think somebody got it to install, but if I remember correctly, USB support is not yet in stable releases.

I just tried it and doesn't boot. Would really like to have it on a flash drive, shame it doesn't work yet... :\

If you get the livecd image, you can use Yumi to chainload the bootloader from the ISO file. Just run Yumi under Windows and have it add the iso to your thumb drive.

The second they launch a stable version I can use to replace windows XP embedded will be a real milestone.

Lots of plant touchscreens and industrial computers run a stripped down version of windows NT 3.51, win2k or XP (XP embedded is just a stripped down XP) which dont need any bells and whistles, just a solid network connection and the ability to run the win32 api. I have over 30 of these panel pc's around various factories running touch screen apps in place of explorer.exe so the desktop is not important to me and the demand is expanding all the time.

Thanks for the time and effort lads, I'm downloading it now to give it a go :-)

Not sure why you need ReactOS if that's the only requirement. You should just put stripped-down Linux installs in their place and run WINE, on which almost all applications work perfectly well. ReactOS in fact is mostly a kernel project. The majority of their userland support comes from WINE. The difference is that they are trying to clone the kernel to provide the truest Windows-like experience (including driver compatibility).

If you don't need the hardware support the reality is that you'll be much better off using a newer version of WINE on a Linux installation than you would be using a ReactOS snapshot.

Driver compatibility is likely to be more important as a lot of embedded systems use proprietary and certified drivers which will never work on Linux.

Not to mention, WINE is far from perfect. I'd argue that ReactOS, albeit with chunks of it coming from WINE is still far more likely to behave like a real Windows machine.

At the end of the day, it'll probably come down to cost and an in situ replacement of the OS without any application rewrite or hardware changes is cheapest.

>Not to mention, WINE is far from perfect. I'd argue that ReactOS, albeit with chunks of it coming from WINE is still far more likely to behave like a real Windows machine.

I would really like to know why you think this. ReactOS's userspace support is WINE almost entirely. They may write extra win32 applications like explorer.exe for use with ReactOS, but they do not use a separate runtime translator. They may modify portions of WINE to work with their ring 0 graphics code or whatever, but almost all of the serious work is done by WINE.

WINE is tested broadly on Linux platforms and the bugs usually get solved quickly. ReactOS is barely tested at all and I would not expect the kernel to be remotely stable for a long time yet. Why do you think ReactOS, which is usually not using a very recent WINE base (at least not in snapshots), would be more stable than Linux + WINE?

Thanks for the effort, indeed. It's a hope for a more sane free world we will live in!

Is the cost of a Windows XP Embedded license really a significant part of the total cost of a custom touchscreen computer install in a factory? I'm genuinely curious.

To get were they are is an amazing achievement, but it still looks like Window 95.

Edit: that's in the FAQ https://www.reactos.org/en/about_userfaq.html#changegui

It's a Free, binary-compatible clone of WinNT. You can now run (old) Windows apps with no license fees paid to MS, and without any proprietary code on your system.

edit2: did I mention it's more secure by default than a normal desktop Windows install? WinNT has some pretty nice security mechanisms, if you use them right. As a trivial example, users aren't admins by default.

From the release notes: "ReactOS now has the infrastructure needed to theme the user interface and shell, allowing users to install and use something besides the classic Windows theme."

Another thing noted on the homepage to keep in mind -- "Please bear in mind that ReactOS 0.3.14 is still in alpha stage, meaning it is not feature-complete and is not recommended for everyday use." That being said, I gave it another shot in VirtualBox with the recommended guest settings (Windows NT 4) and I was pleasantly surprised of how smooth it felt. This project has come a long way.

Reply on news.ycombinator.com